deploy: publish on the LAN address so Caddy on .35 can reach it

127.0.0.1:8096 is unreachable from .35, which is where Caddy runs. Every other
service on .13 is published this way (8095, 8092, 3909, 4200) and proxied by the
same Caddy. Port is LAN-only; Caddy remains the sole public listener.

Makefile

@@ -23,7 +23,8 @@ BIN := bin
 DIST := dist
 
 .PHONY: build dist generate dev test validate validate-self-test fmt vet clean help \
-        corpus corpus-sync corpus-rebuild corpus-audit corpus-verify corpus-demo corpus-test
+        corpus corpus-sync corpus-rebuild corpus-offline corpus-audit corpus-verify \
+        corpus-demo corpus-test corpus-graph corpus-scan corpus-sources
 
 help:
 	@echo "make build            templ generate, compile bin/, no dist"
@@ -36,11 +37,15 @@ help:
 	@echo "make clean            remove bin/ and dist/"
 	@echo ""
 	@echo "corpus (Phase 4, python3 -m corpus.cli):"
-	@echo "make corpus           rebuild the corpus from content/ (drops the collection)"
+	@echo "make corpus           rebuild the pool from the estate (drops both collections)"
+	@echo "make corpus-offline   rebuild content/ only, no Gitea (recovery and CI)"
 	@echo "make corpus-sync      incremental update; only changed documents are re-embedded"
 	@echo "make corpus-audit     freshness + drift audit (exit 1 on findings)"
 	@echo "make corpus-verify    check the index against content/ and Qdrant"
 	@echo "make corpus-demo      same query with and without an area filter"
+	@echo "make corpus-graph Q=  query graph nodes and edges (with their source note)"
+	@echo "make corpus-scan      search the pool for credential-shaped content"
+	@echo "make corpus-sources   show the platform guides and the last harvest"
 	@echo "make corpus-test      corpus unit tests (no container, no network)"
 
 build: generate
@@ -87,6 +92,9 @@ corpus: corpus-rebuild
 corpus-rebuild:
 	@$(CORPUS) rebuild
 
+corpus-offline:
+	@$(CORPUS) --offline rebuild
+
 corpus-sync:
 	@$(CORPUS) sync
 
@@ -99,6 +107,16 @@ corpus-verify:
 corpus-demo:
 	@$(CORPUS) demo-filter --query "distributed audio endpoints" --area edge
 
+corpus-graph:
+	@test -n "$(Q)" || { echo "usage: make corpus-graph Q='photo pipeline'"; exit 2; }
+	@$(CORPUS) graph --query "$(Q)"
+
+corpus-scan:
+	@$(CORPUS) scan
+
+corpus-sources:
+	@$(CORPUS) sources
+
 corpus-test:
 	@$(PYTHON) -m unittest discover -s corpus/tests -t .
 

README.md

@@ -58,6 +58,21 @@ keeps acceptance test A1 true.
 The two things that are not files are `/search` (FTS5 over page titles and body text) and
 `/api/stage/stream` (the presentation panel driven by an event stream instead of a route).
 
+The retrieval corpus (Phase 4) is built separately, because it writes to a service rather
+than to `dist/`:
+
+```sh
+docker compose -f deploy/qdrant.yml up -d   # on the runtime host, loopback only
+make corpus          # content/ -> chunks -> Qdrant + SQLite FTS5
+make corpus-verify   # does the index agree with content/ and with Qdrant?
+make corpus-audit    # stale documents, index drift, contract drift
+make corpus-test     # 76 unit tests, no container and no network
+```
+
+The site never calls Qdrant, so acceptance test A1 is unaffected: stopping the corpus
+services leaves every page working. Operator manual and measurements:
+`docs/phase-4-notes.md`.
+
 ## Ground rules
 
 - **Do not touch other systems.** No stopping containers, no editing other projects, no `nixos-rebuild`, no Caddy changes except the single block specified in Phase 3 — and that one Sam applies.
@@ -72,7 +87,7 @@ The two things that are not files are `/search` (FTS5 over page titles and body
 - [x] Phase 1 — content source of truth (40 documents across six areas, `make validate`)
 - [x] Phase 2 — static site (`make dist`, 45 pages, 86 stage fragments, FTS5 search)
 - [ ] Phase 3 — public deployment
-- [ ] Phase 4 — corpus
+- [x] Phase 4 — corpus and retrieval (51 chunks, FTS5 + Qdrant, freshness audit)
 - [ ] Phase 5 — AI overlay
 - [ ] Phase 6 — guardrails
 - [ ] Phase 7 — polish

corpus/__init__.py

@@ -0,0 +1,22 @@
+"""Phase 4 — the retrievable corpus.
+
+`content/` is the source of truth. This package turns it into the two indexes
+the AI console reads:
+
+* **vectors** in Qdrant, one point per chunk, with the payload that lets the
+  query path filter *before* searching (`area`, `confidence`);
+* **keywords** in SQLite FTS5, one row per chunk, so exact tokens (repo names,
+  `LoRaWAN`, `MQTT`, dates) are not lost the way they are with vectors alone
+  (decision D10, hybrid retrieval).
+
+Everything here is driven by `python3 -m corpus.cli`; `docs/phase-4-notes.md`
+is the operator manual and records what was measured. The pipeline is
+Provider-agnostic by design: `corpus/embeddings.py` is the only place that knows
+how a vector is made, and the provider + model + dimension are recorded per
+collection so a provider change is a governed migration, not silent corruption.
+
+Standard library only, plus the repository's own validator (`tools/`). No new
+dependency, no Node, no framework (AGENTS.md section 4.1).
+"""
+
+__all__ = ["config", "loader", "chunker", "payloads", "embeddings", "qdrant", "store", "pipeline", "freshness"]

corpus/__main__.py

@@ -0,0 +1,8 @@
+"""`python3 -m corpus` is the same as `python3 -m corpus.cli`."""
+
+import sys
+
+from .cli import main
+
+if __name__ == "__main__":
+    sys.exit(main())

corpus/areas.py

@@ -0,0 +1,94 @@
+"""Which resume area a harvested document feeds.
+
+The payload `area` is what lets the query path filter *before* it searches, so a
+harvested file has to be classified once, deterministically, and never promoted
+to make a page read better.
+
+The classification is keyword evidence, in this order: the path, the repository
+name, the description and topics, then the languages. The result is constrained
+to the platform guide's `fits` list: a platform cannot feed an area it does not
+declare. When nothing matches, the first `fits` entry wins, which is why every
+guide lists its strongest area first.
+
+This is a heuristic and it is labelled as one. It is not a claim about Sam's
+career; the curated `content/` tree carries the claims.
+"""
+
+from __future__ import annotations
+
+import re
+
+from .payloads import AREAS
+
+# Order matters: the first area with a hit wins, so the more specific evidence
+# (an ESP32 repository is edge, not platform) is checked before the general.
+_RULES = (
+    ("edge", (
+        "esp32", "esp8266", "arduino", "lora", "lorawan", "mqtt", "sensor",
+        "zigbee", "z-wave", "relay", "gpio", "pir", "modbus", "iot",
+        "firmware", "breadboard", "raspberry pi", "rpi", "camera", "cam",
+        "temperature", "watering", "fish feeder", "driveway", "light",
+        "snapcast", "mopidy", "audio", "voice bridge",
+    )),
+    ("ai", (
+        "ai", "llm", "agent", "agentic", "rag", "embedding", "vector",
+        "model", "ml", "prompt", "graphify", "omniroute", "paperclip",
+        "langgraph", "prefect", "whisper", "piper", "chat", "memory",
+        "knowledge graph", "vision", "triage",
+    )),
+    ("platform", (
+        "nixos", "nix", "docker", "compose", "kubernetes", "k8s", "deploy",
+        "deployment", "backup", "borg", "kopia", "restic", "network", "dns",
+        "email", "caddy", "proxmox", "postgres", "terraform", "ansible",
+        "linux", "systemd", "homelab", "server", "infrastructure", "git",
+        "gitea", "monitor", "logging", "security", "firewall", "vpn",
+        "tailscale", "storage", "garage", "s3",
+    )),
+    ("products", (
+        "product", "app", "web", "dashboard", "react", "laravel", "flutter",
+        "kotlin", "calendar", "photo", "kontra", "store", "resume", "site",
+        "gps", "tracker", "api", "frontend", "ui", "ux",
+    )),
+    ("history", (
+        "history", "agency", "client", "archive", "old", "legacy", "2019",
+        "2018", "2020", "2021", "career", "resume copy",
+    )),
+    ("about", ("about", "profile", "bio", "personal", "cv")),
+)
+
+_WORD_SPLIT = re.compile(r"[^0-9a-z]+")
+
+
+def area_for(evidence: str, fits, fallback: str = None) -> str:
+    """Classify one document from its path, name, description and topics."""
+    haystack = " " + _WORD_SPLIT.sub(" ", (evidence or "").lower()) + " "
+    allowed = [area for area in (fits or AREAS) if area in AREAS]
+    if not allowed:
+        allowed = list(AREAS)
+    for area, needles in _RULES:
+        if area not in allowed:
+            continue
+        for needle in needles:
+            if f" {needle} " in haystack or f" {needle}s " in haystack:
+                return area
+    if fallback in allowed:
+        return fallback
+    return allowed[0]
+
+
+def area_for_repo(repo: dict, fits) -> str:
+    topics = " ".join(repo.get("topics") or [])
+    evidence = " ".join(
+        str(part)
+        for part in (
+            repo.get("full_name", ""),
+            repo.get("name", ""),
+            repo.get("description", ""),
+            topics,
+        )
+    )
+    return area_for(evidence, fits, fallback=(fits or [None])[0])
+
+
+def area_for_note(path: str, fits) -> str:
+    return area_for(path, fits, fallback=(fits or [None])[0])

corpus/chunker.py

@@ -0,0 +1,408 @@
+"""Code-aware chunking.
+
+The rules come from `docs/phase-4-corpus.md` §4.1:
+
+* 400–800 tokens per chunk, about 12% overlap;
+* **never split inside a fenced block, a table or a front-matter block.**
+
+The last rule is structural here, not a post-hoc check: the body is parsed into
+*blocks* first and chunks are packed out of whole blocks, so a boundary can only
+ever fall between blocks. Fenced code, tables, headings, list items and
+paragraphs are all blocks; the large ones cannot be subdivided.
+
+Front-matter is removed by `corpus.loader` before this module sees the body, and
+parsing a body that still has front-matter is an error, so it cannot be split
+either.
+
+`tokens` are estimated with a word-and-punctuation regex. A real BPE tokenizer
+counts 15–30% more for prose and more again for code, so a chunk that measures
+800 here is a conservative upper bound.
+"""
+
+from __future__ import annotations
+
+import re
+from dataclasses import dataclass
+
+# Packing targets (docs/phase-4-corpus.md §4.1).
+TOKEN_TARGET = 600
+TOKEN_MAX = 800
+TOKEN_MIN = 400
+OVERLAP_RATIO = 0.12
+
+FENCE_RE = re.compile(r"^ {0,3}(`{3,}|~{3,})")
+HEADING_RE = re.compile(r"^ {0,3}#{1,6}\s")
+LIST_RE = re.compile(r"^(\s*)([-*+]|\d{1,3}[.)])\s+")
+RULE_RE = re.compile(r"^ {0,3}([-*_])(\s*\1){2,}\s*$")
+TOKEN_RE = re.compile(r"\w+|[^\w\s]", re.UNICODE)
+
+PROTECTED_KINDS = ("fence", "table")
+
+
+class ChunkError(RuntimeError):
+    """The body could not be chunked without breaking the rules."""
+
+
+@dataclass(frozen=True)
+class Block:
+    """An indivisible unit of Markdown. Never split, only packed."""
+
+    kind: str  # fence | table | heading | list | quote | para | rule
+    text: str
+    tokens: int
+    line_start: int  # 1-based, relative to the body this block came from
+    line_end: int
+
+    @property
+    def protected(self) -> bool:
+        return self.kind in PROTECTED_KINDS
+
+
+@dataclass(frozen=True)
+class Chunk:
+    index: int
+    text: str
+    tokens: int
+    line_start: int
+    line_end: int
+    blocks: tuple
+    oversize: bool = False
+
+    @property
+    def kinds(self) -> tuple:
+        return tuple(block.kind for block in self.blocks)
+
+
+def estimate_tokens(text: str) -> int:
+    """Word-plus-punctuation token estimate. Deterministic and cheap."""
+    return len(TOKEN_RE.findall(text))
+
+
+def parse_blocks(body: str) -> list:
+    """Split a Markdown body into atomic blocks.
+
+    Raises ChunkError if the body still carries a front-matter block: that is a
+    loader bug, and silently chunking it would split something that must not be
+    split.
+    """
+    if body.lstrip().startswith("---"):
+        raise ChunkError(
+            "body still starts with a front-matter block; strip front-matter "
+            "before chunking (corpus.loader does this)"
+        )
+
+    lines = body.split("\n")
+    blocks = []
+    i = 0
+    total = len(lines)
+    while i < total:
+        raw = lines[i]
+        stripped = raw.strip()
+        if not stripped:
+            i += 1
+            continue
+        start = i
+        fence = FENCE_RE.match(raw)
+        if fence:
+            i = _read_fence(lines, i, fence.group(1))
+            kind = "fence"
+        elif stripped.startswith("|"):
+            while i < total and lines[i].strip().startswith("|"):
+                i += 1
+            kind = "table"
+        elif HEADING_RE.match(raw):
+            i += 1
+            kind = "heading"
+        elif RULE_RE.match(raw):
+            i += 1
+            kind = "rule"
+        elif LIST_RE.match(raw):
+            i = _read_list_item(lines, i)
+            kind = "list"
+        elif stripped.startswith(">"):
+            while i < total and lines[i].strip().startswith(">"):
+                i += 1
+            kind = "quote"
+        else:
+            i = _read_paragraph(lines, i)
+            kind = "para"
+        text = "\n".join(lines[start:i])
+        blocks.append(
+            Block(kind=kind, text=text, tokens=estimate_tokens(text), line_start=start + 1, line_end=i)
+        )
+    return blocks
+
+
+def _read_fence(lines: list, i: int, marker: str) -> int:
+    """Return the index just past a fenced block. Unterminated runs to EOF."""
+    char = marker[0]
+    width = len(marker)
+    i += 1
+    while i < len(lines):
+        candidate = FENCE_RE.match(lines[i])
+        if candidate and candidate.group(1)[0] == char and len(candidate.group(1)) >= width:
+            return i + 1
+        i += 1
+    return len(lines)
+
+
+def _read_list_item(lines: list, i: int) -> int:
+    """Return the index just past one top-level list item, continuations included."""
+    base_indent = len(LIST_RE.match(lines[i]).group(1))
+    i += 1
+    while i < len(lines):
+        raw = lines[i]
+        if not raw.strip():
+            # A blank line continues the item only if an indented line follows.
+            if i + 1 < len(lines) and lines[i + 1].startswith(" " * (base_indent + 1)):
+                i += 1
+                continue
+            return i
+        indent = len(raw) - len(raw.lstrip())
+        if indent > base_indent or LIST_RE.match(raw) and len(LIST_RE.match(raw).group(1)) > base_indent:
+            i += 1
+            continue
+        if LIST_RE.match(raw):
+            return i
+        if indent <= base_indent:
+            return i
+        i += 1
+    return i
+
+
+def _read_paragraph(lines: list, i: int) -> int:
+    i += 1
+    while i < len(lines):
+        raw = lines[i]
+        stripped = raw.strip()
+        if not stripped:
+            return i
+        if FENCE_RE.match(raw) or HEADING_RE.match(raw) or RULE_RE.match(raw):
+            return i
+        if stripped.startswith("|") or stripped.startswith(">") or LIST_RE.match(raw):
+            return i
+        i += 1
+    return i
+
+
+def chunk_blocks(blocks: list) -> list:
+    """Pack blocks into windows of at most TOKEN_MAX, with ~12% overlap."""
+    chunks = []
+    total = len(blocks)
+    if total == 0:
+        return chunks
+    i = 0
+    while i < total:
+        start = i
+        tokens = 0
+        end = i
+        while end < total:
+            block = blocks[end]
+            if end > start:
+                if tokens + block.tokens > TOKEN_MAX:
+                    break
+                # Prefer breaking in front of a heading.
+                if block.kind == "heading" and tokens >= int(TOKEN_TARGET * 0.75):
+                    break
+            tokens += block.tokens
+            end += 1
+            if tokens >= TOKEN_TARGET and end < total and blocks[end].kind == "heading":
+                break
+        window = blocks[start:end]
+        chunks.append(_make_chunk(len(chunks), window))
+        if end >= total:
+            break
+        # Back up over trailing blocks until roughly OVERLAP_RATIO of the window
+        # repeats, leaving at least one block behind so the walk always advances.
+        budget = max(1, int(round(TOKEN_TARGET * OVERLAP_RATIO)))
+        carried = 0
+        back = end
+        while back - 1 > start and carried + blocks[back - 1].tokens <= budget * 2:
+            carried += blocks[back - 1].tokens
+            back -= 1
+            if carried >= budget:
+                break
+        i = max(back, start + 1)
+    return chunks
+
+
+def _make_chunk(index: int, window: list) -> Chunk:
+    text = "\n\n".join(block.text for block in window)
+    return Chunk(
+        index=index,
+        text=text,
+        tokens=sum(block.tokens for block in window),
+        line_start=window[0].line_start,
+        line_end=window[-1].line_end,
+        blocks=tuple(window),
+        oversize=len(window) == 1 and window[0].tokens > TOKEN_MAX,
+    )
+
+
+def chunk_document(body: str) -> list:
+    """Chunk one document body. The pipeline calls this and nothing else."""
+    chunks = chunk_blocks(parse_blocks(body))
+    verify_no_split(chunks, protected_regions(body))
+    return chunks
+
+
+def protected_regions(body: str) -> list:
+    """Line ranges (1-based, inclusive) that a chunk boundary must not cross.
+
+    Recomputed from the body text alone, independently of the packing code, so
+    the check is not the implementation grading its own homework.
+    """
+    regions = []
+    lines = body.split("\n")
+    i = 0
+    while i < len(lines):
+        raw = lines[i]
+        fence = FENCE_RE.match(raw)
+        if fence:
+            end = _read_fence(lines, i, fence.group(1))
+            regions.append((i + 1, end))
+            i = end
+            continue
+        if raw.strip().startswith("|"):
+            start = i
+            while i < len(lines) and lines[i].strip().startswith("|"):
+                i += 1
+            regions.append((start + 1, i))
+            continue
+        i += 1
+    return regions
+
+
+def verify_no_split(chunks: list, regions: list) -> None:
+    """Raise ChunkError if any chunk boundary cuts a protected region."""
+    for chunk in chunks:
+        for first, last in regions:
+            spans_start = chunk.line_start <= first <= chunk.line_end
+            spans_end = chunk.line_start <= last <= chunk.line_end
+            covers = chunk.line_start <= first and chunk.line_end >= last
+            if (spans_start or spans_end) and not covers:
+                raise ChunkError(
+                    f"chunk {chunk.index} splits protected lines {first}-{last} "
+                    f"(chunk covers {chunk.line_start}-{chunk.line_end})"
+                )
+
+
+# -- code -----------------------------------------------------------------
+
+
+def _code_depth(line: str, depth: int) -> int:
+    """Approximate bracket depth for one line, ignoring strings and comments.
+
+    This is a heuristic lexer, not a parser: it is here to answer one question —
+    "is this a safe line to break after?" — and it errs towards *not* breaking.
+    A line with an ambiguous quote or an unterminated bracket simply keeps the
+    chunk open, which can only produce a larger chunk, never a broken one.
+    """
+    index = 0
+    quote = None
+    length = len(line)
+    while index < length:
+        char = line[index]
+        if quote:
+            if char == "\\":
+                index += 2
+                continue
+            if line.startswith(quote, index):
+                index += len(quote)
+                quote = None
+                continue
+            index += 1
+            continue
+        if line.startswith("//", index) or line.startswith("#", index) and line.lstrip().startswith("#"):
+            break
+        if char in "\"'`":
+            triple = char * 3
+            if line.startswith(triple, index):
+                quote = triple
+                index += 3
+                continue
+            quote = char
+            index += 1
+            continue
+        if char in "([{":
+            depth += 1
+        elif char in ")]}":
+            depth = max(0, depth - 1)
+        index += 1
+    return depth
+
+
+def code_segments(text: str) -> list:
+    """Split code into whole top-level segments. A function is never cut."""
+    lines = text.split("\n")
+    segments = []
+    current = []
+    depth = 0
+    start = 1
+    for number, line in enumerate(lines, start=1):
+        if not current:
+            start = number
+        depth = _code_depth(line, depth)
+        current.append(line)
+        # A blank line at bracket depth zero is the only automatic boundary.
+        if depth == 0 and not line.strip() and len(current) > 1:
+            segments.append((start, number, "\n".join(current)))
+            current = []
+    if current:
+        segments.append((start, len(lines), "\n".join(current)))
+    return [(start, end, body) for start, end, body in segments if body.strip()]
+
+
+def chunk_code(text: str) -> list:
+    """Chunk a source file without cutting a function, class or block.
+
+    Segments are whole top-level units. A unit larger than the ceiling is kept
+    as one oversize chunk and flagged, exactly like an oversized Markdown fence:
+    cutting it would break the rule this function exists to keep.
+
+    Overlap repeats the previous unit, never part of one, so an overlapping
+    chunk is still syntactically whole.
+    """
+    segments = code_segments(text)
+    if not segments:
+        raise ChunkError("source file produced no segments (empty?)")
+
+    chunks = []
+    current = []  # list of (start, end, body)
+    tokens = 0
+
+    def flush():
+        nonlocal current, tokens
+        if not current:
+            return
+        body = "\n".join(segment[2] for segment in current).strip("\n")
+        chunks.append(
+            Chunk(
+                index=len(chunks),
+                text=body,
+                tokens=estimate_tokens(body),
+                line_start=current[0][0],
+                line_end=current[-1][1],
+                blocks=tuple(
+                    Block("code", segment[2], estimate_tokens(segment[2]), segment[0], segment[1])
+                    for segment in current
+                ),
+                oversize=len(current) == 1 and estimate_tokens(current[0][2]) > TOKEN_MAX,
+            )
+        )
+        current = []
+        tokens = 0
+
+    for segment in segments:
+        segment_tokens = estimate_tokens(segment[2])
+        if current and tokens + segment_tokens > TOKEN_MAX:
+            previous = current[-1]
+            flush()
+            budget = int(round(TOKEN_TARGET * OVERLAP_RATIO))
+            if estimate_tokens(previous[2]) <= budget:
+                current = [previous]
+                tokens = estimate_tokens(previous[2])
+        current.append(segment)
+        tokens += segment_tokens
+    flush()
+    return chunks

corpus/cli.py

@@ -0,0 +1,377 @@
+"""Command line entry point: `python3 -m corpus.cli <command>`.
+
+Commands
+
+    sync          incremental update of the corpus (the normal command)
+    rebuild       drop the collection and rebuild it from content/ (recovery)
+    audit         freshness audit; reports stale documents, index drift
+    verify        does the index agree with content/ and with Qdrant?
+    demo-filter   the same query with and without an `area` filter
+    hybrid        vectors + FTS5 fused with reciprocal rank fusion
+    measure       Qdrant memory at 500 / 5,000 / 20,000 chunks
+    info          what this build thinks it is configured to do
+
+Exit codes: 0 success, 1 findings or verification problems, 2 configuration or
+usage error. Every command accepts `--json` where a machine-readable result is
+useful (Prefect reads JSON, a human reads the table).
+
+Configuration is environment-first (`corpus/config.py`); the flags exist so a
+one-off run can point at a scratch Qdrant without exporting variables.
+"""
+
+from __future__ import annotations
+
+import argparse
+import datetime
+import json
+import os
+import sys
+
+from . import config as config_module
+from . import freshness, guides, pipeline, store as store_module
+from .config import Config, ConfigError
+from .qdrant import Qdrant, QdrantError
+from .sources import graph as graph_module
+
+
+def main(argv=None) -> int:
+    parser = _build_parser()
+    args = parser.parse_args(argv)
+    if args.command is None:
+        parser.print_help()
+        return 2
+
+    overrides = {
+        "root": args.root,
+        "qdrant_url": args.qdrant_url,
+        "collection": args.collection,
+        "provider": args.provider,
+        "model": args.model,
+        "dim": args.dim,
+        "db_path": args.db,
+        "data_dir": args.data,
+        "offline": True if args.offline else None,
+        "platforms": tuple(part.strip() for part in args.platforms.split(",") if part.strip())
+        if args.platforms
+        else None,
+        "graph_path": args.graph_path or None,
+    }
+    try:
+        cfg = Config.from_env(**overrides)
+    except ConfigError as exc:
+        print(f"configuration error: {exc}", file=sys.stderr)
+        return 2
+
+    log = (lambda _msg: None) if args.quiet else (lambda msg: print(msg, file=sys.stderr))
+
+    try:
+        handler = {
+            "sync": _cmd_sync,
+            "rebuild": _cmd_rebuild,
+            "audit": _cmd_audit,
+            "verify": _cmd_verify,
+            "demo-filter": _cmd_demo_filter,
+            "hybrid": _cmd_hybrid,
+            "measure": _cmd_measure,
+            "info": _cmd_info,
+            "graph": _cmd_graph,
+            "graph-neighbours": _cmd_graph_neighbours,
+            "scan": _cmd_scan,
+            "sources": _cmd_sources,
+        }[args.command]
+        return handler(cfg, args, log)
+    except (pipeline.PipelineError, ConfigError) as exc:
+        print(f"error: {exc}", file=sys.stderr)
+        return 2
+    except QdrantError as exc:
+        print(f"error: {exc}", file=sys.stderr)
+        return 2
+
+
+def _build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(
+        prog="python3 -m corpus.cli",
+        description="Build and audit the Ai Resume retrieval corpus (docs/phase-4-corpus.md).",
+    )
+    parser.add_argument("--root", help="repository root (default: parent of corpus/)")
+    parser.add_argument("--qdrant-url", help="Qdrant REST URL (default http://127.0.0.1:6333)")
+    parser.add_argument("--collection", help="Qdrant collection name (default resume_corpus)")
+    parser.add_argument("--provider", choices=sorted(config_module.PROVIDERS), help="embedding provider")
+    parser.add_argument("--model", help="embedding model name")
+    parser.add_argument("--dim", type=int, help="embedding dimension")
+    parser.add_argument("--db", help="SQLite corpus database path")
+    parser.add_argument("--data", help="corpus data directory")
+    parser.add_argument("--quiet", action="store_true", help="progress on stderr only becomes silent")
+    parser.add_argument("--offline", action="store_true", help="read content/ only; do not harvest the estate")
+    parser.add_argument("--platforms", help="comma-separated platforms to harvest (default: all enabled)")
+    parser.add_argument("--graph-path", help="read graph.json from this path instead of Gitea")
+
+    sub = parser.add_subparsers(dest="command")
+
+    sync = sub.add_parser("sync", help="incremental update")
+    sync.add_argument("--json", action="store_true")
+
+    rebuild = sub.add_parser("rebuild", help="drop the collection and rebuild from content/")
+    rebuild.add_argument("--json", action="store_true")
+
+    audit = sub.add_parser("audit", help="freshness and drift audit")
+    audit.add_argument("--json", action="store_true")
+    audit.add_argument("--today", help="audit as of this date (YYYY-MM-DD), for tests")
+    audit.add_argument("--fail-on-findings", action="store_true", help="exit 1 when anything is found")
+    audit.add_argument("--alert", action="store_true", help="POST the report to RESUME_APPRISE_URL")
+
+    verify = sub.add_parser("verify", help="check the index against content/ and Qdrant")
+    verify.add_argument("--json", action="store_true")
+
+    demo = sub.add_parser("demo-filter", help="same query with and without an area filter")
+    demo.add_argument("--query", required=True)
+    demo.add_argument("--area", help="area to filter on (ai, platform, edge, products, history, about)")
+    demo.add_argument("--limit", type=int, default=5)
+    demo.add_argument("--json", action="store_true")
+
+    hybrid = sub.add_parser("hybrid", help="vectors + FTS5 fused with RRF")
+    hybrid.add_argument("--query", required=True)
+    hybrid.add_argument("--area")
+    hybrid.add_argument("--limit", type=int, default=5)
+    hybrid.add_argument("--json", action="store_true")
+
+    measure = sub.add_parser("measure", help="Qdrant memory at three corpus sizes")
+    measure.add_argument("--sizes", default="500,5000,20000")
+    measure.add_argument("--container", default="", help="Qdrant container name for docker stats")
+    measure.add_argument("--keep", action="store_true", help="keep the measurement collection")
+    measure.add_argument("--json", action="store_true")
+
+    graph = sub.add_parser("graph", help="query graph nodes and edges by keyword or vector")
+    graph.add_argument("--query", required=True)
+    graph.add_argument("--area")
+    graph.add_argument("--limit", type=int, default=10)
+    graph.add_argument("--json", action="store_true")
+
+    neighbours = sub.add_parser("graph-neighbours", help="edges attached to a graph node id")
+    neighbours.add_argument("--node", required=True)
+    neighbours.add_argument("--limit", type=int, default=50)
+    neighbours.add_argument("--json", action="store_true")
+
+    scan = sub.add_parser("scan", help="search the finished pool for credential-shaped content")
+    scan.add_argument("--json", action="store_true")
+
+    sources = sub.add_parser("sources", help="show the platform guides and the last harvest")
+    sources.add_argument("--json", action="store_true")
+
+    sub.add_parser("info", help="print the resolved configuration")
+    return parser
+
+
+# -- commands --------------------------------------------------------------
+
+
+def _cmd_sync(cfg, args, log) -> int:
+    stats = pipeline.run(cfg, mode="sync", log=log)
+    if args.json:
+        print(json.dumps(stats.to_dict(), indent=2, sort_keys=True))
+    else:
+        print(stats.summary())
+    return 0
+
+
+def _cmd_rebuild(cfg, args, log) -> int:
+    stats = pipeline.run(cfg, mode="rebuild", log=log)
+    if args.json:
+        print(json.dumps(stats.to_dict(), indent=2, sort_keys=True))
+    else:
+        print(stats.summary())
+    return 0
+
+
+def _cmd_audit(cfg, args, log) -> int:
+    today = datetime.date.fromisoformat(args.today) if args.today else None
+    report = freshness.audit(cfg, today=today, log=log)
+    if args.json:
+        print(json.dumps(report, indent=2, sort_keys=True))
+    else:
+        print(freshness.render_text(report))
+    if args.alert:
+        try:
+            sent = freshness.alert(report, cfg.apprise_url)
+            if not sent:
+                log("alert: RESUME_APPRISE_URL is not set; nothing sent")
+        except Exception as exc:  # noqa: BLE001 - an alert failure must not hide the audit
+            print(f"warning: could not send the Apprise alert: {exc}", file=sys.stderr)
+    if args.fail_on_findings and not report["ok"]:
+        return 1
+    return 0
+
+
+def _cmd_verify(cfg, args, log) -> int:
+    problems = pipeline.verify(cfg, log=log)
+    if args.json:
+        print(json.dumps({"ok": not problems, "problems": problems}, indent=2))
+    else:
+        if problems:
+            print(f"FAIL: {len(problems)} problem(s)")
+            for problem in problems:
+                print(f"  - {problem}")
+        else:
+            print("OK: the index agrees with content/ and Qdrant")
+    return 1 if problems else 0
+
+
+def _cmd_demo_filter(cfg, args, log) -> int:
+    result = pipeline.demo_area_filter(cfg, args.query, area=args.area, limit=args.limit)
+    if args.json:
+        print(json.dumps(result, indent=2, sort_keys=True))
+        return 0
+    print(f"query: {result['query']!r}")
+    print("")
+    print(f"without a filter ({len(result['unfiltered'])} result(s)):")
+    for row in result["unfiltered"]:
+        print(f"  {row['score']:.3f}  [{row['area']}/{row['confidence']}] {row['title']}  {row['url']}")
+    print("")
+    if not args.area:
+        print("no --area given, so the filtered set is empty; re-run with e.g. --area edge")
+        return 0
+    print(f"with area={args.area!r} ({len(result['filtered'])} result(s)):")
+    for row in result["filtered"]:
+        print(f"  {row['score']:.3f}  [{row['area']}/{row['confidence']}] {row['title']}  {row['url']}")
+    wrong = [row for row in result["filtered"] if row["area"] != args.area]
+    if wrong:
+        print(f"FAIL: {len(wrong)} filtered result(s) are not in area {args.area!r}")
+        return 1
+    return 0
+
+
+def _cmd_hybrid(cfg, args, log) -> int:
+    rows = pipeline.hybrid_search(cfg, args.query, area=args.area, limit=args.limit)
+    if args.json:
+        print(json.dumps(rows, indent=2, sort_keys=True, default=str))
+        return 0
+    print(f"hybrid (RRF) query={args.query!r} area={args.area!r}: {len(rows)} result(s)")
+    for row in rows:
+        print(f"  {row['score']:.4f}  {row.get('title')}  {row.get('url')}  [{row['id'][:8]}]")
+    return 0
+
+
+def _cmd_measure(cfg, args, log) -> int:
+    from . import measure as measure_module
+
+    sizes = [int(part) for part in args.sizes.split(",") if part.strip()]
+    report = measure_module.measure(
+        cfg, sizes=sizes, container=args.container, keep=args.keep, log=log
+    )
+    if args.json:
+        print(json.dumps(report, indent=2, sort_keys=True))
+        return 0
+    print(measure_module.render_text(report))
+    return 0
+
+
+def _cmd_graph(cfg, args, log) -> int:
+    with store_module.Store(cfg.db_path) as store:
+        report = graph_module.query(cfg, store, args.query, area=args.area, limit=args.limit)
+    if args.json:
+        print(json.dumps(report, indent=2, sort_keys=True, default=str))
+        return 0
+    print(f"graph query {args.query!r} area={args.area!r}")
+    print("")
+    print(f"keyword ({len(report['keyword'])} row(s)):")
+    for row in report["keyword"]:
+        kind = row.get("kind")
+        print(f"  [{kind}] {row.get('label')}  ->  {row.get('source_note')}  {row.get('origin_url')}")
+    print("")
+    print(f"vector ({len(report['vector'])} row(s)):")
+    for row in report["vector"]:
+        print(f"  {row['score']:.3f}  [{row['kind']}] {row['label']}  ->  {row['source_note']}")
+    if report.get("vector_note"):
+        print(f"  note: {report['vector_note']}")
+    return 0 if report["ok"] else 1
+
+
+def _cmd_graph_neighbours(cfg, args, log) -> int:
+    with store_module.Store(cfg.db_path) as store:
+        node = store.graph_node(args.node)
+        edges = store.graph_neighbours(args.node, limit=args.limit)
+    if args.json:
+        print(json.dumps({"node": node, "edges": edges}, indent=2, sort_keys=True, default=str))
+        return 0
+    if not node:
+        print(f"no graph node with id {args.node!r}")
+        return 1
+    print(f"{node.get('label')}  ({node.get('node_type')})")
+    print(f"  note: {node.get('source_note')}")
+    print(f"  origin: {node.get('origin_url')}")
+    for edge in edges:
+        print(
+            f"  {edge.get('source_label')} --{edge.get('relation')}--> {edge.get('target_label')}"
+            f"  [{edge.get('source_note')}]"
+        )
+    return 0
+
+
+def _cmd_scan(cfg, args, log) -> int:
+    report = pipeline.scan_pool(cfg)
+    if args.json:
+        print(json.dumps(report, indent=2, sort_keys=True))
+    else:
+        print(
+            f"scanned {report['chunks']} chunk(s): "
+            + ("clean" if report["ok"] else f"{len(report['hits'])} hit(s)")
+        )
+        for hit in report["hits"]:
+            print(f"  {hit}")
+    return 0 if report["ok"] else 1
+
+
+def _cmd_sources(cfg, args, log) -> int:
+    loaded = guides.load_guides(cfg.root)
+    public = guides.build_public_manifest(loaded)
+    last = {}
+    if os.path.exists(cfg.db_path):
+        with store_module.Store(cfg.db_path) as store:
+            last = store.get_meta("sources") or {}
+    if args.json:
+        print(json.dumps({"guides": public, "last_harvest": last}, indent=2, sort_keys=True))
+        return 0
+    for row in public["platforms"]:
+        state = "enabled" if row["enabled"] else "disabled"
+        print(f"{row['name']} ({state}) — {row['what']}")
+        print(f"  base: {row['base']}  fits: {', '.join(row['fits'])}  read: {', '.join(row['read'])}")
+        print(f"  include: {json.dumps(row['include'])}")
+        print(f"  exclude: {json.dumps(row['exclude'])}")
+        seen = last.get(row["name"]) or {}
+        if seen:
+            print(
+                f"  last harvest: {seen.get('documents', 0)} document(s), "
+                f"newest {seen.get('newest') or '—'}"
+            )
+        print("")
+    return 0
+
+
+def _cmd_info(cfg, args, log) -> int:
+    print(cfg.describe())
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    print("")
+    print(f"qdrant: {'answering, version ' + qdrant.version() if qdrant.health() else 'NOT answering'}")
+    if qdrant.collection_exists(cfg.collection):
+        print(f"collection {cfg.collection}: {qdrant.point_count(cfg.collection)} point(s)")
+    else:
+        print(f"collection {cfg.collection}: does not exist")
+    if os.path.exists(cfg.db_path):
+        with store_module.Store(cfg.db_path) as store:
+            print(f"store: {store.chunk_count()} chunk(s) across {store.source_count()} source(s)")
+            print(f"contract: {json.dumps(store.get_meta('collection'), sort_keys=True)}")
+            print(f"graph: {json.dumps(store.graph_counts(), sort_keys=True)}")
+            sources = store.get_meta("sources") or {}
+            for name in sorted(sources):
+                row = sources[name]
+                print(
+                    f"source {name}: {row.get('documents', 0)} document(s), "
+                    f"newest {row.get('newest') or '—'}"
+                )
+    else:
+        print(f"store: {cfg.db_path} does not exist")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

corpus/config.py

@@ -0,0 +1,206 @@
+"""Configuration for the corpus pipeline. Environment only.
+
+Nothing in this file holds a credential. The OmniRoute key is read from the
+process environment, which systemd populates from
+`~/.config/environment.d/10-secrets.conf` (AGENTS.md section 4.4). `describe()`
+is the only way the pipeline prints its configuration, and it redacts the key.
+"""
+
+from __future__ import annotations
+
+import os
+from dataclasses import dataclass
+
+# Bumped whenever chunk boundaries or payload shape change. A stored corpus
+# carrying a different version is stale by definition and gets rebuilt.
+CHUNKER_VERSION = "blocks-v2-code-v1"
+
+# Provider defaults. `local-hashing` is a deterministic, offline provider used
+# by the tests, by `make corpus-test`, and by any rebuild on a host with no
+# model gateway. It is NOT a semantic embedding model: it exists so the whole
+# pipeline can be exercised without a provider decision, and so a rebuild never
+# depends on the network. The production provider is `omniroute`.
+PROVIDERS = {
+    "local-hashing": ("local-hashing-v1", 256),
+    "omniroute": ("text-embedding-3-small", 1536),
+}
+
+DEFAULT_QDRANT_URL = "http://127.0.0.1:6333"
+DEFAULT_COLLECTION = "resume_corpus"
+DEFAULT_PROVIDER = "local-hashing"
+DEFAULT_OMNIROUTE_URL = "https://omniroute.lab.audasmedia.com.au"
+DEFAULT_GITEA_URL = "https://gitea.lab.audasmedia.com.au"
+DEFAULT_PLATFORMS = ("gitea", "obsidian", "graphify")
+DEFAULT_GRAPH_COLLECTION = "resume_graph"
+
+# A source that has not changed in this window is reported by the audit. The
+# estate is active; silence is evidence that ingestion stopped, not that the
+# work stopped.
+DEFAULT_SOURCE_SILENCE_DAYS = 30
+
+# Freshness horizon: a document whose `last_verified` is older than this is
+# reported stale by the nightly audit. docs/phase-4-notes.md section 5.
+DEFAULT_MAX_AGE_DAYS = 180
+DEFAULT_WARN_DAYS = 30
+
+
+class ConfigError(RuntimeError):
+    """A configuration value is missing or unusable."""
+
+
+@dataclass(frozen=True)
+class Config:
+    """Everything the pipeline needs, resolved once."""
+
+    root: str
+    content_dir: str
+    data_dir: str
+    db_path: str
+    qdrant_url: str
+    collection: str
+    provider: str
+    model: str
+    dim: int
+    api_url: str
+    api_key: str
+    batch_size: int
+    timeout: float
+    max_age_days: int
+    warn_days: int
+    apprise_url: str
+    gitea_url: str = DEFAULT_GITEA_URL
+    gitea_token: str = ""
+    platforms: tuple = DEFAULT_PLATFORMS
+    offline: bool = False
+    graph_path: str = ""
+    graph_collection: str = DEFAULT_GRAPH_COLLECTION
+    source_silence_days: int = DEFAULT_SOURCE_SILENCE_DAYS
+    vectors_on_disk: bool = False
+
+    # -- construction -----------------------------------------------------
+
+    @classmethod
+    def from_env(cls, root: str | None = None, **overrides) -> "Config":
+        """Build a Config from the environment, then apply explicit overrides.
+
+        Overrides exist so the CLI can accept flags without a second code path;
+        they are applied last so a flag always wins over the environment.
+        """
+        root = os.path.abspath(root or os.environ.get("RESUME_ROOT") or _default_root())
+        content_dir = os.environ.get("RESUME_CONTENT") or os.path.join(root, "content")
+        data_dir = os.environ.get("RESUME_CORPUS_DATA") or os.path.join(root, "data", "corpus")
+        provider = os.environ.get("RESUME_EMBED_PROVIDER") or DEFAULT_PROVIDER
+        if provider not in PROVIDERS:
+            raise ConfigError(
+                f"RESUME_EMBED_PROVIDER={provider!r} is not one of {sorted(PROVIDERS)}"
+            )
+        default_model, default_dim = PROVIDERS[provider]
+
+        values = dict(
+            root=root,
+            content_dir=os.path.abspath(content_dir),
+            data_dir=os.path.abspath(data_dir),
+            db_path=os.path.abspath(
+                os.environ.get("RESUME_CORPUS_DB") or os.path.join(data_dir, "corpus.db")
+            ),
+            qdrant_url=(os.environ.get("RESUME_QDRANT_URL") or DEFAULT_QDRANT_URL).rstrip("/"),
+            collection=os.environ.get("RESUME_COLLECTION") or DEFAULT_COLLECTION,
+            provider=provider,
+            model=os.environ.get("RESUME_EMBED_MODEL") or default_model,
+            dim=int(os.environ.get("RESUME_EMBED_DIM") or default_dim),
+            api_url=(os.environ.get("OMNIROUTE_URL") or DEFAULT_OMNIROUTE_URL).rstrip("/"),
+            api_key=os.environ.get("OMNIROUTE_API_KEY") or "",
+            batch_size=int(os.environ.get("RESUME_EMBED_BATCH") or 32),
+            timeout=float(os.environ.get("RESUME_QDRANT_TIMEOUT") or 30.0),
+            max_age_days=int(
+                os.environ.get("RESUME_MAX_AGE_DAYS") or DEFAULT_MAX_AGE_DAYS
+            ),
+            warn_days=int(os.environ.get("RESUME_WARN_DAYS") or DEFAULT_WARN_DAYS),
+            apprise_url=os.environ.get("RESUME_APPRISE_URL") or "",
+            gitea_url=(os.environ.get("GITEA_URL") or os.environ.get("RESUME_GITEA_URL") or DEFAULT_GITEA_URL).rstrip("/"),
+            gitea_token=os.environ.get("GITEA_TOKEN") or "",
+            platforms=tuple(
+                part.strip()
+                for part in (os.environ.get("RESUME_PLATFORMS") or ",".join(DEFAULT_PLATFORMS)).split(",")
+                if part.strip()
+            ),
+            offline=(os.environ.get("RESUME_OFFLINE") or "").lower() in ("1", "true", "yes"),
+            graph_path=os.environ.get("RESUME_GRAPH_PATH") or "",
+            graph_collection=os.environ.get("RESUME_GRAPH_COLLECTION") or DEFAULT_GRAPH_COLLECTION,
+            source_silence_days=int(
+                os.environ.get("RESUME_SOURCE_SILENCE_DAYS") or DEFAULT_SOURCE_SILENCE_DAYS
+            ),
+            vectors_on_disk=(os.environ.get("RESUME_QDRANT_VECTORS_ON_DISK") or "").lower()
+            in ("1", "true", "yes"),
+        )
+        values.update({k: v for k, v in overrides.items() if v not in (None, "")})
+        cfg = cls(**values)
+        cfg.check()
+        return cfg
+
+    # -- validation -------------------------------------------------------
+
+    def check(self) -> None:
+        if self.dim <= 0:
+            raise ConfigError(f"embedding dimension must be positive, got {self.dim}")
+        if self.batch_size <= 0:
+            raise ConfigError(f"batch size must be positive, got {self.batch_size}")
+        if self.provider == "omniroute" and not self.api_key:
+            raise ConfigError(
+                "provider 'omniroute' needs OMNIROUTE_API_KEY in the environment "
+                "(~/.config/environment.d/10-secrets.conf). It is never read from this repo."
+            )
+        if not os.path.isdir(self.content_dir):
+            raise ConfigError(f"content directory not found: {self.content_dir}")
+        unknown = [name for name in self.platforms if name not in ("gitea", "obsidian", "graphify")]
+        if unknown:
+            raise ConfigError(
+                f"RESUME_PLATFORMS contains unknown platform(s) {unknown}; "
+                "expected gitea, obsidian, graphify"
+            )
+        if not self.offline and self.gitea_token == "":
+            # Only a hard error when a platform that needs Gitea is selected; the
+            # token itself is never printed or stored.
+            if set(self.platforms) & {"gitea", "obsidian", "graphify"}:
+                raise ConfigError(
+                    "GITEA_TOKEN is not set. Harvesting needs read access to Gitea "
+                    "(~/.config/environment.d/10-secrets.conf), or set RESUME_OFFLINE=1 "
+                    "to rebuild only the curated tier."
+                )
+
+    # -- presentation -----------------------------------------------------
+
+    def describe(self) -> str:
+        """Human-readable configuration. Never includes the API key value."""
+        key = "set" if self.api_key else "unset"
+        return (
+            f"root={self.root}\n"
+            f"content={self.content_dir}\n"
+            f"data={self.data_dir}\n"
+            f"db={self.db_path}\n"
+            f"qdrant={self.qdrant_url} collection={self.collection}\n"
+            f"provider={self.provider} model={self.model} dim={self.dim} "
+            f"batch={self.batch_size}\n"
+            f"api_url={self.api_url} api_key={key}\n"
+            f"freshness=max_age_days:{self.max_age_days} warn_days:{self.warn_days}\n"
+            f"estate=platforms:{','.join(self.platforms)} offline:{self.offline} "
+            f"gitea:{self.gitea_url} token:{'set' if self.gitea_token else 'unset'} "
+            f"graph_path:{self.graph_path or 'from gitea'} silence_days:{self.source_silence_days}\n"
+            f"qdrant vectors_on_disk={self.vectors_on_disk} graph_collection={self.graph_collection}\n"
+            f"chunker={CHUNKER_VERSION}"
+        )
+
+    # -- derived paths ----------------------------------------------------
+
+    @property
+    def chunker_version(self) -> str:
+        return CHUNKER_VERSION
+
+    @property
+    def meta_path(self) -> str:
+        return os.path.join(self.data_dir, "collection.json")
+
+
+def _default_root() -> str:
+    """The repository root: the parent of this package."""
+    return os.path.dirname(os.path.dirname(os.path.abspath(__file__)))

corpus/embeddings.py

@@ -0,0 +1,241 @@
+"""The one place a vector is made.
+
+`docs/phase-4-corpus.md` §4.3: *"Record the embedding model and dimension per
+collection. A provider change is a governed migration, not a silent corruption.
+Provider decision pending Sam — build provider-agnostic behind one interface."*
+
+So: one interface (`Provider`), two implementations, and the provider name,
+model and dimension are written to the collection metadata by the pipeline. A
+later run with a different model refuses to mix vectors into the same
+collection; it reports a governed migration instead.
+
+**local-hashing** is deterministic, offline, and *not semantic*. It hashes word
+tokens into a fixed-width vector. It exists so that the pipeline, its tests and
+a from-scratch rebuild never depend on a network or on an unmade decision. It is
+the default for `make corpus-test` and for the Qdrant memory measurement.
+
+**omniroute** is the production path: an OpenAI-compatible `/v1/embeddings`
+call through OmniRoute, which is the only permitted model gateway
+(AGENTS.md §4.5). No provider SDK, no direct vendor call, key from the
+environment only.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import math
+import time
+import urllib.error
+import urllib.request
+
+
+class EmbeddingError(RuntimeError):
+    """Embedding failed. The pipeline aborts rather than writing partial vectors."""
+
+
+class Provider:
+    """Interface every embedding provider implements."""
+
+    name = "provider"
+    model = "model"
+    dim = 0
+
+    def embed(self, texts: list) -> list:
+        raise NotImplementedError
+
+    def stats(self) -> dict:
+        """Call counters. The DoD proves incremental mode with these numbers."""
+        return {"calls": 0, "texts": 0}
+
+    def reset_stats(self) -> None:
+        pass
+
+    def describe(self) -> dict:
+        return {"provider": self.name, "model": self.model, "dim": self.dim}
+
+
+class LocalHashingProvider(Provider):
+    """Deterministic feature hashing. Offline, fast, not semantic.
+
+    Every token is hashed with sha256 (never Python's salted `hash()`), so the
+    same text produces the same vector on every host and every run. Vectors are
+    L2-normalised, which suits cosine distance.
+    """
+
+    name = "local-hashing"
+
+    def __init__(self, model: str = "local-hashing-v1", dim: int = 256) -> None:
+        if dim <= 0:
+            raise EmbeddingError(f"local-hashing needs a positive dimension, got {dim}")
+        self.model = model
+        self.dim = dim
+        self._calls = 0
+        self._texts = 0
+
+    def embed(self, texts: list) -> list:
+        self._calls += 1
+        self._texts += len(texts)
+        return [self._vector(text) for text in texts]
+
+    def _vector(self, text: str) -> list:
+        vector = [0.0] * self.dim
+        for token in _tokens(text):
+            digest = hashlib.sha256(token.encode("utf-8")).digest()
+            index = int.from_bytes(digest[:8], "big") % self.dim
+            sign = 1.0 if digest[8] & 1 else -1.0
+            vector[index] += sign
+        norm = math.sqrt(sum(value * value for value in vector))
+        if norm == 0.0:
+            # An empty text still needs a unit vector of the right width.
+            vector[0] = 1.0
+            return vector
+        return [value / norm for value in vector]
+
+    def stats(self) -> dict:
+        return {"calls": self._calls, "texts": self._texts}
+
+    def reset_stats(self) -> None:
+        self._calls = 0
+        self._texts = 0
+
+
+class OmniRouteProvider(Provider):
+    """OpenAI-compatible `/v1/embeddings` through OmniRoute.
+
+    Retries on 429 and 5xx with exponential backoff (the risk table in
+    `docs/phase-4-corpus.md` calls this out for bulk ingest). The key is never
+    logged, never put in an error message, and never read from this repository.
+    """
+
+    name = "omniroute"
+
+    def __init__(
+        self,
+        api_url: str,
+        api_key: str,
+        model: str,
+        dim: int,
+        batch_size: int = 32,
+        timeout: float = 30.0,
+        attempts: int = 4,
+        backoff: float = 1.5,
+    ) -> None:
+        if not api_key:
+            raise EmbeddingError(
+                "OmniRoute needs OMNIROUTE_API_KEY in the environment "
+                "(~/.config/environment.d/10-secrets.conf)"
+            )
+        self.api_url = api_url.rstrip("/")
+        self._api_key = api_key
+        self.model = model
+        self.dim = dim
+        self.batch_size = max(1, batch_size)
+        self.timeout = timeout
+        self.attempts = max(1, attempts)
+        self.backoff = backoff
+        self._calls = 0
+        self._texts = 0
+
+    def embed(self, texts: list) -> list:
+        out = []
+        for start in range(0, len(texts), self.batch_size):
+            batch = texts[start : start + self.batch_size]
+            out.extend(self._embed_batch(batch))
+        return out
+
+    def _embed_batch(self, batch: list) -> list:
+        body = json.dumps({"model": self.model, "input": batch}).encode("utf-8")
+        last_error = None
+        for attempt in range(self.attempts):
+            request = urllib.request.Request(
+                f"{self.api_url}/v1/embeddings",
+                data=body,
+                method="POST",
+                headers={
+                    "Content-Type": "application/json",
+                    "Authorization": f"Bearer {self._api_key}",
+                },
+            )
+            self._calls += 1
+            try:
+                with urllib.request.urlopen(request, timeout=self.timeout) as response:
+                    payload = json.loads(response.read().decode("utf-8"))
+                vectors = _vectors_from_response(payload, self.model, self.dim)
+                self._texts += len(batch)
+                return vectors
+            except urllib.error.HTTPError as exc:
+                last_error = f"HTTP {exc.code} from {self.api_url}/v1/embeddings"
+                exc.close()
+                # 4xx other than 429 is our fault: do not hammer the gateway.
+                if exc.code != 429 and exc.code < 500:
+                    raise EmbeddingError(f"{last_error} (not retryable)") from None
+            except (urllib.error.URLError, TimeoutError) as exc:
+                last_error = f"cannot reach {self.api_url}: {exc.reason}"
+            except EmbeddingError:
+                raise
+            if attempt + 1 < self.attempts:
+                time.sleep(self.backoff ** attempt)
+        raise EmbeddingError(f"{last_error} after {self.attempts} attempts")
+
+    def stats(self) -> dict:
+        return {"calls": self._calls, "texts": self._texts}
+
+    def reset_stats(self) -> None:
+        self._calls = 0
+        self._texts = 0
+
+
+def _vectors_from_response(payload: dict, model: str, dim: int) -> list:
+    data = payload.get("data")
+    if not isinstance(data, list) or not data:
+        raise EmbeddingError("embedding response has no 'data' array")
+    ordered = sorted(data, key=lambda item: item.get("index", 0))
+    vectors = []
+    for item in ordered:
+        vector = item.get("embedding")
+        if not isinstance(vector, list) or not vector:
+            raise EmbeddingError("embedding response item has no 'embedding' array")
+        if dim and len(vector) != dim:
+            raise EmbeddingError(
+                f"provider returned dimension {len(vector)} but the collection is "
+                f"configured for {dim}; refusing to mix dimensions"
+            )
+        vectors.append([float(value) for value in vector])
+    returned_model = payload.get("model")
+    if returned_model and model and returned_model != model:
+        raise EmbeddingError(
+            f"provider answered as model {returned_model!r}, not {model!r}; "
+            "a model change must be a governed migration"
+        )
+    return vectors
+
+
+def build_provider(cfg) -> Provider:
+    """Construct the provider named by the configuration."""
+    if cfg.provider == "local-hashing":
+        return LocalHashingProvider(model=cfg.model, dim=cfg.dim)
+    if cfg.provider == "omniroute":
+        return OmniRouteProvider(
+            api_url=cfg.api_url,
+            api_key=cfg.api_key,
+            model=cfg.model,
+            dim=cfg.dim,
+            batch_size=cfg.batch_size,
+            timeout=cfg.timeout,
+        )
+    raise EmbeddingError(f"unknown embedding provider {cfg.provider!r}")
+
+
+def _tokens(text: str) -> list:
+    out = []
+    current = []
+    for char in text.lower():
+        if char.isalnum():
+            current.append(char)
+        elif current:
+            out.append("".join(current))
+            current = []
+    if current:
+        out.append("".join(current))
+    return out

corpus/exclusions.py

@@ -0,0 +1,256 @@
+"""What must never reach the pool.
+
+AGENTS.md §4.4 and `docs/phase-4-corpus.md` §6 make this a hard gate, not a
+suggestion: node_modules, virtualenvs, git internals, `.trash`, `.obsidian`, the
+SilverBullet auth file, credential-shaped files and binaries are refused. The
+platform guide excludes are *additional*; they cannot widen this list.
+
+Two different outcomes, because they are different things:
+
+* **reject** — the file is not read at all. The reason is recorded and reported
+  so a run says *why* it skipped something.
+* **image** — a media file is recorded as an image node reference (path only,
+  never bytes, never extracted text). Phase 4 §1.
+
+Credential-shaped *content* is a third thing: a file with a valid path can still
+contain a private key or an API token. `scan_text` rejects the whole document
+rather than redacting part of it, and the same scan runs over the finished pool
+so "no secret in the pool" is verified by search, not asserted.
+"""
+
+from __future__ import annotations
+
+import re
+
+from . import globs
+
+# Directory names that are never descended into. Checked per path component, so
+# a file deep inside one is refused by the same rule as its parent.
+EXCLUDED_DIRS = frozenset(
+    {
+        "node_modules",
+        ".venv",
+        "venv",
+        ".git",
+        ".trash",
+        ".obsidian",
+        ".pi",
+        "__pycache__",
+        ".mypy_cache",
+        ".pytest_cache",
+        ".ruff_cache",
+        "graphify-out",
+        "graphify-out-temp",
+        ".next",
+        ".cache",
+        "site-packages",
+    }
+)
+
+# Basenames (and simple suffix patterns) that are credential-shaped by name.
+CREDENTIAL_FILES = (
+    ".silverbullet.auth.json",
+    ".env",
+    ".env.*",
+    "*.env",
+    "*.env.*",
+    ".netrc",
+    ".git-credentials",
+    ".npmrc",
+    ".pypirc",
+    "credentials",
+    "credentials.*",
+    "secrets.*",
+    "secret.*",
+    "auth.json",
+    "*.pem",
+    "*.key",
+    "*.p12",
+    "*.pfx",
+    "*.jks",
+    "*.keystore",
+    "*.kdbx",
+    "*.age",
+    "id_rsa",
+    "id_ed25519",
+    "id_ecdsa",
+    "*.lock",
+    "*.lockb",
+)
+
+# Media is never text. It becomes an image node reference and nothing else.
+IMAGE_EXTENSIONS = frozenset(
+    {"png", "jpg", "jpeg", "gif", "webp", "bmp", "tif", "tiff", "heic", "avif", "ico"}
+)
+VIDEO_EXTENSIONS = frozenset({"mp4", "mov", "mkv", "webm", "avi", "m4v", "mpg", "mpeg"})
+AUDIO_EXTENSIONS = frozenset({"mp3", "wav", "flac", "ogg", "m4a", "aac", "opus", "wma"})
+MEDIA_EXTENSIONS = IMAGE_EXTENSIONS | VIDEO_EXTENSIONS | AUDIO_EXTENSIONS
+
+# Everything else that is bytes, not prose. Rejected with a reason; not indexed
+# and not recorded as media.
+BINARY_EXTENSIONS = frozenset(
+    {
+        "pdf", "zip", "tar", "gz", "bz2", "xz", "7z", "rar", "iso", "dmg",
+        "exe", "dll", "so", "dylib", "o", "a", "class", "jar", "war",
+        "pyc", "pyo", "wasm", "ttf", "otf", "woff", "woff2", "eot",
+        "sqlite", "sqlite3", "db", "mdb", "bin", "dat", "pkl", "npy", "npz",
+        "stl", "step", "obj", "blend", "psd", "ai", "sketch", "fig",
+    }
+)
+
+# Text-ish extensions the estate actually contains. A file with an unknown
+# extension is admitted only when it decodes as UTF-8 and does not look binary.
+TEXT_EXTENSIONS = frozenset(
+    {
+        "md", "markdown", "mdx", "txt", "rst", "adoc", "org",
+        "py", "go", "ts", "tsx", "js", "jsx", "mjs", "cjs", "rs", "rb",
+        "php", "kt", "kts", "java", "cs", "c", "h", "cc", "cpp", "hpp",
+        "ino", "lua", "pl", "r", "scala", "swift", "dart", "ex", "exs",
+        "sh", "bash", "zsh", "fish", "ps1", "bat",
+        "nix", "toml", "yml", "yaml", "ini", "conf", "cfg", "properties",
+        "json", "jsonc", "xml", "html", "htm", "css", "scss", "sass", "less",
+        "sql", "graphql", "proto", "tf", "hcl", "dockerfile", "makefile",
+        "gitignore", "dockerignore", "editorconfig", "service", "timer",
+        "csv", "tsv", "svg", "tex", "patch", "diff", "env.example",
+        "example", "sample", "template", "j2", "tmpl",
+    }
+)
+
+_CREDENTIAL_PATTERNS = (
+    ("private-key", re.compile(r"-----BEGIN (?:[A-Z]+ )?PRIVATE KEY-----")),
+    ("aws-access-key", re.compile(r"\bAKIA[0-9A-Z]{16}\b")),
+    ("github-token", re.compile(r"\bgh[pousr]_[A-Za-z0-9]{20,}\b")),
+    ("slack-token", re.compile(r"\bxox[baprs]-[A-Za-z0-9-]{10,}\b")),
+    ("openai-key", re.compile(r"\bsk-[A-Za-z0-9]{20,}\b")),
+    ("anthropic-key", re.compile(r"\bsk-ant-[A-Za-z0-9_\-]{20,}\b")),
+    ("jwt", re.compile(r"\beyJ[A-Za-z0-9_\-]{10,}\.[A-Za-z0-9_\-]{10,}\.[A-Za-z0-9_\-]{10,}\b")),
+    ("age-key", re.compile(r"\bAGE-SECRET-KEY-1[0-9A-Z]{20,}\b")),
+    (
+        "assigned-secret",
+        re.compile(
+            r"(?im)^\s*(?:export\s+)?[A-Z0-9_]*(?:SECRET|TOKEN|PASSWORD|PASSWD|API_KEY|"
+            r"ACCESS_KEY|PRIVATE_KEY)[A-Z0-9_]*\s*[:=]\s*[\"']?[^\s\"'#]{8,}"
+        ),
+    ),
+)
+
+# A `.env.example`-style file may document the *name* of a variable. It may not
+# carry a value. This filter keeps `GITEA_TOKEN=` and `TOKEN=changeme` out.
+_PLACEHOLDERS = {"", "changeme", "change-me", "xxx", "todo", "placeholder", "your-token-here", "redacted"}
+
+
+class Rejection:
+    """One file that was not admitted, and the rule that refused it."""
+
+    __slots__ = ("path", "platform", "reason", "detail")
+
+    def __init__(self, path: str, platform: str, reason: str, detail: str = "") -> None:
+        self.path = path
+        self.platform = platform
+        self.reason = reason
+        self.detail = detail
+
+    def to_dict(self) -> dict:
+        return {
+            "path": self.path,
+            "platform": self.platform,
+            "reason": self.reason,
+            "detail": self.detail,
+        }
+
+
+def extension(path: str) -> str:
+    base = path.rsplit("/", 1)[-1].lower()
+    if "." not in base:
+        return base
+    return base.rsplit(".", 1)[-1]
+
+
+def basename(path: str) -> str:
+    return path.rsplit("/", 1)[-1]
+
+
+def classify(path: str, guide=None, size: int | None = None) -> tuple:
+    """Return `(kind, reason)` for a path.
+
+    `kind` is one of `text`, `image`, `reject`. `reason` is empty for text and
+    names the rule for the other two. Guide excludes are applied here so a
+    caller cannot forget them.
+    """
+    path = globs.normalise(path)
+    if not path:
+        return "reject", "empty path"
+
+    parts = path.split("/")
+    for part in parts[:-1]:
+        if part.lower() in EXCLUDED_DIRS:
+            return "reject", f"excluded directory {part!r}"
+
+    name = basename(path)
+    lowered = name.lower()
+    if lowered in EXCLUDED_DIRS:
+        return "reject", f"excluded directory {lowered!r}"
+    for pattern in CREDENTIAL_FILES:
+        if globs.matches(lowered, pattern) or globs.matches(name, pattern):
+            return "reject", "credential-shaped file name"
+
+    if guide is not None:
+        exclude = getattr(guide, "exclude", {}) or {}
+        if globs.under_any(path, exclude.get("paths")):
+            return "reject", "excluded by platform guide"
+        if globs.matches_any(path, exclude.get("files") or ()):
+            return "reject", "excluded by platform guide"
+
+    ext = extension(path)
+    if ext in MEDIA_EXTENSIONS:
+        return "image", f"media ({ext}) recorded as an image reference"
+    if ext in BINARY_EXTENSIONS:
+        return "reject", f"binary ({ext})"
+
+    if size is not None:
+        limit = int(((getattr(guide, "limits", {}) or {}).get("max_file_kb") or 200) * 1024)
+        if size > limit:
+            return "reject", f"oversize ({size} bytes > {limit})"
+
+    if guide is not None:
+        include = getattr(guide, "include", {}) or {}
+        include_paths = include.get("paths") or []
+        if include_paths and not globs.under_any(path, include_paths):
+            return "reject", "not under an included path"
+
+    if ext not in TEXT_EXTENSIONS and name.lower() not in TEXT_EXTENSIONS:
+        # Unknown extension: admitted only if the caller proves it is UTF-8 text.
+        return "unknown", ""
+    return "text", ""
+
+
+def looks_binary(data: bytes) -> bool:
+    if b"\x00" in data[:4096]:
+        return True
+    try:
+        data.decode("utf-8")
+    except UnicodeDecodeError:
+        return True
+    return False
+
+
+def scan_text(text: str) -> list:
+    """Return `[(kind, line_number, excerpt)]` for credential-shaped content.
+
+    The excerpt is short and never includes the matched value: a rejection
+    report must not become the leak it is reporting.
+    """
+    hits = []
+    for number, line in enumerate(text.splitlines(), start=1):
+        for kind, pattern in _CREDENTIAL_PATTERNS:
+            match = pattern.search(line)
+            if not match:
+                continue
+            value = match.group(0)
+            if kind == "assigned-secret":
+                assigned = value.split("=", 1)[-1].split(":", 1)[-1].strip().strip("\"'")
+                if assigned.lower() in _PLACEHOLDERS:
+                    continue
+            hits.append((kind, number, line.strip()[:60].split("=", 1)[0]))
+            break
+    return hits

corpus/freshness.py

@@ -0,0 +1,279 @@
+"""Freshness audit — the thing that stops the corpus rotting silently.
+
+The risk table in `docs/phase-4-corpus.md` names it: *"Silent staleness… a
+stale index is otherwise an invisible quality failure."* So the audit is a first
+class command, it runs nightly from Prefect, and it alerts through the existing
+Apprise service when it finds something.
+
+It reports four kinds of finding, and they are different problems:
+
+* **stale** — a document whose `last_verified` is older than the horizon. The
+  claim on the page may no longer be true.
+* **due soon** — inside the warning window. Re-verify before it becomes stale.
+* **index drift** — the source changed (or disappeared, or was never indexed)
+  since it was embedded. Retrieval is answering from an older world.
+* **contract drift** — the collection was built with a different model,
+  dimension or chunker than the running configuration.
+
+Exit code is 0 for a clean audit and 1 when findings exist and `--fail-on-findings`
+was passed, so the Prefect deployment can alert on the result without parsing
+text.
+"""
+
+from __future__ import annotations
+
+import datetime
+import json
+import urllib.request
+
+from . import loader, store as store_module
+from .config import CHUNKER_VERSION
+from .pipeline import _contract
+from .qdrant import Qdrant
+from .sources import estate as estate_module
+
+
+def audit(cfg, today: datetime.date = None, log=None) -> dict:
+    """Return the audit report as a plain dict (JSON-serialisable)."""
+    log = log or (lambda _msg: None)
+    today = today or datetime.date.today()
+
+    # The audit reads the same estate the pipeline reads, so "the source has
+    # stopped changing" is measured against the source, not against the index.
+    harvest = None
+    try:
+        harvest = estate_module.harvest(cfg, log=log)
+        documents = harvest.documents
+    except estate_module.EstateError as exc:
+        documents = loader.load_documents(cfg)
+        estate_error = str(exc)
+    else:
+        estate_error = ""
+
+    findings = {
+        "stale": [],
+        "due_soon": [],
+        "index_drift": [],
+        "contract_drift": [],
+        "stale_source": [],
+        "source_missing": [],
+        "source_unavailable": [],
+    }
+    if estate_error:
+        findings["source_unavailable"].append({"platform": "estate", "problem": estate_error})
+
+    if harvest is not None:
+        for name, row in sorted((harvest.sources or {}).items()):
+            if not row.get("enabled"):
+                continue
+            # Curated content is measured by last_verified above; source silence
+            # is about the harvested estate, which should change on its own.
+            if name == "curated":
+                continue
+            if row.get("missing") or not row.get("documents"):
+                findings["source_missing"].append(
+                    {
+                        "platform": name,
+                        "problem": "no documents were harvested",
+                        "repo": row.get("repo") or row.get("base") or "",
+                    }
+                )
+                continue
+            newest = row.get("newest") or ""
+            try:
+                changed = datetime.date.fromisoformat(newest[:10])
+            except ValueError:
+                findings["source_missing"].append(
+                    {"platform": name, "problem": f"no usable last_modified ({newest!r})"}
+                )
+                continue
+            age = (today - changed).days
+            if age > cfg.source_silence_days:
+                findings["stale_source"].append(
+                    {
+                        "platform": name,
+                        "repo": row.get("repo") or row.get("base") or "",
+                        "newest": newest,
+                        "age_days": age,
+                        "problem": f"the source has not changed in {age} days",
+                    }
+                )
+
+    for doc in documents:
+        try:
+            verified = datetime.date.fromisoformat(doc.last_verified)
+        except ValueError:
+            findings["index_drift"].append(
+                {"source_path": doc.path, "problem": f"invalid last_verified {doc.last_verified!r}"}
+            )
+            continue
+        age = (today - verified).days
+        entry = {
+            "source_path": doc.path,
+            "title": doc.title,
+            "area": doc.area,
+            "confidence": doc.confidence,
+            "last_verified": doc.last_verified,
+            "age_days": age,
+        }
+        if age > cfg.max_age_days:
+            findings["stale"].append(entry)
+        elif age > cfg.max_age_days - cfg.warn_days:
+            findings["due_soon"].append(entry)
+
+    with store_module.Store(cfg.db_path) as store:
+        manifest = store.manifest()
+        on_disk = {doc.path: doc for doc in documents}
+
+        for doc in documents:
+            row = manifest.get(doc.path)
+            if row is None:
+                findings["index_drift"].append(
+                    {"source_path": doc.path, "problem": "not indexed (run sync)"}
+                )
+                continue
+            if row["content_hash"] != doc.digest:
+                findings["index_drift"].append(
+                    {"source_path": doc.path, "problem": "source changed since it was embedded"}
+                )
+            if row["chunker_version"] != CHUNKER_VERSION:
+                findings["index_drift"].append(
+                    {
+                        "source_path": doc.path,
+                        "problem": f"chunked with {row['chunker_version']}, current {CHUNKER_VERSION}",
+                    }
+                )
+
+        for source in sorted(set(manifest) - set(on_disk)):
+            findings["index_drift"].append(
+                {"source_path": source, "problem": "indexed but no longer in content/"}
+            )
+
+        graph_counts = store.graph_counts()
+        stored_contract = store.get_meta("collection")
+        if stored_contract != _contract(cfg):
+            findings["contract_drift"].append(
+                {"problem": "collection contract does not match this configuration",
+                 "stored": stored_contract, "configured": _contract(cfg)}
+            )
+
+        qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+        reachable = qdrant.health()
+        points = None
+        if reachable and qdrant.collection_exists(cfg.collection):
+            points = qdrant.point_count(cfg.collection)
+        elif not reachable:
+            findings["contract_drift"].append(
+                {"problem": f"Qdrant is not answering at {cfg.qdrant_url}"}
+            )
+
+        report = {
+            "generated_at": datetime.datetime.now(datetime.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
+            "today": today.isoformat(),
+            "horizon": {
+                "max_age_days": cfg.max_age_days,
+                "warn_days": cfg.warn_days,
+            },
+            "corpus": {
+                "documents": len(documents),
+                "chunks": store.chunk_count(),
+                "sources": store.source_count(),
+                "collection": cfg.collection,
+                "collection_points": points,
+                "graph_collection": cfg.graph_collection,
+                "graph_nodes": graph_counts["nodes"],
+                "graph_edges": graph_counts["edges"],
+                "provider": cfg.provider,
+                "model": cfg.model,
+                "dim": cfg.dim,
+                "chunker_version": CHUNKER_VERSION,
+            },
+            "findings": findings,
+            "sources": (harvest.sources if harvest is not None else {}),
+            "warnings": (harvest.warnings if harvest is not None else []),
+            "summary": {
+                "stale": len(findings["stale"]),
+                "due_soon": len(findings["due_soon"]),
+                "index_drift": len(findings["index_drift"]),
+                "contract_drift": len(findings["contract_drift"]),
+                "stale_source": len(findings["stale_source"]),
+                "source_missing": len(findings["source_missing"]),
+                "source_unavailable": len(findings["source_unavailable"]),
+            },
+            # `due_soon` is a warning, not a failure: the document is still
+            # inside its verification horizon. Stale, drift and contract
+            # problems are failures.
+            "ok": not (
+                findings["stale"]
+                or findings["index_drift"]
+                or findings["contract_drift"]
+                or findings["stale_source"]
+                or findings["source_missing"]
+                or findings["source_unavailable"]
+            ),
+        }
+    log(f"audit: {json.dumps(report['summary'], sort_keys=True)}")
+    return report
+
+
+def render_text(report: dict) -> str:
+    """Human-readable audit, for a terminal or an Apprise body."""
+    lines = []
+    corpus = report["corpus"]
+    graph = ""
+    if corpus.get("graph_nodes") is not None:
+        graph = f", graph {corpus['graph_nodes']} node(s) / {corpus['graph_edges']} edge(s)"
+    lines.append(
+        f"corpus: {corpus['documents']} document(s), {corpus['chunks']} chunk(s), "
+        f"{corpus['collection_points']} vector(s) in {corpus['collection']}{graph}"
+    )
+    lines.append(
+        f"horizon: last_verified within {report['horizon']['max_age_days']} days "
+        f"(warning at {report['horizon']['max_age_days'] - report['horizon']['warn_days']} days)"
+    )
+    findings = report["findings"]
+    for kind, label in (
+        ("stale", "STALE — past the verification horizon"),
+        ("due_soon", "DUE SOON — verify before it goes stale"),
+        ("index_drift", "INDEX DRIFT — the index does not match the sources"),
+        ("contract_drift", "CONTRACT DRIFT — rebuild required"),
+        ("stale_source", "STALE SOURCE — the estate has stopped changing"),
+        ("source_missing", "SOURCE MISSING — an enabled platform produced nothing"),
+        ("source_unavailable", "SOURCE UNAVAILABLE — the estate could not be read"),
+    ):
+        rows = findings.get(kind) or []
+        if not rows:
+            continue
+        lines.append("")
+        lines.append(f"{label} ({len(rows)})")
+        for row in rows:
+            if "problem" in row:
+                lines.append(f"  - {row.get('source_path', '<collection>')}: {row['problem']}")
+            else:
+                lines.append(
+                    f"  - {row['source_path']} ({row['area']}, {row['confidence']}): "
+                    f"last_verified {row['last_verified']} ({row['age_days']} days ago)"
+                )
+    if report["ok"]:
+        lines.append("")
+        lines.append(
+            "OK: nothing stale, no drift, every source current."
+            + (" (see warnings above)" if findings["due_soon"] else "")
+        )
+    return "\n".join(lines)
+
+
+def alert(report: dict, apprise_url: str, timeout: float = 10.0) -> bool:
+    """POST the audit to the existing Apprise service. No-op without a URL."""
+    if not apprise_url:
+        return False
+    title = "ai-resume corpus audit: findings" if not report["ok"] else "ai-resume corpus audit: OK"
+    body = json.dumps({"title": title, "body": render_text(report), "type": "warning" if not report["ok"] else "info"})
+    request = urllib.request.Request(
+        apprise_url,
+        data=body.encode("utf-8"),
+        method="POST",
+        headers={"Content-Type": "application/json"},
+    )
+    with urllib.request.urlopen(request, timeout=timeout):
+        return True

corpus/globs.py

@@ -0,0 +1,89 @@
+"""Path globs, with the one extension the guides need: `**/` matches nothing.
+
+`fnmatch` is close but wrong for this job in two ways: it lets `*` cross a `/`,
+and it has no `**`. Platform guides are read by a human and written as ordinary
+gitignore-style globs (`**/node_modules/**`, `200 projects/**`), so this module
+implements exactly that little language and nothing else.
+
+Matching is always against a normalised, repository-relative POSIX path.
+"""
+
+from __future__ import annotations
+
+import re
+
+_CACHE: dict = {}
+
+
+def normalise(path: str) -> str:
+    """Repository-relative, POSIX separators, no leading `./` or `/`."""
+    path = str(path).replace("\\", "/").strip()
+    while path.startswith("./"):
+        path = path[2:]
+    return path.lstrip("/")
+
+
+def compile_glob(pattern: str):
+    """Compile one glob to a regular expression. Cached."""
+    cached = _CACHE.get(pattern)
+    if cached is not None:
+        return cached
+
+    pattern = normalise(pattern)
+    out = ["^"]
+    index = 0
+    while index < len(pattern):
+        char = pattern[index]
+        if char == "*":
+            if pattern[index : index + 2] == "**":
+                index += 2
+                if pattern[index : index + 1] == "/":
+                    # `**/` — any number of whole directories, including none.
+                    out.append("(?:[^/]+/)*")
+                    index += 1
+                else:
+                    out.append(".*")
+                continue
+            out.append("[^/]*")
+        elif char == "?":
+            out.append("[^/]")
+        elif char == "[":
+            close = pattern.find("]", index + 1)
+            if close == -1:
+                out.append(re.escape(char))
+            else:
+                body = pattern[index + 1 : close]
+                if body.startswith("!"):
+                    body = "^" + body[1:]
+                out.append("[" + body + "]")
+                index = close
+        else:
+            out.append(re.escape(char))
+        index += 1
+    out.append("$")
+    compiled = re.compile("".join(out))
+    _CACHE[pattern] = compiled
+    return compiled
+
+
+def matches(path: str, pattern: str) -> bool:
+    return bool(compile_glob(pattern).match(normalise(path)))
+
+
+def matches_any(path: str, patterns) -> bool:
+    return any(matches(path, pattern) for pattern in (patterns or ()))
+
+
+def under_any(path: str, patterns) -> bool:
+    """True when `path` is a pattern's directory itself or inside it.
+
+    `200 projects/**` should also exclude the bare directory entry `200 projects`
+    that a tree listing returns, not only the files underneath it.
+    """
+    path = normalise(path)
+    for pattern in patterns or ():
+        if matches(path, pattern):
+            return True
+        if pattern.endswith("/**") and path == normalise(pattern[:-3]):
+            return True
+    return False

corpus/guides.py

@@ -0,0 +1,227 @@
+"""Platform guides: one small file per platform, configuration and public copy.
+
+`platforms/<name>.yml` says what a platform is, where it lives, what to include,
+what to exclude, which resume area it feeds, and whether it is enabled. The same
+wording is shown on the public site, so every value here is written to be read
+by a visitor, not just by the pipeline.
+
+The schema is fixed and validated at load time. An unknown area, a missing
+`what`, or a guide whose filename and `name` disagree is a configuration error,
+not a default. A disabled guide is loaded and shown, but not harvested.
+
+There is exactly one YAML dialect in this repository and one parser for it
+(`tools/validate_content.py`); this module imports that parser rather than
+adding a second one.
+"""
+
+from __future__ import annotations
+
+import os
+from dataclasses import dataclass, field
+
+from . import loader
+from .payloads import AREAS  # one area enum for guides and payloads
+
+# The platforms the pipeline knows how to harvest. `curated` is content/ and is
+# always on: it is tier 1 and has no guide because it is not an external estate.
+KINDS = ("gitea", "obsidian", "graphify")
+REQUIRED = ("name", "enabled", "what", "fits", "read")
+DEFAULT_LIMITS = {
+    "max_documents": 4000,
+    "max_file_kb": 200,
+    "max_files_per_repo": 40,
+    "max_commits": 20,
+    "max_tree_entries": 3000,
+    "max_nodes": 20000,
+    "max_edges": 40000,
+}
+
+
+class GuideError(RuntimeError):
+    """A platform guide is missing, malformed, or inconsistent."""
+
+
+@dataclass
+class Guide:
+    name: str
+    enabled: bool
+    what: str
+    fits: list
+    read: list
+    base: str = ""
+    auth: str = ""
+    include: dict = field(default_factory=dict)
+    exclude: dict = field(default_factory=dict)
+    limits: dict = field(default_factory=dict)
+    public: bool = True
+    path: str = ""  # repository-relative, for the public page and error messages
+
+    # -- derived ----------------------------------------------------------
+
+    @property
+    def kind(self) -> str:
+        return self.name
+
+    @property
+    def primary_area(self) -> str:
+        return self.fits[0]
+
+    def limit(self, name: str) -> int:
+        return int(self.limits.get(name) or DEFAULT_LIMITS.get(name) or 0)
+
+    def to_public_dict(self) -> dict:
+        """What the site shows. No auth variable values, only its name."""
+        return {
+            "name": self.name,
+            "enabled": self.enabled,
+            "what": self.what,
+            "fits": list(self.fits),
+            "read": list(self.read),
+            "base": self.base,
+            "include": {
+                "repos": _as_list((self.include or {}).get("repos")),
+                "paths": _as_list((self.include or {}).get("paths")),
+            },
+            "exclude": {
+                "repos": _as_list((self.exclude or {}).get("repos")),
+                "paths": _as_list((self.exclude or {}).get("paths")),
+            },
+            "public": self.public,
+            "path": self.path,
+        }
+
+    def to_dict(self) -> dict:
+        data = self.to_public_dict()
+        data["limits"] = dict(self.limits)
+        data["auth"] = self.auth  # the variable name, never its value
+        return data
+
+
+def load_guides(root: str, directory: str = None) -> dict:
+    """Load every guide under platforms/, keyed by name. Sorted, deterministic."""
+    directory = directory or os.path.join(root, "platforms")
+    if not os.path.isdir(directory):
+        raise GuideError(f"platform guide directory not found: {directory}")
+
+    validator = loader.load_validator(root)
+    guides = {}
+    for filename in sorted(os.listdir(directory)):
+        if not filename.endswith((".yml", ".yaml")):
+            continue
+        path = os.path.join(directory, filename)
+        rel = os.path.relpath(path, root).replace(os.sep, "/")
+        guide = _load_one(validator, path, rel, os.path.splitext(filename)[0])
+        if guide.name in guides:
+            raise GuideError(f"{rel}: duplicate platform name {guide.name!r}")
+        guides[guide.name] = guide
+
+    missing = sorted(set(KINDS) - set(guides))
+    if missing:
+        raise GuideError(
+            f"no platform guide for {', '.join(missing)} under {directory}. "
+            "One small file per platform is required (docs/phase-4-corpus.md §2)."
+        )
+    return guides
+
+
+def _load_one(validator, path: str, rel: str, stem: str) -> Guide:
+    with open(path, "r", encoding="utf-8") as handle:
+        text = handle.read()
+    lines = [(number, line) for number, line in enumerate(text.splitlines(), start=1)]
+    try:
+        node = validator.parse_front_matter(lines)
+    except validator.YamlError as exc:
+        raise GuideError(f"{rel}:{exc.line}: invalid YAML: {exc.msg}") from exc
+    values = loader.to_python(node)
+    if not isinstance(values, dict):
+        raise GuideError(f"{rel}: a platform guide must be a mapping")
+
+    for key in REQUIRED:
+        if values.get(key) in (None, "", []):
+            raise GuideError(f"{rel}: missing required key {key!r}")
+
+    name = str(values["name"]).strip()
+    if name != stem:
+        raise GuideError(f"{rel}: name {name!r} does not match the filename {stem!r}")
+    if name not in KINDS:
+        raise GuideError(f"{rel}: unknown platform {name!r}; expected one of {list(KINDS)}")
+
+    fits = [str(area) for area in values["fits"]]
+    unknown = [area for area in fits if area not in AREAS]
+    if unknown:
+        raise GuideError(f"{rel}: fits contains unknown area(s) {unknown}; allowed {list(AREAS)}")
+
+    read = [str(item) for item in values["read"]]
+    if not read:
+        raise GuideError(f"{rel}: read must name at least one thing to harvest")
+
+    include = _mapping(values.get("include"), f"{rel}: include")
+    exclude = _mapping(values.get("exclude"), f"{rel}: exclude")
+    limits = _mapping(values.get("limits"), f"{rel}: limits")
+    for key, value in limits.items():
+        if isinstance(value, list):
+            # `graph_paths` is the one list-valued limit: the places a graph may
+            # be committed, tried in order.
+            continue
+        try:
+            limits[key] = int(value)
+        except (TypeError, ValueError):
+            raise GuideError(f"{rel}: limit {key!r} must be an integer, got {value!r}") from None
+
+    return Guide(
+        name=name,
+        enabled=bool(values.get("enabled")),
+        what=str(values["what"]).strip(),
+        fits=fits,
+        read=read,
+        base=str(values.get("base") or "").strip(),
+        auth=str(values.get("auth") or "").strip(),
+        include=include,
+        exclude=exclude,
+        limits=limits,
+        public=bool(values.get("public", True)),
+        path=rel,
+    )
+
+
+def _mapping(value, where: str) -> dict:
+    if value is None:
+        return {}
+    if not isinstance(value, dict):
+        raise GuideError(f"{where} must be a mapping")
+    out = {}
+    for key, item in value.items():
+        if isinstance(item, list):
+            out[key] = [str(entry) for entry in item]
+        else:
+            out[key] = item
+    return out
+
+
+def _as_list(value) -> list:
+    """`repos: all` is a scalar; the public manifest still emits a list."""
+    if value in (None, ""):
+        return []
+    if isinstance(value, list):
+        return [str(item) for item in value]
+    return [str(value)]
+
+
+def enabled_guides(guides: dict) -> list:
+    return [guide for _, guide in sorted(guides.items()) if guide.enabled]
+
+
+def auth_value(guide: Guide, environ=None) -> str:
+    """Read a guide's token from the environment. Never from a file in the repo."""
+    if not guide.auth:
+        return ""
+    environ = os.environ if environ is None else environ
+    return environ.get(guide.auth, "")
+
+
+def build_public_manifest(guides: dict) -> dict:
+    """The guides as the public site shows them. Deterministic JSON shape."""
+    return {
+        "version": 1,
+        "platforms": [guide.to_public_dict() for _, guide in sorted(guides.items())],
+    }

corpus/loader.py

@@ -0,0 +1,180 @@
+"""Load `content/` into documents, reusing the Phase 1 validator's parser.
+
+There is exactly one front-matter dialect in this repository and exactly one
+parser for it: `tools/validate_content.py`. This module imports that parser
+rather than reimplementing it, so the corpus can never disagree with the gate
+about what a document says.
+
+The gate itself is `validate_tree` from the same file. The pipeline runs it
+before it chunks anything: an invalid content tree stops the run, it never
+reaches the vector store.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import importlib.util
+import os
+import sys
+from dataclasses import dataclass, field
+
+VALIDATOR_NAME = "validate_content"
+
+
+class ContentError(RuntimeError):
+    """The content tree is unreadable or invalid; the run must stop."""
+
+
+def load_validator(root: str):
+    """Import tools/validate_content.py as a module (cached per root)."""
+    cached = sys.modules.get(VALIDATOR_NAME)
+    if cached is not None and getattr(cached, "__validator_root__", None) == root:
+        return cached
+    path = os.path.join(root, "tools", "validate_content.py")
+    if not os.path.exists(path):
+        raise ContentError(f"validator not found at {path}")
+    spec = importlib.util.spec_from_file_location(VALIDATOR_NAME, path)
+    if spec is None or spec.loader is None:
+        raise ContentError(f"cannot load validator at {path}")
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    module.__validator_root__ = root  # type: ignore[attr-defined]
+    sys.modules[VALIDATOR_NAME] = module
+    return module
+
+
+def to_python(node):
+    """Convert a validator front-matter Node into plain Python values."""
+    if node is None or node.kind == "null":
+        return None
+    if node.kind == "scalar":
+        return node.value
+    if node.kind == "seq":
+        return [to_python(item) for item in node.value]
+    if node.kind == "map":
+        return {key: to_python(value) for key, value in node.value.items()}
+    raise ContentError(f"unexpected front-matter node kind {node.kind!r}")
+
+
+@dataclass
+class Document:
+    """One Markdown file plus the front-matter the schema guarantees."""
+
+    path: str  # repository-relative, posix, e.g. "content/edge/l3/edge-zigbee.md"
+    source_id: str
+    area: str
+    level: int
+    title: str
+    claim: str
+    confidence: str
+    period: str
+    tech: list
+    metrics: list
+    artefacts: list
+    public: bool
+    last_verified: str
+    tags: list
+    body: str
+    digest: str  # sha256 of the file bytes: the incremental-mode key
+    url: str  # canonical page URL, matching internal/content.PageURL
+
+    # Provenance (docs/phase-4-corpus.md §3). Curated content is tier 1 and has
+    # no external origin; harvested documents set these in corpus/sources/.
+    platform: str = "curated"
+    tier: int = 1
+    origin_url: str = ""
+    last_modified: str = ""
+    source_type: str = ""  # empty means: derive from tags (payloads.source_type_for)
+    kind: str = "markdown"  # "markdown" or "code", selects the chunker
+    repo_path: str = ""  # origin-relative path; `path` above is the unique source key
+
+    # Filled in by the pipeline as it goes; not part of the schema.
+    chunk_count: int = field(default=0)
+
+    @property
+    def artefact_urls(self) -> list:
+        return [str(a.get("url", "")) for a in self.artefacts if a.get("url")]
+
+
+def page_url(area: str, doc_id: str, level: int) -> str:
+    """Canonical URL for a document. Mirrors internal/content.PageURL exactly."""
+    if level == 1:
+        return f"/{area}/"
+    if level == 2:
+        return f"/{area}/{doc_id}/"
+    return f"/{area}/a/{doc_id}/"
+
+
+def digest_of(path: str) -> str:
+    """sha256 of the raw file bytes — the same digest for any platform."""
+    with open(path, "rb") as handle:
+        return hashlib.sha256(handle.read()).hexdigest()
+
+
+def load_documents(cfg) -> list:
+    """Load every document under content/, sorted by path.
+
+    Raises ContentError when the gate fails, so a broken content tree never
+    produces a half-correct index.
+    """
+    validator = load_validator(cfg.root)
+    violations, count = validator.validate_tree(cfg.root, cfg.content_dir, quiet=True)
+    if violations:
+        detail = "\n  ".join(violations[:20])
+        more = "" if len(violations) <= 20 else f"\n  … and {len(violations) - 20} more"
+        raise ContentError(
+            f"content gate failed with {len(violations)} violation(s) "
+            f"across {count} document(s). Run `make validate`:\n  {detail}{more}"
+        )
+
+    documents = []
+    for path in validator.iter_documents(cfg.content_dir):
+        documents.append(_load_one(cfg, validator, path))
+    documents.sort(key=lambda d: d.path)
+    if not documents:
+        raise ContentError(f"no documents found under {cfg.content_dir}")
+    return documents
+
+
+def _load_one(cfg, validator, path: str) -> Document:
+    rel = os.path.relpath(path, cfg.root).replace(os.sep, "/")
+    fm_lines, body, problem = validator.split_front_matter(path)
+    if problem:
+        raise ContentError(f"{rel}: {problem}")
+    try:
+        root_node = validator.parse_front_matter(fm_lines)
+    except validator.YamlError as exc:
+        raise ContentError(f"{rel}:{exc.line}: invalid YAML: {exc.msg}") from exc
+    fields = to_python(root_node)
+    if not isinstance(fields, dict):
+        raise ContentError(f"{rel}: front-matter must be a mapping")
+
+    level = int(fields["level"])
+    doc = Document(
+        path=rel,
+        source_id=str(fields["id"]),
+        area=str(fields["area"]),
+        level=level,
+        title=str(fields["title"]),
+        claim=str(fields["claim"]),
+        confidence=str(fields["confidence"]),
+        period=str(fields["period"]),
+        tech=list(fields.get("tech") or []),
+        metrics=list(fields.get("metrics") or []),
+        artefacts=list(fields.get("artefacts") or []),
+        public=bool(fields.get("public")),
+        last_verified=str(fields["last_verified"]),
+        tags=list(fields.get("tags") or []),
+        body=body,
+        digest=digest_of(path),
+        url=page_url(str(fields["area"]), str(fields["id"]), level),
+        platform="curated",
+        tier=1,
+        # A curated claim is traceable to the page it appears on. The URL is
+        # site-relative because the public origin is a deployment concern.
+        origin_url=page_url(str(fields["area"]), str(fields["id"]), level),
+        last_modified=f"{fields['last_verified']}T00:00:00Z",
+        kind="markdown",
+        repo_path=rel,
+    )
+    return doc

corpus/measure.py

@@ -0,0 +1,242 @@
+"""Qdrant memory at 500 / 5,000 / 20,000 chunks.
+
+`docs/phase-4-corpus.md` requires the measurement and `reference/stack.md` is
+where the number is recorded. The corpus itself is small; the point of the
+measurement is to show the *shape* of the growth so the 400 m cap is a known
+quantity rather than a hope.
+
+Method:
+
+1. Create an empty collection and sample memory — the baseline.
+2. For each size: recreate the collection, upsert that many chunks with payloads
+   shaped exactly like production payloads (same fields, same text lengths),
+   wait for the write to settle, then sample.
+
+Memory is read from Qdrant's own Prometheus `app_process_resident_memory_bytes`,
+which is the process that would be capped, not the container sandbox. When
+`--container` is also given the container's counter is recorded next to it.
+
+The measurement collection is dropped afterwards unless `--keep` is passed, and
+it is never the production collection name.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import re
+import statistics
+import subprocess
+import time
+import uuid
+
+from . import chunker, embeddings, loader, payloads
+from .qdrant import Qdrant
+
+# Qdrant 1.20 exposes memory_resident_bytes; older builds used the Prometheus
+# default app_process_resident_memory_bytes. Accept either.
+RSS_METRIC = re.compile(
+    r"^(?:memory_resident_bytes|app_process_resident_memory_bytes)\s+([0-9.eE+]+)",
+    re.MULTILINE,
+)
+INGEST_BATCH = 250
+INGEST_SETTLE = 0.4
+MEMUSAGE = re.compile(r"([0-9.]+)\s*([KMG]i?B)", re.IGNORECASE)
+CAP_MIB = 400  # reference/stack.md memory budget for Qdrant
+
+
+def measure(cfg, sizes=(500, 5000, 20000), container: str = "", keep: bool = False, log=None) -> dict:
+    log = log or (lambda _msg: None)
+    collection = f"{cfg.collection}_measure"
+    provider = embeddings.build_provider(cfg)
+    qdrant = Qdrant(cfg.qdrant_url, timeout=max(cfg.timeout, 60.0))
+    if not qdrant.health():
+        raise RuntimeError(f"Qdrant is not answering at {cfg.qdrant_url}")
+
+    templates = _templates(cfg, provider)
+    log(f"measure: {len(templates)} payload template(s), dim={cfg.dim}, collection={collection}")
+
+    qdrant.drop_collection(collection)
+    _create(qdrant, cfg, collection)
+    time.sleep(1.0)
+    baseline = _sample(qdrant, container)
+
+    samples = []
+    for size in sorted(int(value) for value in sizes):
+        qdrant.drop_collection(collection)
+        _create(qdrant, cfg, collection)
+        points = _points(templates, size, provider)
+        started = time.time()
+        # Throttled ingest: the peak during a bulk write is what trips the
+        # cgroup limit, not the settled size. One batch at a time, sampling as we
+        # go, so the measurement reports the peak and not just the end state.
+        sent = 0
+        peak = baseline["rss_mib"]
+        for start in range(0, len(points), INGEST_BATCH):
+            sent += qdrant.upsert(collection, points[start : start + INGEST_BATCH], batch=INGEST_BATCH)
+            time.sleep(INGEST_SETTLE)
+            reading = _rss_mib(qdrant)
+            if reading is not None:
+                peak = max(peak, reading)
+        elapsed = round(time.time() - started, 2)
+        points_now = qdrant.point_count(collection)
+        time.sleep(2.0)
+        reading = _sample(qdrant, container)
+        growth = reading["rss_mib"] - baseline["rss_mib"]
+        samples.append(
+            {
+                "chunks_requested": size,
+                "points": points_now,
+                "sent": sent,
+                "upsert_seconds": elapsed,
+                "rss_mib": reading["rss_mib"],
+                "peak_rss_mib": round(peak, 1),
+                "container_mib": reading["container_mib"],
+                "growth_mib": round(growth, 1),
+                "bytes_per_chunk": int(growth * 1024 * 1024 / size) if size else None,
+            }
+        )
+        log(
+            f"measure: {size} chunk(s) -> {points_now} point(s), rss {reading['rss_mib']:.1f} MiB "
+            f"(+{growth:.1f} MiB), upsert {elapsed}s"
+        )
+
+    if not keep:
+        qdrant.drop_collection(collection)
+
+    report = {
+        "collection": collection,
+        "dim": cfg.dim,
+        "distance": "Cosine",
+        "provider": cfg.provider,
+        "model": cfg.model,
+        "qdrant_version": qdrant.version(),
+        "cap_mib": CAP_MIB,
+        "baseline": baseline,
+        "samples": samples,
+        "within_cap": all(
+            max(sample["rss_mib"], sample["peak_rss_mib"]) <= CAP_MIB for sample in samples
+        ),
+        "measured_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
+        "kept": keep,
+    }
+    return report
+
+
+def _create(qdrant: Qdrant, cfg, collection: str) -> None:
+    """Same collection settings as production, so the memory reading means something."""
+    qdrant.create_collection(
+        collection, cfg.dim, vectors_on_disk=cfg.vectors_on_disk
+    )
+
+
+def _templates(cfg, provider) -> list:
+    """Real payloads from content/, so field count and text length are realistic."""
+    documents = loader.load_documents(cfg)
+    out = []
+    for doc in documents:
+        for chunk in chunker.chunk_document(doc.body):
+            payload = payloads.build_payload(doc, chunk, cfg)
+            payload["id"] = "template"
+            out.append(payload)
+    if not out:
+        raise RuntimeError("no content to build a measurement from")
+    return out
+
+
+def _points(templates: list, size: int, provider) -> list:
+    """`size` points with unique ids, unique source paths and real payload shape."""
+    points = []
+    texts = []
+    for index in range(size):
+        template = templates[index % len(templates)]
+        payload = dict(template)
+        payload["id"] = str(uuid.UUID(hex=hashlib.sha256(f"measure:{index}".encode()).hexdigest()[:32]))
+        payload["source_path"] = f"measure/{template['area']}/{index}.md"
+        payload["chunk_index"] = index
+        points.append({"id": payload["id"], "payload": payload})
+        texts.append(f"{index} {payload['text']}")
+    vectors = provider.embed(texts)
+    for point, vector in zip(points, vectors):
+        point["vector"] = vector
+    return points
+
+
+def _sample(qdrant: Qdrant, container: str) -> dict:
+    """Three readings two seconds apart; report the median so a GC pause is not the number."""
+    readings = []
+    for attempt in range(3):
+        readings.append(_rss_mib(qdrant))
+        if attempt < 2:
+            time.sleep(2.0)
+    readings = [value for value in readings if value is not None]
+    rss = round(statistics.median(readings), 1) if readings else 0.0
+    return {"rss_mib": rss, "container_mib": _container_mib(container) if container else None}
+
+
+def _rss_mib(qdrant: Qdrant):
+    text = qdrant.metrics()
+    if not text:
+        return None
+    match = RSS_METRIC.search(text)
+    if not match:
+        return None
+    return float(match.group(1)) / (1024 * 1024)
+
+
+def _container_mib(container: str):
+    """Container memory from `docker stats`. None when docker or the container is absent."""
+    try:
+        completed = subprocess.run(
+            ["docker", "stats", "--no-stream", "--format", "{{.MemUsage}}", container],
+            capture_output=True,
+            text=True,
+            timeout=20,
+            check=False,
+        )
+    except (OSError, subprocess.SubprocessError):
+        return None
+    if completed.returncode != 0:
+        return None
+    match = MEMUSAGE.search(completed.stdout)
+    if not match:
+        return None
+    value = float(match.group(1))
+    unit = match.group(2).upper()
+    if unit.startswith("G"):
+        value *= 1024
+    elif unit.startswith("K"):
+        value /= 1024
+    return round(value, 1)
+
+
+def render_text(report: dict) -> str:
+    lines = [
+        f"Qdrant {report['qdrant_version']} · collection {report['collection']} · "
+        f"dim {report['dim']} cosine · provider {report['provider']}/{report['model']}",
+        f"baseline (empty collection): {report['baseline']['rss_mib']:.1f} MiB",
+        "",
+        f"{'chunks':>8}  {'points':>8}  {'rss MiB':>8}  {'peak':>7}  {'growth':>7}  {'B/chunk':>8}  {'upsert s':>8}",
+    ]
+    for sample in report["samples"]:
+        per_chunk = sample["bytes_per_chunk"] or 0
+        lines.append(
+            f"{sample['chunks_requested']:>8}  {sample['points']:>8}  {sample['rss_mib']:>8.1f}  "
+            f"{sample['peak_rss_mib']:>7.1f}  {sample['growth_mib']:>+7.1f}  {per_chunk:>8}  "
+            f"{sample['upsert_seconds']:>8}"
+        )
+    lines.append("")
+    verdict = "within" if report["within_cap"] else "OVER"
+    lines.append(f"cap {report['cap_mib']} MiB: {verdict} cap at every measured size")
+    if report["kept"]:
+        lines.append(f"collection {report['collection']} was kept for inspection")
+    return "\n".join(lines)
+
+
+if __name__ == "__main__":
+    import sys
+
+    from .config import Config
+
+    cfg = Config.from_env(dim=int(sys.argv[1]) if len(sys.argv) > 1 else None)
+    print(render_text(measure(cfg, container="", log=print)))

corpus/payloads.py

@@ -0,0 +1,202 @@
+"""Chunk payloads: the thing the retrieval filter runs against.
+
+`docs/phase-4-corpus.md` §4.2 fixes the payload. `area` and `confidence` are the
+two that matter: they let the query path narrow *before* it searches, and they
+let an answer qualify unfinished work instead of presenting it as delivered.
+
+**A missing or invalid field fails the run.** Nothing in this module supplies a
+default value: `validate_payload` raises on the first field that is absent,
+empty or out of enum, and the pipeline reports which document and chunk it was.
+That is the DoD item "a missing field fails the run rather than defaulting".
+"""
+
+from __future__ import annotations
+
+import datetime
+
+AREAS = ("ai", "platform", "edge", "products", "history", "about")
+LEVELS = (1, 2, 3)
+# docs/phase-4-corpus.md §3 fixes the harvested types; adr and metric were
+# already used by curated content and stay valid.
+SOURCE_TYPES = (
+    "doc", "code", "commit", "graph_node", "graph_edge", "claim", "adr", "metric",
+)
+CONFIDENCE = ("delivered", "running", "in-progress", "abandoned", "historic")
+PLATFORMS = ("curated", "gitea", "obsidian", "graphify")
+TIERS = (1, 2)
+
+# Fields every payload must carry (docs/phase-4-corpus.md §4.2), plus the ones
+# the index needs to answer "where did this come from?" without a second lookup.
+REQUIRED_FIELDS = (
+    "area",
+    "level",
+    "source_type",
+    "confidence",
+    "period",
+    "artefacts",
+    "last_verified",
+    "text",
+    # docs/phase-4-corpus.md §3: provenance is part of the payload, not an
+    # optional extra. A chunk that cannot say where it came from is not admitted.
+    "platform",
+    "tier",
+    "origin_url",
+    "path",
+    "last_modified",
+)
+PROVENANCE_FIELDS = (
+    "id",
+    "source_path",
+    "source_id",
+    "chunk_index",
+    "title",
+    "claim",
+    "url",
+    "content_hash",
+    "embedding_model",
+    "embedding_dim",
+    "chunker_version",
+)
+
+
+class PayloadError(RuntimeError):
+    """A payload is incomplete or out of enum. The run must stop."""
+
+
+def source_type_for(doc) -> str:
+    """Classify a document.
+
+    Harvested documents carry an explicit `source_type`. Curated content does
+    not: the schema is frozen and has no such field, so it is derived from the
+    tags the schema does allow.
+    """
+    explicit = str(getattr(doc, "source_type", "") or "").strip()
+    if explicit:
+        return explicit
+    tags = {str(tag).lower() for tag in doc.tags}
+    if "adr" in tags:
+        return "adr"
+    if "metric" in tags:
+        return "metric"
+    if "code" in tags:
+        return "code"
+    return "doc"
+
+
+def build_payload(doc, chunk, cfg) -> dict:
+    """Build one chunk payload from a document and a chunk.
+
+    Values are copied, never guessed: if the document does not carry a field,
+    the payload does not gain one.
+    """
+    payload = {
+        # -- the retrieval filter (docs/phase-4-corpus.md §4.2) ------------
+        "area": doc.area,
+        "level": doc.level,
+        "source_type": source_type_for(doc),
+        "confidence": doc.confidence,
+        "period": doc.period,
+        "artefacts": [str(a.get("url", "")) for a in doc.artefacts],
+        "last_verified": doc.last_verified,
+        "text": chunk.text,
+        # -- provenance (docs/phase-4-corpus.md §3) ------------------------
+        "platform": doc.platform,
+        "tier": int(doc.tier),
+        "origin_url": doc.origin_url,
+        "path": getattr(doc, "repo_path", "") or doc.path,
+        "last_modified": doc.last_modified,
+        # -- provenance ----------------------------------------------------
+        "id": None,  # filled by the pipeline once the id is computed
+        "source_path": doc.path,
+        "source_id": doc.source_id,
+        "chunk_index": chunk.index,
+        "title": doc.title,
+        "claim": doc.claim,
+        "url": doc.url,
+        "content_hash": doc.digest,
+        "embedding_model": cfg.model,
+        "embedding_dim": cfg.dim,
+        "chunker_version": cfg.chunker_version,
+        # -- supporting evidence, used by the answer to cite ---------------
+        "public": bool(doc.public),
+        "tech": list(doc.tech),
+        "tags": list(doc.tags),
+        "chunk_tokens": chunk.tokens,
+    }
+    return payload
+
+
+def validate_payload(payload: dict, where: str = "chunk") -> None:
+    """Raise PayloadError if a required field is missing, empty or out of enum."""
+    for field in REQUIRED_FIELDS + PROVENANCE_FIELDS:
+        if field == "id" and payload.get("id") is None:
+            raise PayloadError(f"{where}: payload field 'id' was not set")
+        if field not in payload:
+            raise PayloadError(f"{where}: payload is missing required field {field!r}")
+
+    area = payload["area"]
+    if area not in AREAS:
+        raise PayloadError(f"{where}: area {area!r} is not one of {list(AREAS)}")
+    if payload["level"] not in LEVELS:
+        raise PayloadError(f"{where}: level {payload['level']!r} is not one of {list(LEVELS)}")
+    if payload["source_type"] not in SOURCE_TYPES:
+        raise PayloadError(
+            f"{where}: source_type {payload['source_type']!r} is not one of {list(SOURCE_TYPES)}"
+        )
+    if payload["confidence"] not in CONFIDENCE:
+        raise PayloadError(
+            f"{where}: confidence {payload['confidence']!r} is not one of {list(CONFIDENCE)}"
+        )
+
+    if payload["platform"] not in PLATFORMS:
+        raise PayloadError(
+            f"{where}: platform {payload['platform']!r} is not one of {list(PLATFORMS)}"
+        )
+    if payload["tier"] not in TIERS:
+        raise PayloadError(f"{where}: tier {payload['tier']!r} is not one of {list(TIERS)}")
+
+    for field in (
+        "period", "title", "claim", "url", "source_path", "source_id",
+        "embedding_model", "origin_url", "path",
+    ):
+        value = payload[field]
+        if not isinstance(value, str) or not value.strip():
+            raise PayloadError(f"{where}: field {field!r} must be a non-empty string")
+
+    text = payload["text"]
+    if not isinstance(text, str) or not text.strip():
+        raise PayloadError(f"{where}: field 'text' must be non-empty")
+
+    if not isinstance(payload["artefacts"], list):
+        raise PayloadError(f"{where}: field 'artefacts' must be a list (empty is allowed)")
+
+    _check_date(payload["last_verified"], where)
+    _check_timestamp(payload["last_modified"], where)
+    if not isinstance(payload["embedding_dim"], int) or payload["embedding_dim"] <= 0:
+        raise PayloadError(f"{where}: field 'embedding_dim' must be a positive integer")
+    if not isinstance(payload["chunk_index"], int) or payload["chunk_index"] < 0:
+        raise PayloadError(f"{where}: field 'chunk_index' must be a non-negative integer")
+    if not isinstance(payload["content_hash"], str) or len(payload["content_hash"]) != 64:
+        raise PayloadError(f"{where}: field 'content_hash' must be a sha256 hex digest")
+
+
+def _check_timestamp(value: str, where: str) -> None:
+    """`last_modified` is an ISO-8601 instant, e.g. 2026-10-14T05:00:00Z."""
+    text = str(value)
+    if text.endswith("Z"):
+        text = text[:-1] + "+00:00"
+    try:
+        datetime.datetime.fromisoformat(text)
+    except ValueError as exc:
+        raise PayloadError(
+            f"{where}: field 'last_modified' must be an ISO-8601 timestamp, got {value!r}"
+        ) from exc
+
+
+def _check_date(value: str, where: str) -> None:
+    try:
+        datetime.date.fromisoformat(str(value))
+    except ValueError as exc:
+        raise PayloadError(
+            f"{where}: field 'last_verified' must be an ISO date (YYYY-MM-DD), got {value!r}"
+        ) from exc

corpus/pipeline.py

@@ -0,0 +1,679 @@
+"""The ingestion pipeline: estate -> normalise -> chunk -> enrich -> embed -> upsert.
+
+One entry point, `run`, with two modes:
+
+* **rebuild** — drop the collections, empty the store, embed everything. This is
+  the "one command rebuilds the pool from an empty volume" path.
+* **sync** — the default. Only documents whose content hash, embedding model,
+  embedding dimension or chunker version changed are re-chunked and re-embedded.
+  Everything else is left alone, including the graph when graph.json is byte-for-byte
+  unchanged.
+
+The order matters. Every payload for every changed document is built and
+validated *before* a single vector is written, so an invalid payload stops the
+run with an empty diff rather than leaving half an index behind. Embedding is one
+flat pass over the changed chunks, batched, so the provider call count is a
+direct measure of how much work incremental mode actually did.
+
+Deterministic ids (`sha256(source_path + chunk_index + content_hash)`) mean
+re-running updates points instead of duplicating them, and the id is a valid
+UUID so Qdrant accepts it as a point id without a mapping table.
+
+Two Qdrant collections, one store:
+
+* `resume_corpus` — one point per chunk of a document, from every platform;
+* `resume_graph` — one point per graph node and one per graph edge, joined to the
+  note it came from by `source_note`.
+
+Neither is ever read by the public site (acceptance test A1): the site serves
+pre-rendered files and does not open a socket to Qdrant.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import time
+import uuid
+from dataclasses import asdict, dataclass, field
+
+from . import chunker, embeddings, exclusions, payloads, store as store_module
+from .config import CHUNKER_VERSION
+from .qdrant import Qdrant, QdrantError
+from .sources import estate as estate_module
+from .sources import graph as graph_module
+
+# How many texts one embed call carries. The provider may batch smaller
+# internally (OmniRoute does); this bounds memory on the ingest host.
+EMBED_BATCH = 128
+
+
+class PipelineError(RuntimeError):
+    """The run cannot continue without corrupting the index."""
+
+
+def chunk_id(source_path: str, chunk_index: int, content_hash: str) -> str:
+    """Deterministic point id: sha256(source_path + chunk_index + content_hash).
+
+    Rendered as a UUID because Qdrant only accepts unsigned integers or UUIDs as
+    point ids; the first 128 bits of the digest keep collisions improbable.
+    """
+    digest = hashlib.sha256(
+        f"{source_path}\x00{chunk_index}\x00{content_hash}".encode("utf-8")
+    ).hexdigest()
+    return str(uuid.UUID(hex=digest[:32]))
+
+
+@dataclass
+class RunStats:
+    mode: str = "sync"
+    documents: int = 0
+    added: int = 0
+    updated: int = 0
+    unchanged: int = 0
+    removed: int = 0
+    chunks_written: int = 0
+    chunks_deleted: int = 0
+    embed_calls: int = 0
+    texts_embedded: int = 0
+    points_sent: int = 0
+    collection_points: int = 0
+    store_chunks: int = 0
+    graph_nodes: int = 0
+    graph_edges: int = 0
+    graph_points: int = 0
+    graph_rebuilt: bool = False
+    rejections: int = 0
+    images: int = 0
+    duration_s: float = 0.0
+    collection: str = ""
+    graph_collection: str = ""
+    provider: str = ""
+    model: str = ""
+    dim: int = 0
+    chunker_version: str = CHUNKER_VERSION
+    platforms: dict = field(default_factory=dict)
+    rejection_summary: dict = field(default_factory=dict)
+    warnings: list = field(default_factory=list)
+    notes: list = field(default_factory=list)
+
+    def summary(self) -> str:
+        rows = [
+            f"mode={self.mode} documents={self.documents} chunks={self.store_chunks}",
+            f"documents added={self.added} updated={self.updated} unchanged={self.unchanged} removed={self.removed}",
+            f"chunks written={self.chunks_written} deleted={self.chunks_deleted}",
+            f"embedding calls={self.embed_calls} texts={self.texts_embedded}",
+            f"points sent={self.points_sent} collection points={self.collection_points}",
+            f"graph nodes={self.graph_nodes} edges={self.graph_edges} points={self.graph_points} "
+            f"rebuilt={self.graph_rebuilt}",
+            f"rejections={self.rejections} image references={self.images}",
+            f"platforms={json.dumps(self.platforms, sort_keys=True)}",
+            f"provider={self.provider} model={self.model} dim={self.dim} chunker={self.chunker_version}",
+            f"collection={self.collection} graph_collection={self.graph_collection} "
+            f"duration={self.duration_s:.2f}s",
+        ]
+        if self.rejection_summary:
+            rows.append(f"rejected: {json.dumps(self.rejection_summary, sort_keys=True)}")
+        rows.extend(self.notes)
+        rows.extend(f"warning: {warning}" for warning in self.warnings)
+        return "\n".join(rows)
+
+    def to_dict(self) -> dict:
+        return asdict(self)
+
+
+def run(cfg, mode: str = "sync", log=None) -> RunStats:
+    """Build or refresh the pool. Returns what happened; raises on any error."""
+    if mode not in ("sync", "rebuild"):
+        raise PipelineError(f"unknown mode {mode!r}; expected 'sync' or 'rebuild'")
+    started = time.time()
+    log = log or (lambda _msg: None)
+    stats = RunStats(
+        mode=mode,
+        collection=cfg.collection,
+        graph_collection=cfg.graph_collection,
+        provider=cfg.provider,
+        model=cfg.model,
+        dim=cfg.dim,
+        chunker_version=cfg.chunker_version,
+    )
+
+    # 1. Read the estate. Curated content is validated first (loader does it),
+    #    so a broken content tree stops the run before Gitea is touched.
+    harvest = estate_module.harvest(cfg, log=log)
+    documents = harvest.documents
+    stats.documents = len(documents)
+    stats.rejections = len(harvest.rejections)
+    stats.images = len(harvest.images)
+    stats.rejection_summary = harvest.rejection_summary()
+    stats.warnings = list(harvest.warnings)
+    stats.platforms = {
+        name: {"documents": row.get("documents", 0)} for name, row in harvest.sources.items()
+    }
+    log(f"estate: {len(documents)} document(s), {stats.rejections} rejection(s), {stats.images} image reference(s)")
+
+    provider = embeddings.build_provider(cfg)
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    if not qdrant.health():
+        raise PipelineError(
+            f"Qdrant is not answering at {cfg.qdrant_url}. Start it "
+            "(deploy/qdrant.yml) or point RESUME_QDRANT_URL at a running instance."
+        )
+
+    with store_module.Store(cfg.db_path) as store:
+        contract = _contract(cfg)
+        mode = _check_contract(store, qdrant, cfg, contract, mode)
+        _ensure_collection(qdrant, cfg.collection, cfg)
+
+        manifest = store.manifest()
+        changed, removed = _classify(documents, manifest, cfg, mode)
+        stats.unchanged = len(documents) - len(changed)
+        stats.removed = len(removed)
+        log(
+            f"incremental plan: {len(changed)} to embed, "
+            f"{stats.unchanged} unchanged, {len(removed)} removed"
+        )
+
+        for path in removed:
+            stats.chunks_deleted += store.delete_document(path)
+            qdrant.delete_by_source(cfg.collection, path)
+            log(f"removed {path}")
+
+        # 2. Build and validate every payload before any write.
+        prepared = []  # [(doc, payloads)]
+        for doc in changed:
+            prepared.append((doc, _prepare(doc, cfg)))
+        all_payloads = [payload for _, rows in prepared for payload in rows]
+
+        # 3. Embed in one flat, batched pass. This is the number the DoD uses to
+        #    prove incremental mode: it is proportional to changed chunks only.
+        vectors = _embed_all(provider, [payload["text"] for payload in all_payloads])
+        if len(vectors) != len(all_payloads):
+            raise PipelineError(
+                f"provider returned {len(vectors)} vectors for {len(all_payloads)} chunks"
+            )
+
+        cursor = 0
+        platform_changes = stats.platforms
+        for doc, rows in prepared:
+            points = []
+            for payload in rows:
+                vector = vectors[cursor]
+                cursor += 1
+                if len(vector) != cfg.dim:
+                    raise PipelineError(
+                        f"{doc.path}: vector dimension {len(vector)} != configured {cfg.dim}"
+                    )
+                points.append({"id": payload["id"], "vector": vector, "payload": payload})
+
+            # Remove what this document used to have, then write the new rows and
+            # points. A crash between the two leaves an index that reports as
+            # incomplete, never one that claims to be whole.
+            stats.chunks_deleted += store.delete_document(doc.path)
+            qdrant.delete_by_source(cfg.collection, doc.path)
+            stats.points_sent += qdrant.upsert(cfg.collection, points)
+            store.write_document(doc, rows)
+            stats.chunks_written += len(rows)
+
+            bucket = platform_changes.setdefault(doc.platform, {"documents": 0})
+            key = "updated" if doc.path in manifest else "added"
+            bucket[key] = bucket.get(key, 0) + 1
+            if doc.path in manifest:
+                stats.updated += 1
+                log(f"updated {doc.path} ({len(points)} chunk(s))")
+            else:
+                stats.added += 1
+                log(f"added {doc.path} ({len(points)} chunk(s))")
+
+        provider_stats = provider.stats()
+        stats.embed_calls = provider_stats["calls"]
+        stats.texts_embedded = provider_stats["texts"]
+
+        # 4. Graph: its own collection, rebuilt only when graph.json changed.
+        graph_stats = _sync_graph(qdrant, store, cfg, harvest, contract, mode, provider, log)
+        stats.graph_nodes = graph_stats["nodes"]
+        stats.graph_edges = graph_stats["edges"]
+        stats.graph_points = graph_stats["points"]
+        stats.graph_rebuilt = graph_stats["rebuilt"]
+        stats.notes.extend(graph_stats["notes"])
+
+        # 5. Consistency and the credential scan. Both are checks, not hopes.
+        stats.store_chunks = store.chunk_count()
+        stats.collection_points = qdrant.point_count(cfg.collection)
+        if stats.collection_points != stats.store_chunks:
+            raise PipelineError(
+                "index is inconsistent after the run: "
+                f"{stats.collection_points} point(s) in Qdrant but "
+                f"{stats.store_chunks} chunk(s) in SQLite. Run `python3 -m corpus.cli verify`."
+            )
+        _scan_pool(store)
+        _scan_graph(store)
+
+        store.set_meta("collection", contract)
+        store.set_meta("sources", harvest.sources)
+        store.set_meta("rejections", harvest.rejections[:200])
+        store.set_meta("images", harvest.images[:500])
+        store.set_meta(
+            "last_run",
+            {
+                **stats.to_dict(),
+                "finished_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
+            },
+        )
+        _write_collection_file(cfg, store, stats, harvest)
+
+    stats.duration_s = round(time.time() - started, 3)
+    return stats
+
+
+# -- planning -------------------------------------------------------------
+
+
+def _contract(cfg) -> dict:
+    """What this collection *is*. Recorded so a change is a migration, not drift."""
+    return {
+        "provider": cfg.provider,
+        "model": cfg.model,
+        "dim": cfg.dim,
+        "chunker_version": cfg.chunker_version,
+        "distance": "Cosine",
+    }
+
+
+def _check_contract(store, qdrant, cfg, contract: dict, mode: str) -> str:
+    if mode == "rebuild":
+        qdrant.drop_collection(cfg.collection)
+        qdrant.drop_collection(cfg.graph_collection)
+        if store.schema_stale:
+            store.rebuild_schema()
+        else:
+            store.reset()
+        store.set_meta("schema_version", store_module.SCHEMA_VERSION)
+        return mode
+
+    if store.schema_stale:
+        raise PipelineError(
+            "the corpus database was written by an older payload shape and has no "
+            "provenance columns. It is derived data: run `python3 -m corpus.cli rebuild`."
+        )
+
+    stored = store.get_meta("collection")
+    if stored is None:
+        if qdrant.collection_exists(cfg.collection) and qdrant.point_count(cfg.collection) > 0:
+            raise PipelineError(
+                f"collection {cfg.collection!r} holds points but this store has no "
+                "collection contract (was data/corpus/corpus.db deleted?). "
+                "Confirm you want to overwrite it: python3 -m corpus.cli rebuild"
+            )
+        return mode
+    if stored != contract:
+        raise PipelineError(
+            "the collection contract does not match this configuration — refusing to "
+            "mix vectors in one collection (docs/phase-4-corpus.md §4).\n"
+            f"  stored: {json.dumps(stored, sort_keys=True)}\n"
+            f"  config: {json.dumps(contract, sort_keys=True)}\n"
+            "A provider or chunker change is a governed migration: run "
+            "`python3 -m corpus.cli rebuild` once Sam has approved the change."
+        )
+    return mode
+
+
+def _ensure_collection(qdrant: Qdrant, name: str, cfg) -> None:
+    if not qdrant.collection_exists(name):
+        qdrant.create_collection(name, cfg.dim, vectors_on_disk=cfg.vectors_on_disk)
+        return
+    info = qdrant.collection_info(name)
+    vectors = ((info.get("config") or {}).get("params") or {}).get("vectors")
+    size = None
+    if isinstance(vectors, dict):
+        size = vectors.get("size")
+    if size is not None and int(size) != int(cfg.dim):
+        raise PipelineError(
+            f"collection {name!r} has vectors of size {size} but this run "
+            f"is configured for {cfg.dim}. That is a governed migration: rebuild."
+        )
+
+
+def _classify(documents: list, manifest: dict, cfg, mode: str):
+    """Which documents need work, and which are gone."""
+    changed = []
+    on_disk = set()
+    for doc in documents:
+        on_disk.add(doc.path)
+        if mode == "rebuild":
+            changed.append(doc)
+            continue
+        entry = manifest.get(doc.path)
+        if entry is None:
+            changed.append(doc)
+            continue
+        if (
+            entry["content_hash"] != doc.digest
+            or entry["embedding_model"] != cfg.model
+            or int(entry["embedding_dim"]) != int(cfg.dim)
+            or entry["chunker_version"] != cfg.chunker_version
+        ):
+            changed.append(doc)
+    removed = sorted(path for path in manifest if path not in on_disk)
+    return changed, removed
+
+
+def _prepare(doc, cfg) -> list:
+    """Chunk one document and build validated payloads. No network, no writes."""
+    try:
+        if getattr(doc, "kind", "markdown") == "code":
+            chunks = chunker.chunk_code(doc.body)
+        else:
+            chunks = chunker.chunk_document(doc.body)
+    except chunker.ChunkError as exc:
+        raise PipelineError(f"{doc.path}: {exc}") from exc
+    if not chunks:
+        raise PipelineError(f"{doc.path}: produced no chunks (empty body?)")
+    doc.chunk_count = len(chunks)
+
+    out = []
+    for chunk in chunks:
+        payload = payloads.build_payload(doc, chunk, cfg)
+        payload["id"] = chunk_id(doc.path, chunk.index, doc.digest)
+        try:
+            payloads.validate_payload(payload, where=f"{doc.path}#chunk-{chunk.index}")
+        except payloads.PayloadError as exc:
+            raise PipelineError(str(exc)) from exc
+        out.append(payload)
+    return out
+
+
+def _embed_all(provider, texts: list) -> list:
+    out = []
+    for start in range(0, len(texts), EMBED_BATCH):
+        out.extend(provider.embed(texts[start : start + EMBED_BATCH]))
+    return out
+
+
+# -- graph ----------------------------------------------------------------
+
+
+def _sync_graph(qdrant, store, cfg, harvest, contract, mode, provider, log) -> dict:
+    graph = harvest.graph
+    stored = store.get_meta("graph") or {}
+    notes = []
+    if graph is None:
+        if mode == "rebuild":
+            qdrant.drop_collection(cfg.graph_collection)
+            store.write_graph([], [])
+            store.set_meta("graph", {})
+            notes.append("graph: no graph.json was available; the graph collection is empty")
+            return {"nodes": 0, "edges": 0, "points": 0, "rebuilt": True, "notes": notes}
+        counts = store.graph_counts()
+        notes.append(
+            f"graph: graph.json was not available this run; kept the previous graph "
+            f"({counts['nodes']} nodes, {counts['edges']} edges)"
+        )
+        return {**counts, "points": counts["nodes"] + counts["edges"], "rebuilt": False, "notes": notes}
+
+    graph_contract = {
+        "digest": graph.digest,
+        "source_path": graph.source_path,
+        "provider": contract["provider"],
+        "model": contract["model"],
+        "dim": contract["dim"],
+    }
+    rebuild = mode == "rebuild" or stored != graph_contract
+    if not rebuild:
+        counts = store.graph_counts()
+        return {
+            **counts,
+            "points": counts["nodes"] + counts["edges"],
+            "rebuilt": False,
+            "notes": [f"graph: unchanged ({graph.digest[:12]}), not re-embedded"],
+        }
+
+    if not qdrant.collection_exists(cfg.graph_collection):
+        _ensure_collection(qdrant, cfg.graph_collection, cfg)
+    else:
+        qdrant.drop_collection(cfg.graph_collection)
+        qdrant.create_collection(cfg.graph_collection, cfg.dim, vectors_on_disk=cfg.vectors_on_disk)
+
+    rows = [("node", row) for row in graph.nodes] + [("edge", row) for row in graph.edges]
+    payloads_for_graph = [graph_module.point_payload(row, kind) for kind, row in rows]
+    vectors = _embed_all(provider, [payload["text"] for payload in payloads_for_graph])
+    points = [
+        {"id": row["point_id"], "vector": vector, "payload": payload}
+        for (_, row), vector, payload in zip(rows, vectors, payloads_for_graph)
+    ]
+    sent = qdrant.upsert(cfg.graph_collection, points)
+    store.write_graph(graph.nodes, graph.edges)
+    store.set_meta("graph", graph_contract)
+    log(f"graph: rebuilt {len(graph.nodes)} node(s) and {len(graph.edges)} edge(s), {sent} point(s)")
+    return {
+        "nodes": len(graph.nodes),
+        "edges": len(graph.edges),
+        "points": sent,
+        "rebuilt": True,
+        "notes": notes,
+    }
+
+
+# -- checks ---------------------------------------------------------------
+
+
+def _scan_pool(store) -> None:
+    """Search the finished pool for credential-shaped content. Fail, do not warn."""
+    hits = []
+    for source_path, chunk_index, text in store.pool_texts():
+        for kind, line, _excerpt in exclusions.scan_text(text):
+            hits.append((source_path, chunk_index, kind, line))
+    if hits:
+        detail = ", ".join(f"{path}#chunk-{index} ({kind} line {line})" for path, index, kind, line in hits[:10])
+        raise PipelineError(
+            f"credential-shaped content reached the pool in {len(hits)} place(s): {detail}. "
+            "The run is failed rather than published; fix the source or the exclusion rule."
+        )
+
+
+def _scan_graph(store) -> None:
+    rows = store.db.execute("SELECT id, text FROM graph_nodes UNION ALL SELECT id, text FROM graph_edges")
+    hits = []
+    for row in rows:
+        for kind, line, _excerpt in exclusions.scan_text(row["text"]):
+            hits.append((row["id"], kind, line))
+    if hits:
+        detail = ", ".join(f"{node} ({kind} line {line})" for node, kind, line in hits[:10])
+        raise PipelineError(f"credential-shaped content reached the graph in {len(hits)} place(s): {detail}")
+
+
+def _write_collection_file(cfg, store, stats: RunStats, harvest) -> None:
+    """Mirror the contract and the last run to data/corpus/collection.json."""
+    body = {
+        "collection": cfg.collection,
+        "contract": _contract(cfg),
+        "qdrant_url": cfg.qdrant_url,
+        "chunks": store.chunk_count(),
+        "sources": store.source_count(),
+        "graph_collection": cfg.graph_collection,
+        "graph": store.graph_counts(),
+        "platforms": harvest.sources,
+        "rejections": harvest.rejection_summary(),
+        "images": len(harvest.images),
+        "last_run": store.get_meta("last_run"),
+    }
+    with open(cfg.meta_path, "w", encoding="utf-8") as handle:
+        json.dump(body, handle, indent=2, sort_keys=True)
+        handle.write("\n")
+
+
+# -- verification and demonstration ---------------------------------------
+
+
+def verify(cfg, log=None) -> list:
+    """Return a list of problems. Empty means the index agrees with the sources."""
+    log = log or (lambda _msg: None)
+    problems = []
+    harvest = estate_module.harvest(cfg)
+    documents = harvest.documents
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    if not qdrant.health():
+        return [f"Qdrant is not answering at {cfg.qdrant_url}"]
+
+    contract = _contract(cfg)
+    with store_module.Store(cfg.db_path) as store:
+        stored = store.get_meta("collection")
+        if stored != contract:
+            problems.append(f"collection contract mismatch: stored={stored} config={contract}")
+
+        if not qdrant.collection_exists(cfg.collection):
+            problems.append(f"collection {cfg.collection!r} does not exist")
+            return problems
+
+        info = qdrant.collection_info(cfg.collection)
+        size = (((info.get("config") or {}).get("params") or {}).get("vectors") or {}).get("size")
+        if size is not None and int(size) != cfg.dim:
+            problems.append(f"collection vector size {size} != configured {cfg.dim}")
+
+        points_by_source = qdrant.scroll_sources(cfg.collection)
+        rows = store.chunk_payloads()
+        by_source = {}
+        for row in rows:
+            by_source.setdefault(row["source_path"], []).append(row)
+            try:
+                payloads.validate_payload(row, where=f"{row['source_path']}#chunk-{row['chunk_index']}")
+            except payloads.PayloadError as exc:
+                problems.append(str(exc))
+
+        on_disk = {doc.path: doc for doc in documents}
+        for doc in documents:
+            payload_rows = by_source.get(doc.path, [])
+            if not payload_rows:
+                problems.append(f"{doc.path}: no chunks in the index (run sync)")
+                continue
+            if payload_rows[0]["content_hash"] != doc.digest:
+                problems.append(f"{doc.path}: source changed since it was embedded")
+            expected = len(payload_rows)
+            actual = points_by_source.get(doc.path, 0)
+            if actual != expected:
+                problems.append(
+                    f"{doc.path}: {expected} chunk row(s) but {actual} vector(s) in Qdrant"
+                )
+
+        for source in sorted(set(by_source) - set(on_disk)):
+            problems.append(f"{source}: indexed but no longer in content/ or the estate (run sync)")
+
+        total_points = sum(points_by_source.values())
+        if total_points != store.chunk_count():
+            problems.append(
+                f"{total_points} vector(s) in Qdrant but {store.chunk_count()} chunk(s) in SQLite"
+            )
+
+        # Graph: every row must resolve to a note, and Qdrant must agree.
+        counts = store.graph_counts()
+        graph_meta = store.get_meta("graph") or {}
+        if harvest.graph is not None:
+            if not graph_meta:
+                problems.append("graph: graph.json is available but the graph was never indexed")
+            if qdrant.collection_exists(cfg.graph_collection):
+                graph_points = qdrant.point_count(cfg.graph_collection)
+                if graph_points != counts["nodes"] + counts["edges"]:
+                    problems.append(
+                        f"graph: {graph_points} point(s) in Qdrant but "
+                        f"{counts['nodes'] + counts['edges']} row(s) in SQLite"
+                    )
+            elif counts["nodes"] or counts["edges"]:
+                problems.append(f"graph: collection {cfg.graph_collection!r} does not exist")
+            for row in store.db.execute("SELECT id, source_note, origin_url FROM graph_nodes LIMIT 5"):
+                if not row["source_note"] or not row["origin_url"]:
+                    problems.append(f"graph node {row['id']}: no source note or origin")
+        missing_notes = [
+            row["id"]
+            for row in store.db.execute(
+                "SELECT id FROM graph_nodes WHERE source_note = '' OR origin_url = '' LIMIT 10"
+            )
+        ]
+        if missing_notes:
+            problems.append(f"graph: {len(missing_notes)} node(s) carry no source note")
+
+        log(
+            f"verify: {store.chunk_count()} chunk(s) across {store.source_count()} source(s), "
+            f"{total_points} vector(s) in Qdrant, {counts['nodes']} node(s), {counts['edges']} edge(s)"
+        )
+    for warning in harvest.warnings:
+        problems.append(f"estate warning: {warning}")
+    return problems
+
+
+def demo_area_filter(cfg, query: str, area: str = None, limit: int = 5) -> dict:
+    """Same query with and without an `area` filter — the DoD's filter evidence."""
+    provider = embeddings.build_provider(cfg)
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    vector = provider.embed([query])[0]
+    unfiltered = qdrant.search(cfg.collection, vector, limit=limit)
+    filtered = qdrant.search(cfg.collection, vector, limit=limit, area=area) if area else []
+    return {
+        "query": query,
+        "area": area,
+        "unfiltered": [_hit(row) for row in unfiltered],
+        "filtered": [_hit(row) for row in filtered],
+    }
+
+
+def _hit(row: dict) -> dict:
+    payload = row.get("payload") or {}
+    return {
+        "score": round(row["score"], 4),
+        "area": payload.get("area"),
+        "platform": payload.get("platform"),
+        "tier": payload.get("tier"),
+        "confidence": payload.get("confidence"),
+        "title": payload.get("title"),
+        "url": payload.get("url"),
+        "source_path": payload.get("source_path"),
+        "chunk_index": payload.get("chunk_index"),
+        "preview": (payload.get("text") or "").replace("\n", " ")[:120],
+    }
+
+
+def hybrid_search(cfg, query: str, area: str = None, limit: int = 5, platform: str = None, tier: int = None) -> list:
+    """Vectors + FTS5 fused with reciprocal rank fusion (decision D10).
+
+    Phase 5 owns the query graph; this is here because Phase 4 is the phase that
+    has to prove the two halves line up — same ids, same metadata, one ranking.
+    """
+    provider = embeddings.build_provider(cfg)
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    vector = provider.embed([query])[0]
+    ranked = {}
+    for rank, hit in enumerate(
+        qdrant.search(cfg.collection, vector, limit=limit * 4, area=area, platform=platform, tier=tier)
+    ):
+        entry = ranked.setdefault(hit["id"], {"id": hit["id"], "payload": hit["payload"], "score": 0.0})
+        entry["score"] += 1.0 / (60 + rank)
+    with store_module.Store(cfg.db_path) as store:
+        for rank, hit in enumerate(
+            store.keyword_search(query, area=area, platform=platform, tier=tier, limit=limit * 4)
+        ):
+            entry = ranked.setdefault(hit["id"], {"id": hit["id"], "payload": {}, "score": 0.0})
+            entry.setdefault("keyword", hit["snippet"])
+            entry["score"] += 1.0 / (60 + rank)
+    ordered = sorted(ranked.values(), key=lambda row: row["score"], reverse=True)[:limit]
+    for row in ordered:
+        payload = row.get("payload") or {}
+        row["title"] = payload.get("title")
+        row["url"] = payload.get("url")
+    return ordered
+
+
+def scan_pool(cfg) -> dict:
+    """Run the credential scan without writing anything. Used by `make corpus-scan`."""
+    with store_module.Store(cfg.db_path) as store:
+        hits = []
+        texts = store.pool_texts()
+        for source_path, chunk_index, text in texts:
+            for kind, line, _excerpt in exclusions.scan_text(text):
+                hits.append(
+                    {"source_path": source_path, "chunk": chunk_index, "kind": kind, "line": line}
+                )
+        try:
+            _scan_graph(store)
+        except PipelineError as exc:
+            hits.append({"source_path": "graph", "kind": "credential-shaped content", "detail": str(exc)})
+        chunks = len(texts)
+    return {"ok": not hits, "hits": hits, "chunks": chunks}

corpus/qdrant.py

@@ -0,0 +1,277 @@
+"""Qdrant REST client — standard library only.
+
+No `qdrant-client` dependency: the four calls this pipeline needs are HTTP
+requests, and AGENTS.md §4.1 freezes the stack. `_request` is the single
+transport hook, which is what the tests replace to run the whole pipeline
+without a container.
+
+Qdrant is bound to `127.0.0.1:6333` on the runtime host (`.13`) and is never
+publicly reachable; Caddy is the only public listener.
+"""
+
+from __future__ import annotations
+
+import http.client
+import json
+import time
+import urllib.error
+import urllib.request
+
+DISTANCE_COSINE = "Cosine"
+
+# The runtime host (.13) is memory-bound: 2.3 GiB free and 8 GiB of swap already
+# in use, so Qdrant gets a 400 m cap and must live inside it
+# (reference/stack.md). Measured on 2026-10-26: the *default* HNSW and payload
+# settings are OOM-killed at the cap between 5,000 and 20,000 chunks of 1024
+# dimensions. Keeping the HNSW graph and the payloads on disk, and using a
+# smaller graph (m=8, ef_construct=64), fits the cap because the kernel can
+# reclaim those page-cache pages instead of killing the process. Vectors stay in
+# RAM because they are the part search reads on every query.
+#
+# The trade is build time and a little search latency; the corpus is ~50 chunks
+# today and the public site has no query burst. See docs/phase-4-notes.md §7.
+MEMORY_FRUGAL_DEFAULTS = {
+    "on_disk_payload": True,
+    "hnsw_m": 8,
+    "hnsw_ef_construct": 64,
+    "vectors_on_disk": False,
+}
+
+# Segment behaviour under a cap. `memmap_threshold` makes Qdrant mmap segments
+# this size and larger (page cache the kernel can reclaim instead of OOM), and
+# `indexing_threshold` is raised so a bulk ingest does not build the HNSW graph
+# while it is still writing segments. Measured: 20,000 chunks at 1024
+# dimensions peaked at 296 MiB during ingest and settled to 96 MiB.
+MEMORY_FRUGAL_OPTIMIZERS = {
+    "default_segment_number": 2,
+    "memmap_threshold": 5000,
+    "indexing_threshold": 20000,
+}
+
+
+class QdrantError(RuntimeError):
+    """A Qdrant call failed, or Qdrant is not reachable."""
+
+
+class Qdrant:
+    def __init__(self, url: str = "http://127.0.0.1:6333", timeout: float = 30.0, attempts: int = 3) -> None:
+        self.url = url.rstrip("/")
+        self.timeout = timeout
+        self.attempts = max(1, attempts)
+
+    # -- transport --------------------------------------------------------
+
+    def _request(self, method: str, path: str, body=None) -> dict:
+        """One HTTP call with retry/backoff. The only place that touches the network."""
+        data = None if body is None else json.dumps(body).encode("utf-8")
+        last_error = None
+        for attempt in range(self.attempts):
+            request = urllib.request.Request(
+                f"{self.url}{path}",
+                data=data,
+                method=method,
+                headers={"Content-Type": "application/json"} if data else {},
+            )
+            try:
+                with urllib.request.urlopen(request, timeout=self.timeout) as response:
+                    raw = response.read().decode("utf-8")
+                return json.loads(raw) if raw.strip() else {}
+            except urllib.error.HTTPError as exc:
+                detail = exc.read().decode("utf-8", "replace")[:400]
+                exc.close()
+                if exc.code == 404:
+                    raise QdrantNotFound(f"{method} {path}: {detail}") from None
+                if exc.code != 429 and exc.code < 500:
+                    raise QdrantError(f"{method} {path} failed: HTTP {exc.code} {detail}") from None
+                last_error = f"HTTP {exc.code} {detail}"
+            except (urllib.error.URLError, TimeoutError, http.client.HTTPException, ConnectionError) as exc:
+                # RemoteDisconnected and friends arrive here when the server dies
+                # mid-request (an OOM kill looks exactly like this).
+                last_error = f"cannot reach {self.url}: {getattr(exc, 'reason', exc)}"
+            if attempt + 1 < self.attempts:
+                time.sleep(0.5 * (2 ** attempt))
+        raise QdrantError(f"{method} {path} failed after {self.attempts} attempts: {last_error}")
+
+    # -- cluster ----------------------------------------------------------
+
+    def health(self) -> bool:
+        """True when Qdrant answers. Never raises."""
+        try:
+            self._request("GET", "/")
+            return True
+        except (QdrantError, QdrantNotFound):
+            return False
+
+    def version(self) -> str:
+        response = self._request("GET", "/")
+        return str(response.get("version", "unknown"))
+
+    # -- collections ------------------------------------------------------
+
+    def collection_exists(self, name: str) -> bool:
+        try:
+            self._request("GET", f"/collections/{name}")
+            return True
+        except QdrantNotFound:
+            return False
+
+    def create_collection(
+        self,
+        name: str,
+        dim: int,
+        distance: str = DISTANCE_COSINE,
+        on_disk_payload: bool = None,
+        hnsw_m: int = None,
+        hnsw_ef_construct: int = None,
+        vectors_on_disk: bool = None,
+    ) -> None:
+        """Create a collection sized for a memory-bound host (see MEMORY_FRUGAL_DEFAULTS)."""
+        defaults = MEMORY_FRUGAL_DEFAULTS
+        body = {
+            "vectors": {
+                "size": int(dim),
+                "distance": distance,
+                "on_disk": bool(
+                    defaults["vectors_on_disk"] if vectors_on_disk is None else vectors_on_disk
+                ),
+            },
+            "hnsw_config": {
+                "m": int(defaults["hnsw_m"] if hnsw_m is None else hnsw_m),
+                "ef_construct": int(
+                    defaults["hnsw_ef_construct"] if hnsw_ef_construct is None else hnsw_ef_construct
+                ),
+                "on_disk": True,
+            },
+            "on_disk_payload": bool(
+                defaults["on_disk_payload"] if on_disk_payload is None else on_disk_payload
+            ),
+            "optimizers_config": dict(MEMORY_FRUGAL_OPTIMIZERS),
+        }
+        self._request("PUT", f"/collections/{name}", body)
+
+    def drop_collection(self, name: str) -> None:
+        try:
+            self._request("DELETE", f"/collections/{name}")
+        except QdrantNotFound:
+            pass
+
+    def collection_info(self, name: str) -> dict:
+        response = self._request("GET", f"/collections/{name}")
+        return dict(response.get("result") or {})
+
+    def point_count(self, name: str) -> int:
+        response = self._request("POST", f"/collections/{name}/points/count", {"exact": True})
+        return int((response.get("result") or {}).get("count", 0))
+
+    # -- points -----------------------------------------------------------
+
+    def upsert(self, name: str, points: list, batch: int = 256) -> int:
+        """Upsert points in batches. Returns the number of points sent."""
+        sent = 0
+        for start in range(0, len(points), max(1, batch)):
+            chunk = points[start : start + max(1, batch)]
+            self._request(
+                "PUT",
+                f"/collections/{name}/points?wait=true",
+                {"points": chunk},
+            )
+            sent += len(chunk)
+        return sent
+
+    def delete_by_source(self, name: str, source_path: str) -> None:
+        """Delete every point that came from one document.
+
+        Re-chunking a document can produce fewer chunks than before, so the old
+        points are removed rather than overwritten. Deterministic ids make
+        duplicates impossible; this makes orphans impossible.
+        """
+        self._request(
+            "POST",
+            f"/collections/{name}/points/delete?wait=true",
+            {"filter": {"must": [{"key": "source_path", "match": {"value": source_path}}]}},
+        )
+
+    def search(
+        self,
+        name: str,
+        vector: list,
+        limit: int = 5,
+        area: str = None,
+        confidence: str = None,
+        platform: str = None,
+        tier: int = None,
+        with_payload: bool = True,
+    ) -> list:
+        """Nearest neighbours, optionally filtered *before* the search."""
+        body = {
+            "query": [float(value) for value in vector],
+            "limit": int(limit),
+            "with_payload": bool(with_payload),
+        }
+        filters = _filter_conditions(area, confidence, platform, tier)
+        if filters:
+            body["filter"] = {"must": filters}
+        try:
+            response = self._request("POST", f"/collections/{name}/points/query", body)
+            points = ((response.get("result") or {}).get("points")) or []
+        except QdrantNotFound:
+            # Older Qdrant exposes the same operation as /points/search.
+            response = self._request("POST", f"/collections/{name}/points/search", body)
+            points = response.get("result") or []
+        return [
+            {
+                "id": point.get("id"),
+                "score": float(point.get("score") or 0.0),
+                "payload": point.get("payload") or {},
+            }
+            for point in points
+        ]
+
+    def scroll_sources(self, name: str, limit: int = 1000) -> dict:
+        """Map source_path -> number of vectors, for consistency checks."""
+        counts = {}
+        offset = None
+        while True:
+            body = {"limit": limit, "with_payload": ["source_path"], "with_vector": False}
+            if offset is not None:
+                body["offset"] = offset
+            response = self._request("POST", f"/collections/{name}/points/scroll", body)
+            result = response.get("result") or {}
+            points = result.get("points") or []
+            for point in points:
+                source = (point.get("payload") or {}).get("source_path", "<missing>")
+                counts[source] = counts.get(source, 0) + 1
+            offset = result.get("next_page_offset")
+            if not points or offset is None:
+                return counts
+
+    def metrics(self) -> str:
+        """Prometheus text from Qdrant's /metrics route (used by the memory measurement).
+
+        Returns "" when the route is not exposed, so callers can fall back to
+        reading the container's own memory counters.
+        """
+        try:
+            with urllib.request.urlopen(f"{self.url}/metrics", timeout=self.timeout) as response:
+                return response.read().decode("utf-8", "replace")
+        except (urllib.error.URLError, urllib.error.HTTPError, TimeoutError):
+            return ""
+
+
+class QdrantNotFound(QdrantError):
+    """A 404 from Qdrant: the collection or route does not exist."""
+
+
+def _filter_conditions(
+    area: str = None, confidence: str = None, platform: str = None, tier: int = None
+) -> list:
+    conditions = []
+    if area:
+        conditions.append({"key": "area", "match": {"value": area}})
+    if confidence:
+        conditions.append({"key": "confidence", "match": {"value": confidence}})
+    if platform:
+        conditions.append({"key": "platform", "match": {"value": platform}})
+    if tier is not None:
+        conditions.append({"key": "tier", "match": {"value": int(tier)}})
+    return conditions

corpus/sources/estate.py

@@ -0,0 +1,848 @@
+"""Harvest the estate into one set of documents.
+
+Four sources, one shape (`loader.Document`), one tier rule:
+
+| Platform | Tier | Reached by |
+|---|---|---|
+| `curated` (`content/`) | 1 | the local repository, already validated |
+| `gitea` | 2 | the Gitea API |
+| `obsidian` | 2 | the same Gitea API, one repository |
+| `graphify` | 2 | `graph.json` committed to the vault repository |
+
+`content/` is loaded first and is never displaced: the query layer uses `tier`
+to prefer a curated claim over harvested text that disagrees with it.
+
+Every refused file is recorded with the rule that refused it. The DoD asks for
+this explicitly ("the run reports the rejection"), and it is the difference
+between an exclusion list that works and one nobody can audit.
+"""
+
+from __future__ import annotations
+
+import datetime
+import hashlib
+import json
+import os
+import re
+
+from .. import areas, exclusions, globs, guides, loader
+from . import graph as graph_module
+from .gitea import GiteaClient, GiteaError, GiteaNotFound, file_url, repo_url
+
+PLATFORM_ORDER = ("curated", "gitea", "obsidian", "graphify")
+
+
+class EstateError(RuntimeError):
+    """The estate could not be read. The run stops rather than indexing half of it."""
+
+
+class Harvest:
+    def __init__(self) -> None:
+        self.documents: list = []
+        self.rejections: list = []
+        self.images: list = []
+        self.warnings: list = []
+        self.sources: dict = {}
+        self.graph = None  # graph_module.GraphResult | None
+        self.offline = False
+
+    # -- reporting --------------------------------------------------------
+
+    def add_rejection(self, rejection) -> None:
+        self.rejections.append(rejection.to_dict() if hasattr(rejection, "to_dict") else rejection)
+
+    def rejection_summary(self) -> dict:
+        counts = {}
+        for row in self.rejections:
+            reason = str(row.get("reason") or "unknown")
+            # Keep the summary readable: 200 distinct oversize byte counts are
+            # one finding, not two hundred.
+            if reason.startswith("oversize"):
+                reason = "oversize"
+            elif reason.startswith("binary"):
+                reason = "binary"
+            key = f"{row.get('platform')}/{reason}"
+            counts[key] = counts.get(key, 0) + 1
+        return dict(sorted(counts.items()))
+
+    def to_dict(self) -> dict:
+        return {
+            "documents": len(self.documents),
+            "images": len(self.images),
+            "rejections": len(self.rejections),
+            "rejection_summary": self.rejection_summary(),
+            "warnings": list(self.warnings),
+            "sources": self.sources,
+            "graph": self.graph.counts() if self.graph else None,
+            "offline": self.offline,
+        }
+
+
+def harvest(cfg, log=None) -> Harvest:
+    """Read every enabled platform. Curated content is always read."""
+    log = log or (lambda _msg: None)
+    result = Harvest()
+    result.offline = bool(cfg.offline)
+
+    curated = loader.load_documents(cfg)
+    result.documents.extend(curated)
+    result.sources["curated"] = _source_row(
+        "curated", enabled=True, documents=curated, extra={"path": "content/"}
+    )
+    log(f"curated: {len(curated)} document(s) from content/")
+
+    if cfg.offline:
+        # Not a warning: offline is a deliberate mode, not a failure.
+        log("offline mode: only content/ was read (RESUME_OFFLINE=1)")
+        return result
+
+    all_guides = guides.load_guides(cfg.root)
+    selected = {
+        name: guide
+        for name, guide in all_guides.items()
+        if guide.enabled and name in cfg.platforms
+    }
+    missing = [name for name in cfg.platforms if name not in all_guides]
+    if missing:
+        result.warnings.append(f"platform(s) {missing} have no guide and were skipped")
+
+    needs_gitea = {"gitea", "obsidian", "graphify"} & set(selected)
+    if not needs_gitea:
+        return result
+
+    base = cfg.gitea_url or selected.get("gitea", all_guides.get("gitea")).base
+    client = GiteaClient(base, cfg.gitea_token, timeout=cfg.timeout)
+    try:
+        repos = client.repos()
+    except GiteaError as exc:
+        raise EstateError(f"cannot list Gitea repositories: {exc}") from exc
+    log(f"gitea: {len(repos)} repositories visible")
+
+    vault_repo = _vault_repo(selected)
+    vault_tree = None
+    if vault_repo:
+        try:
+            vault_tree = client.tree(vault_repo, "main")
+        except GiteaNotFound:
+            try:
+                vault_tree = client.tree(vault_repo, "master")
+            except GiteaError as exc:
+                raise EstateError(f"cannot read {vault_repo}: {exc}") from exc
+        except GiteaError as exc:
+            raise EstateError(f"cannot read {vault_repo}: {exc}") from exc
+
+    if "gitea" in selected:
+        _harvest_repos(result, client, selected["gitea"], repos, base, log)
+    if "obsidian" in selected:
+        _harvest_obsidian(result, client, selected["obsidian"], repos, vault_repo, vault_tree, cfg, base, log)
+    if "graphify" in selected:
+        _harvest_graph(result, client, selected["graphify"], repos, vault_repo, vault_tree, cfg, base, log)
+
+    return result
+
+
+# -- gitea repositories ---------------------------------------------------
+
+
+def _harvest_repos(result: Harvest, client, guide, repos, base_url, log) -> None:
+    excluded = set(_as_list((guide.exclude or {}).get("repos")))
+    include = (guide.include or {}).get("repos") or "all"
+    wanted = None if include == "all" or _as_list(include) == [] else set(_as_list(include))
+
+    documents = []
+    seen_repos = set()
+    for repo in repos:
+        full_name = repo.get("full_name") or ""
+        if not full_name or full_name in excluded:
+            continue
+        if wanted is not None and full_name not in wanted:
+            continue
+        seen_repos.add(full_name)
+        try:
+            documents.extend(_harvest_one_repo(result, client, guide, repo, base_url))
+        except GiteaError as exc:
+            result.warnings.append(f"gitea/{full_name}: {exc}")
+            continue
+        if len(documents) >= guide.limit("max_documents"):
+            result.warnings.append(
+                f"gitea: document limit {guide.limit('max_documents')} reached; "
+                "remaining repositories were skipped"
+            )
+            break
+
+    result.documents.extend(documents)
+    result.sources["gitea"] = _source_row(
+        "gitea", enabled=True, documents=documents,
+        extra={"repos": len(seen_repos), "base": base_url},
+    )
+    log(f"gitea: {len(documents)} document(s), {len(result.rejections)} rejection(s) so far")
+
+
+def _harvest_one_repo(result: Harvest, client, guide, repo, base_url) -> list:
+    full_name = repo["full_name"]
+    branch = repo.get("default_branch") or "main"
+    updated = _timestamp(repo.get("updated_at"))
+    created = _timestamp(repo.get("created_at"))
+    area = areas.area_for_repo(repo, guide.fits)
+    topics = [str(topic) for topic in (repo.get("topics") or [])]
+    description = str(repo.get("description") or "").strip()
+    confidence = "abandoned" if repo.get("archived") else "historic"
+
+    documents = []
+    try:
+        entries = client.tree(full_name, branch)
+    except GiteaNotFound:
+        entries = []
+
+    blobs = [entry for entry in entries if entry.get("type") == "blob"]
+    admitted = []
+    for entry in blobs:
+        path = globs.normalise(entry.get("path") or "")
+        kind, reason = exclusions.classify(path, guide, size=entry.get("size"))
+        if kind == "reject":
+            result.add_rejection(exclusions.Rejection(path, "gitea", reason, full_name))
+            continue
+        if kind == "image":
+            result.images.append(
+                {"path": path, "platform": "gitea", "repo": full_name,
+                 "origin_url": file_url(base_url, full_name, branch, path)}
+            )
+            continue
+        admitted.append(entry)
+
+    # The file tree is paths only. It is what lets a question like "does he have
+    # a Prometheus config anywhere?" find a repository whose README never says so.
+    tree_doc = _tree_document(guide, repo, area, updated, admitted, base_url)
+    if tree_doc:
+        documents.append(tree_doc)
+
+    # README first: it is the document a reader would open.
+    readmes = [
+        entry for entry in admitted
+        if _is_readme(entry.get("path") or "") and globs.normalise(entry.get("path") or "").count("/") <= 1
+    ][:3]
+    used = set()
+    for entry in readmes:
+        doc = _file_document(
+            result, client, guide, repo, branch, area, updated, created,
+            confidence, description, topics, entry, base_url,
+        )
+        if doc:
+            documents.append(doc)
+            used.add(globs.normalise(entry["path"]))
+
+    # Selected source files: the evidence behind the README's claims.
+    candidates = [
+        entry for entry in admitted
+        if globs.normalise(entry.get("path") or "") not in used
+        and not _is_readme(entry.get("path") or "")
+        and _source_rank(entry.get("path") or "")
+    ]
+    candidates.sort(key=lambda entry: (_depth(entry.get("path") or ""), globs.normalise(entry["path"])))
+    for entry in candidates[: guide.limit("max_files_per_repo")]:
+        doc = _file_document(
+            result, client, guide, repo, branch, area, updated, created,
+            confidence, description, topics, entry, base_url,
+        )
+        if doc:
+            documents.append(doc)
+
+    # One document per repository for commits and one for languages. These are
+    # short, stable, and answer "what is being worked on?" without a filesystem.
+    commits = _commit_document(result, client, guide, repo, area, updated, created, confidence, base_url)
+    if commits:
+        documents.append(commits)
+    languages = _language_document(result, client, guide, repo, area, updated, created, confidence, base_url)
+    if languages:
+        documents.append(languages)
+
+    if not any(not doc.source_id.endswith(("@tree", "@commits", "@languages")) for doc in documents):
+        # No README and no admitted source file: still admit a summary so the
+        # repository is discoverable and its silence is visible, not invisible.
+        summary = _summary_document(guide, repo, area, updated, created, confidence, description, topics, base_url)
+        if summary:
+            documents.insert(0, summary)
+
+    return documents
+
+
+def _file_document(result, client, guide, repo, branch, area, updated, created, confidence, description, topics, entry, base_url) -> object:
+    path = globs.normalise(entry.get("path") or "")
+    full_name = repo["full_name"]
+    try:
+        raw = client.raw(full_name, branch, path)
+    except GiteaError as exc:
+        result.warnings.append(f"gitea/{full_name}/{path}: {exc}")
+        return None
+    if exclusions.looks_binary(raw):
+        result.add_rejection(exclusions.Rejection(path, "gitea", "binary content", full_name))
+        return None
+    text = raw.decode("utf-8", "replace")
+    if not text.strip():
+        result.add_rejection(exclusions.Rejection(path, "gitea", "empty file", full_name))
+        return None
+    hits = exclusions.scan_text(text)
+    if hits:
+        kinds = sorted({hit[0] for hit in hits})
+        result.add_rejection(
+            exclusions.Rejection(path, "gitea", "credential-shaped content", full_name + ": " + ", ".join(kinds))
+        )
+        return None
+
+    kind = "code" if exclusions.extension(path) not in ("md", "markdown", "mdx", "txt", "rst") else "markdown"
+    title = f"{full_name} — {path}"
+    body = _strip_front_matter(text) if kind == "markdown" else text
+    if not body.strip():
+        result.add_rejection(exclusions.Rejection(path, "gitea", "front-matter only", full_name))
+        return None
+    if description or topics:
+        header = [f"Repository: {full_name}"]
+        if description:
+            header.append(description)
+        if topics:
+            header.append("Topics: " + ", ".join(topics))
+        body = "\n\n".join(["\n".join(header), text])
+    return _make_document(
+        key=f"gitea:{full_name}/{path}",
+        platform="gitea",
+        area=area,
+        title=title,
+        claim=_claim(text, title),
+        body=body,
+        digest=hashlib.sha256(raw).hexdigest(),
+        origin_url=file_url(base_url, full_name, branch, path),
+        path=path,
+        repo_path=path,
+        source_id=f"{full_name}/{path}",
+        source_type="code" if kind == "code" else "doc",
+        last_modified=updated,
+        confidence=confidence,
+        period=_period(created, updated),
+        tags=topics + ["gitea", area],
+        tech=[],
+        public=guide.public,
+        kind=kind,
+    )
+
+
+def _tree_document(guide, repo, area, updated, blobs, base_url) -> object:
+    if not blobs:
+        return None
+    full_name = repo["full_name"]
+    branch = repo.get("default_branch") or "main"
+    limit = guide.limit("max_tree_entries")
+    paths = [globs.normalise(entry.get("path") or "") for entry in blobs][:limit]
+    body = "\n".join(
+        [f"Repository: {full_name}", f"Branch: {branch}", f"Files: {len(blobs)}", "", "File tree:"]
+        + [f"- {path}" for path in paths]
+    )
+    if len(blobs) > limit:
+        body += f"\n- … {len(blobs) - limit} more"
+    digest = hashlib.sha256(body.encode("utf-8")).hexdigest()
+    return _make_document(
+        key=f"gitea:{full_name}@tree",
+        platform="gitea",
+        area=area,
+        title=f"{full_name} — file tree",
+        claim=f"File tree of {full_name}, {len(blobs)} files.",
+        body=body,
+        digest=digest,
+        origin_url=repo_url(base_url, full_name),
+        path="",
+        repo_path="",
+        source_id=f"{full_name}@tree",
+        source_type="doc",
+        last_modified=updated,
+        confidence="abandoned" if repo.get("archived") else "historic",
+        period="",
+        tags=["gitea", "tree", area],
+        tech=[],
+        public=guide.public,
+        kind="markdown",
+    )
+
+
+def _commit_document(result, client, guide, repo, area, updated, created, confidence, base_url) -> object:
+    full_name = repo["full_name"]
+    branch = repo.get("default_branch") or "main"
+    try:
+        commits = client.commits(full_name, branch, limit=guide.limit("max_commits"))
+    except GiteaError:
+        return None
+    if not commits:
+        return None
+    lines = [f"Repository: {full_name}", f"Recent commits on {branch}:", ""]
+    newest = updated
+    for item in commits:
+        commit = item.get("commit") or {}
+        message = (commit.get("message") or "").strip().splitlines()
+        headline = message[0] if message else "(no message)"
+        date = _timestamp(((commit.get("author") or {}).get("date")) or ((commit.get("committer") or {}).get("date")))
+        if date:
+            newest = max(newest, date)
+        lines.append(f"- {date[:10]} {headline}")
+    body = "\n".join(lines)
+    digest = hashlib.sha256(body.encode("utf-8")).hexdigest()
+    return _make_document(
+        key=f"gitea:{full_name}@commits",
+        platform="gitea",
+        area=area,
+        title=f"{full_name} — recent commits",
+        claim=f"Recent commit history for {full_name}.",
+        body=body,
+        digest=digest,
+        origin_url=f"{repo_url(base_url, full_name)}/commits/branch/{branch}",
+        path="",
+        repo_path="",
+        source_id=f"{full_name}@commits",
+        source_type="commit",
+        last_modified=newest,
+        confidence=confidence,
+        period=_period(created, newest),
+        tags=["gitea", "commits", area],
+        tech=[],
+        public=guide.public,
+        kind="markdown",
+    )
+
+
+def _language_document(result, client, guide, repo, area, updated, created, confidence, base_url) -> object:
+    full_name = repo["full_name"]
+    try:
+        languages = client.languages(full_name)
+    except GiteaError:
+        return None
+    if not languages:
+        return None
+    total = sum(int(value) for value in languages.values()) or 1
+    rows = sorted(languages.items(), key=lambda item: int(item[1]), reverse=True)
+    lines = [f"Repository: {full_name}", "", "Language statistics:"]
+    for name, size in rows:
+        lines.append(f"- {name}: {int(size)} bytes ({100.0 * int(size) / total:.1f}%)")
+    body = "\n".join(lines)
+    digest = hashlib.sha256(body.encode("utf-8")).hexdigest()
+    return _make_document(
+        key=f"gitea:{full_name}@languages",
+        platform="gitea",
+        area=area,
+        title=f"{full_name} — languages",
+        claim=f"Language breakdown for {full_name}: " + ", ".join(name for name, _ in rows[:4]) + ".",
+        body=body,
+        digest=digest,
+        origin_url=repo_url(base_url, full_name),
+        path="",
+        repo_path="",
+        source_id=f"{full_name}@languages",
+        source_type="metric",
+        last_modified=updated,
+        confidence=confidence,
+        period=_period(created, updated),
+        tags=["gitea", "languages", area],
+        tech=[name for name, _ in rows[:6]],
+        public=guide.public,
+        kind="markdown",
+    )
+
+
+def _summary_document(guide, repo, area, updated, created, confidence, description, topics, base_url) -> object:
+    full_name = repo["full_name"]
+    body = "\n".join(
+        [f"Repository: {full_name}", f"Description: {description or '(none)'}",
+         f"Topics: {', '.join(topics) if topics else '(none)'}"]
+    )
+    digest = hashlib.sha256(body.encode("utf-8")).hexdigest()
+    return _make_document(
+        key=f"gitea:{full_name}@summary",
+        platform="gitea",
+        area=area,
+        title=full_name,
+        claim=description or f"Repository {full_name}.",
+        body=body,
+        digest=digest,
+        origin_url=repo_url(base_url, full_name),
+        path="",
+        repo_path="",
+        source_id=f"{full_name}@summary",
+        source_type="doc",
+        last_modified=updated,
+        confidence=confidence,
+        period=_period(created, updated),
+        tags=topics + ["gitea", area],
+        tech=[],
+        public=guide.public,
+        kind="markdown",
+    )
+
+
+# -- obsidian vault -------------------------------------------------------
+
+
+def _harvest_obsidian(result, client, guide, repos, vault_repo, tree, cfg, base_url, log) -> None:
+    if not vault_repo or tree is None:
+        result.warnings.append("obsidian: the vault repository could not be listed")
+        result.sources["obsidian"] = _source_row("obsidian", enabled=True, documents=[], extra={"missing": True})
+        return
+    repo = next((row for row in repos if row.get("full_name") == vault_repo), {})
+    branch = repo.get("default_branch") or "main"
+    updated = _timestamp(repo.get("updated_at"))
+    created = _timestamp(repo.get("created_at"))
+    validator = loader.load_validator(cfg.root)
+    documents = []
+
+    blobs = [entry for entry in tree if entry.get("type") == "blob"]
+    for entry in blobs:
+        if len(documents) >= guide.limit("max_documents"):
+            result.warnings.append(
+                f"obsidian: document limit {guide.limit('max_documents')} reached"
+            )
+            break
+        path = globs.normalise(entry.get("path") or "")
+        kind, reason = exclusions.classify(path, guide, size=entry.get("size"))
+        if kind == "reject":
+            # A path outside the include list is expected and is not a finding;
+            # only hard exclusions are reported, so the report stays readable.
+            if reason != "not under an included path":
+                result.add_rejection(exclusions.Rejection(path, "obsidian", reason))
+            continue
+        if kind == "image":
+            result.images.append(
+                {"path": path, "platform": "obsidian",
+                 "origin_url": file_url(base_url, vault_repo, branch, path)}
+            )
+            continue
+        try:
+            raw = client.raw(vault_repo, branch, path)
+        except GiteaError as exc:
+            result.warnings.append(f"obsidian/{path}: {exc}")
+            continue
+        if exclusions.looks_binary(raw):
+            result.add_rejection(exclusions.Rejection(path, "obsidian", "binary content"))
+            continue
+        text = raw.decode("utf-8", "replace")
+        if not text.strip():
+            result.add_rejection(exclusions.Rejection(path, "obsidian", "empty file"))
+            continue
+        hits = exclusions.scan_text(text)
+        if hits:
+            result.add_rejection(
+                exclusions.Rejection(
+                    path, "obsidian", "credential-shaped content",
+                    ", ".join(sorted({hit[0] for hit in hits})),
+                )
+            )
+            continue
+
+        fields, body = _split_note(text, validator)
+        if not body.strip():
+            result.add_rejection(exclusions.Rejection(path, "obsidian", "empty note"))
+            continue
+        title = str(fields.get("title") or _heading(body) or os.path.splitext(os.path.basename(path))[0])
+        tag_values = fields.get("tags") or []
+        if isinstance(tag_values, str):
+            tag_values = [tag_values]
+        tags = [str(tag) for tag in tag_values] + _folder_tags(path)
+        area = areas.area_for_note(path, guide.fits)
+        documents.append(
+            _make_document(
+                key=f"obsidian:{path}",
+                platform="obsidian",
+                area=area,
+                title=title,
+                claim=_claim(body, title),
+                body=body,
+                digest=hashlib.sha256(raw).hexdigest(),
+                origin_url=file_url(base_url, vault_repo, branch, path),
+                path=path,
+                repo_path=path,
+                source_id=path,
+                source_type="doc",
+                last_modified=updated,
+                confidence="historic",
+                period=_period(created, updated),
+                tags=tags + ["obsidian", area],
+                tech=[],
+                public=guide.public,
+                kind="markdown",
+            )
+        )
+
+    result.documents.extend(documents)
+    result.sources["obsidian"] = _source_row(
+        "obsidian", enabled=True, documents=documents,
+        extra={"repo": vault_repo, "base": base_url},
+    )
+    log(f"obsidian: {len(documents)} note(s)")
+
+
+# -- graphify -------------------------------------------------------------
+
+
+def _harvest_graph(result, client, guide, repos, vault_repo, tree, cfg, base_url, log) -> None:
+    if not vault_repo:
+        result.warnings.append("graphify: no vault repository configured in the guide")
+        result.sources["graphify"] = _source_row("graphify", enabled=True, documents=[], extra={"missing": True})
+        return
+    repo = next((row for row in repos if row.get("full_name") == vault_repo), {})
+    branch = repo.get("default_branch") or "main"
+    updated = _timestamp(repo.get("updated_at"))
+    vault_paths = {
+        globs.normalise(entry.get("path") or "")
+        for entry in (tree or [])
+        if entry.get("type") == "blob"
+    }
+
+    raw = None
+    source_path = ""
+    if cfg.graph_path:
+        if not os.path.exists(cfg.graph_path):
+            raise EstateError(f"RESUME_GRAPH_PATH is set but {cfg.graph_path} does not exist")
+        with open(cfg.graph_path, "rb") as handle:
+            raw = handle.read()
+        source_path = cfg.graph_path
+    else:
+        for candidate in guide.limits.get("graph_paths") or ["graph.json"]:
+            try:
+                raw = client.raw(vault_repo, branch, candidate)
+                source_path = f"gitea:{vault_repo}/{candidate}"
+                break
+            except GiteaNotFound:
+                continue
+            except GiteaError as exc:
+                raise EstateError(f"cannot read graph.json from {vault_repo}: {exc}") from exc
+
+    if raw is None:
+        result.warnings.append(
+            f"graphify: graph.json is not committed to {vault_repo} "
+            f"(looked in {guide.limits.get('graph_paths')}). The graph was not ingested; "
+            "the nightly audit will keep reporting this until it is committed."
+        )
+        result.sources["graphify"] = _source_row(
+            "graphify", enabled=True, documents=[], extra={"missing": True, "repo": vault_repo}
+        )
+        return
+
+    graph = graph_module.parse(
+        raw,
+        platform="graphify",
+        base=base_url,
+        repo=vault_repo,
+        ref=branch,
+        last_modified=updated,
+        area=guide.primary_area,
+        vault_paths=vault_paths,
+        max_nodes=guide.limit("max_nodes"),
+        max_edges=guide.limit("max_edges"),
+        source_path=source_path,
+    )
+    result.graph = graph
+    for rejection in graph.rejections:
+        result.add_rejection(rejection)
+    result.images.extend(graph.images)
+    result.sources["graphify"] = _source_row(
+        "graphify", enabled=True, documents=[],
+        extra={
+            "repo": vault_repo,
+            "path": source_path,
+            "nodes": len(graph.nodes),
+            "edges": len(graph.edges),
+            "digest": graph.digest,
+            "built_at_commit": graph.built_at_commit,
+        },
+    )
+    if graph.warnings:
+        result.warnings.extend(f"graphify: {warning}" for warning in graph.warnings)
+    log(
+        f"graphify: {len(graph.nodes)} node(s), {len(graph.edges)} edge(s), "
+        f"{len(graph.images)} image node(s)"
+    )
+
+
+# -- shared helpers -------------------------------------------------------
+
+
+def _make_document(**values) -> loader.Document:
+    """One Document constructor, so every source produces the same shape."""
+    defaults = dict(
+        metrics=[], artefacts=[], tech=[], tags=[], public=True, kind="markdown",
+        platform="gitea", tier=2, level=3, period="", source_type="doc",
+    )
+    defaults.update(values)
+    title = defaults["title"]
+    origin = defaults["origin_url"]
+    path = defaults.get("path") or ""
+    return loader.Document(
+        path=defaults["key"],
+        source_id=defaults.get("source_id") or (path or defaults["key"]),
+        repo_path=defaults.get("repo_path") or path,
+        area=defaults["area"],
+        level=defaults.get("level", 3),
+        title=title,
+        claim=defaults.get("claim") or title,
+        confidence=defaults.get("confidence") or "historic",
+        period=defaults.get("period") or "—",
+        tech=defaults.get("tech") or [],
+        metrics=defaults.get("metrics") or [],
+        artefacts=defaults.get("artefacts") or [],
+        public=bool(defaults.get("public", True)),
+        last_verified=str(defaults["last_modified"])[:10],
+        tags=defaults.get("tags") or [],
+        body=defaults["body"],
+        digest=defaults["digest"],
+        url=origin,
+        platform=defaults["platform"],
+        tier=int(defaults.get("tier", 2)),
+        origin_url=origin,
+        last_modified=defaults["last_modified"],
+        source_type=defaults.get("source_type") or "doc",
+        kind=defaults.get("kind") or "markdown",
+    )
+
+
+def _strip_front_matter(text: str) -> str:
+    """Drop leading YAML blocks. The Note path parses them; files only need them gone.
+
+    Some notes carry two blocks in a row (a stale export, or a block followed by
+    a horizontal rule that closes), so this repeats. It never strips a body that
+    does not have a closing delimiter.
+    """
+    body = text
+    for _ in range(3):
+        lines = body.splitlines()
+        if not lines or lines[0].strip() != "---":
+            return body
+        closing = None
+        for index in range(1, len(lines)):
+            if lines[index].strip() in ("---", "..."):
+                closing = index
+                break
+        if closing is None:
+            return body
+        body = "\n".join(lines[closing + 1 :]).strip()
+    return body
+
+
+def _split_note(text: str, validator) -> tuple:
+    """Tolerant front-matter split. A note without front-matter is still a note."""
+    lines = text.splitlines()
+    if not lines or lines[0].strip() != "---":
+        return {}, text
+    end = None
+    for index in range(1, len(lines)):
+        if lines[index].strip() in ("---", "..."):
+            end = index
+            break
+    if end is None:
+        return {}, text
+    fields = {}
+    try:
+        node = validator.parse_front_matter(
+            [(number, line) for number, line in enumerate(lines[1:end], start=2)]
+        )
+        parsed = loader.to_python(node)
+        if isinstance(parsed, dict):
+            fields = parsed
+    except validator.YamlError:
+        fields = {}
+    # Strip any further leading block as well, so a duplicate front-matter block
+    # never reaches the chunker.
+    return fields, _strip_front_matter("\n".join(lines[end + 1 :]).strip())
+
+
+def _source_row(name: str, enabled: bool, documents: list, extra: dict = None) -> dict:
+    stamps = sorted(doc.last_modified for doc in documents if doc.last_modified)
+    row = {
+        "platform": name,
+        "enabled": enabled,
+        "documents": len(documents),
+        "newest": stamps[-1] if stamps else "",
+        "oldest": stamps[0] if stamps else "",
+        "fetched_at": datetime.datetime.now(datetime.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
+    }
+    row.update(extra or {})
+    return row
+
+
+def _vault_repo(selected: dict) -> str:
+    for name in ("obsidian", "graphify"):
+        guide = selected.get(name)
+        if guide and guide.base and "/" in guide.base and not guide.base.startswith("http"):
+            return guide.base
+    return ""
+
+
+def _as_list(value) -> list:
+    if value in (None, ""):
+        return []
+    if isinstance(value, list):
+        return [str(item) for item in value]
+    return [str(value)]
+
+
+def _timestamp(value) -> str:
+    """Normalise any RFC3339 timestamp to UTC `…Z`. Empty stays empty."""
+    if not value:
+        return "1970-01-01T00:00:00Z"
+    text = str(value).strip()
+    if text.endswith("Z"):
+        text = text[:-1] + "+00:00"
+    try:
+        moment = datetime.datetime.fromisoformat(text)
+    except ValueError:
+        return "1970-01-01T00:00:00Z"
+    if moment.tzinfo is None:
+        moment = moment.replace(tzinfo=datetime.timezone.utc)
+    return moment.astimezone(datetime.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+
+
+def _period(created: str, updated: str) -> str:
+    start = (created or "")[:4]
+    end = (updated or "")[:4]
+    if not start or start == "1970":
+        return end or "—"
+    if start == end:
+        return start
+    return f"{start}–{end}"
+
+
+def _claim(text: str, title: str) -> str:
+    for line in (text or "").splitlines():
+        stripped = line.strip()
+        if not stripped or stripped.startswith(("#", "|", "```", "---", ">")):
+            continue
+        cleaned = re.sub(r"[*_`\[\]]", "", stripped)
+        if len(cleaned) > 20:
+            return cleaned[:180]
+    return title[:180]
+
+
+def _heading(text: str) -> str:
+    for line in (text or "").splitlines():
+        if line.startswith("# "):
+            return line[2:].strip()
+    return ""
+
+
+def _folder_tags(path: str) -> list:
+    parts = [part for part in path.split("/")[:-1] if part and not part[0].isdigit()]
+    return [part.lower() for part in parts[-2:]]
+
+
+def _is_readme(path: str) -> bool:
+    return os.path.basename(path).lower().startswith("readme")
+
+
+def _depth(path: str) -> int:
+    return globs.normalise(path).count("/")
+
+
+def _source_rank(path: str) -> bool:
+    ext = exclusions.extension(path)
+    if ext not in exclusions.TEXT_EXTENSIONS:
+        return False
+    if ext in ("json", "csv", "tsv", "svg"):
+        return False
+    # Prefer files that carry decisions over generated or vendored files.
+    return not any(
+        part.lower() in ("vendor", "third_party", "examples", "testdata", "fixtures", "migrations")
+        for part in globs.normalise(path).split("/")[:-1]
+    )

corpus/sources/gitea.py

@@ -0,0 +1,165 @@
+"""Gitea API client — the only network dependency in the pipeline.
+
+`docs/phase-4-corpus.md` §1: Gitea is reached over HTTPS, and the vault and the
+graph arrive through the same API, so there is no cross-machine copy and no SSH
+from the runtime host. This client is standard library only (`urllib`), it never
+prints or stores the token, and every call names only the path it requested.
+
+Read-only by construction: there is no method here that writes, creates, deletes
+or changes anything in Gitea. The token only needs read scope.
+"""
+
+from __future__ import annotations
+
+import http.client
+import json
+import time
+import urllib.error
+import urllib.parse
+import urllib.request
+
+
+class GiteaError(RuntimeError):
+    """A Gitea call failed. The message never contains the token."""
+
+
+class GiteaNotFound(GiteaError):
+    """The requested repository, branch or file does not exist."""
+
+
+class GiteaClient:
+    def __init__(self, base: str, token: str, timeout: float = 30.0, attempts: int = 3) -> None:
+        if not base:
+            raise GiteaError("a Gitea base URL is required (GITEA_URL or the guide's base)")
+        if not token:
+            raise GiteaError(
+                "a Gitea token is required. It is read from the environment variable "
+                "named by the platform guide (GITEA_TOKEN) and never from this repository."
+            )
+        self.base = base.rstrip("/")
+        self._token = token
+        self.timeout = timeout
+        self.attempts = max(1, attempts)
+        self.calls = 0
+
+    # -- transport --------------------------------------------------------
+
+    def _request(self, path: str, expect_bytes: bool = False):
+        """One GET with retry/backoff. The only place that touches the network."""
+        url = f"{self.base}/api/v1{path}"
+        last_error = None
+        for attempt in range(self.attempts):
+            request = urllib.request.Request(
+                url,
+                method="GET",
+                headers={
+                    "Authorization": f"token {self._token}",
+                    "Accept": "application/vnd.gitea+json, application/json",
+                    "User-Agent": "ai-resume-corpus/1",
+                },
+            )
+            try:
+                self.calls += 1
+                with urllib.request.urlopen(request, timeout=self.timeout) as response:
+                    raw = response.read()
+                if expect_bytes:
+                    return raw
+                return json.loads(raw.decode("utf-8")) if raw.strip() else {}
+            except urllib.error.HTTPError as exc:
+                detail = exc.read().decode("utf-8", "replace")[:200]
+                exc.close()
+                if exc.code == 404:
+                    raise GiteaNotFound(f"GET {path}: not found") from None
+                if exc.code == 401:
+                    raise GiteaError(
+                        f"GET {path}: authentication failed. Check the token named by the "
+                        "platform guide; it is not printed here."
+                    ) from None
+                if exc.code == 403:
+                    raise GiteaError(f"GET {path}: forbidden (the token may lack read scope)") from None
+                if exc.code != 429 and exc.code < 500:
+                    raise GiteaError(f"GET {path}: HTTP {exc.code} {detail}") from None
+                last_error = f"HTTP {exc.code}"
+            except (urllib.error.URLError, TimeoutError, http.client.HTTPException, ConnectionError) as exc:
+                last_error = f"cannot reach {self.base}: {getattr(exc, 'reason', exc)}"
+            if attempt + 1 < self.attempts:
+                time.sleep(0.5 * (2 ** attempt))
+        raise GiteaError(f"GET {path} failed after {self.attempts} attempts: {last_error}")
+
+    # -- repositories -----------------------------------------------------
+
+    def repos(self, limit: int = 50) -> list:
+        """Every repository the token can read, sorted by full name."""
+        out = []
+        page = 1
+        while True:
+            response = self._request(f"/repos/search?limit={limit}&page={page}")
+            rows = response.get("data") or []
+            out.extend(rows)
+            if len(rows) < limit:
+                break
+            page += 1
+            if page > 200:  # a guard, not a limit anyone should reach
+                break
+        return sorted(out, key=lambda repo: repo.get("full_name", ""))
+
+    def repo(self, full_name: str) -> dict:
+        return self._request(f"/repos/{_repo_path(full_name)}")
+
+    # -- contents ---------------------------------------------------------
+
+    def tree(self, full_name: str, ref: str) -> list:
+        """Recursive blob/tree listing. Returns every entry, following pages."""
+        entries = []
+        page = 1
+        while True:
+            response = self._request(
+                f"/repos/{_repo_path(full_name)}/git/trees/{urllib.parse.quote(ref, safe='')}"
+                f"?recursive=true&per_page=1000&page={page}"
+            )
+            rows = response.get("tree") or []
+            entries.extend(rows)
+            if not rows or not response.get("truncated") or len(rows) < 1000:
+                break
+            page += 1
+            if page > 50:
+                break
+        return entries
+
+    def raw(self, full_name: str, ref: str, path: str) -> bytes:
+        quoted = urllib.parse.quote(path, safe="/")
+        return self._request(
+            f"/repos/{_repo_path(full_name)}/raw/{quoted}?ref={urllib.parse.quote(ref, safe='')}",
+            expect_bytes=True,
+        )
+
+    def commits(self, full_name: str, ref: str, limit: int = 20) -> list:
+        return self._request(
+            f"/repos/{_repo_path(full_name)}/commits?sha={urllib.parse.quote(ref, safe='')}"
+            f"&limit={int(limit)}&page=1"
+        )
+
+    def languages(self, full_name: str) -> dict:
+        return self._request(f"/repos/{_repo_path(full_name)}/languages")
+
+    # -- presentation -----------------------------------------------------
+
+    def describe(self) -> dict:
+        return {"base": self.base, "token": "set", "calls": self.calls}
+
+
+def file_url(base: str, full_name: str, ref: str, path: str) -> str:
+    """The URL a reader can open: the file at its branch, in Gitea."""
+    return (
+        f"{base.rstrip('/')}/{full_name}/src/branch/{urllib.parse.quote(ref, safe='')}/"
+        f"{urllib.parse.quote(path, safe='/')}"
+    )
+
+
+def repo_url(base: str, full_name: str) -> str:
+    return f"{base.rstrip('/')}/{full_name}"
+
+
+def _repo_path(full_name: str) -> str:
+    owner, _, name = full_name.partition("/")
+    return f"{urllib.parse.quote(owner, safe='')}/{urllib.parse.quote(name, safe='')}"

corpus/sources/graph.py

@@ -0,0 +1,335 @@
+"""Graphify output: nodes and edges, each joined back to the note it came from.
+
+`docs/phase-4-corpus.md` §4: *"Graph is not chunked. Nodes and edges go into
+their own collection, joined to note chunks by `path`, so a graph answer can name
+the note it came from."*
+
+The graph is not rebuilt here. It already exists and is rebuilt on the build host
+by its own skill; this module consumes `graph.json`.
+
+Two rules from §3 drive the whole file:
+
+* **A node with no resolvable source note is refused, not defaulted.** Graphify
+  emits nodes for files outside the vault (`settings.json`, `.pi/…`). Those are
+  not traceable to a note, so they are rejected and counted.
+* **Media never becomes text.** A media node becomes an image reference: path,
+  origin URL and note, and nothing else.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import re
+import uuid
+
+from .. import areas, exclusions, globs
+from ..sources.gitea import file_url
+
+
+class GraphError(RuntimeError):
+    """graph.json is not a graph this pipeline can admit."""
+
+
+class GraphResult:
+    def __init__(self, platform: str = "graphify"):
+        self.platform = platform
+        self.nodes: list = []
+        self.edges: list = []
+        self.images: list = []
+        self.rejections: list = []
+        self.warnings: list = []
+        self.digest = ""
+        self.source_path = ""
+        self.built_at_commit = ""
+
+    def counts(self) -> dict:
+        return {
+            "nodes": len(self.nodes),
+            "edges": len(self.edges),
+            "images": len(self.images),
+            "rejected": len(self.rejections),
+        }
+
+
+def parse(
+    raw: bytes,
+    *,
+    platform: str,
+    base: str,
+    repo: str,
+    ref: str,
+    last_modified: str,
+    area: str,
+    vault_paths: set,
+    max_nodes: int = 20000,
+    max_edges: int = 40000,
+    source_path: str = "",
+) -> GraphResult:
+    """Turn graph.json bytes into rows that carry their note. Never guesses."""
+    result = GraphResult(platform)
+    result.digest = hashlib.sha256(raw).hexdigest()
+    result.source_path = source_path
+    try:
+        data = json.loads(raw.decode("utf-8"))
+    except (UnicodeDecodeError, json.JSONDecodeError) as exc:
+        raise GraphError(f"graph.json is not valid UTF-8 JSON: {exc}") from exc
+    if not isinstance(data, dict):
+        raise GraphError("graph.json must be a JSON object")
+    result.built_at_commit = str(data.get("built_at_commit") or "")
+
+    raw_nodes = data.get("nodes") or []
+    raw_links = data.get("links") or data.get("edges") or []
+    if not isinstance(raw_nodes, list) or not isinstance(raw_links, list):
+        raise GraphError("graph.json must contain `nodes` and `links` arrays")
+
+    labels = {}
+    notes = {}
+    for item in raw_nodes:
+        if not isinstance(item, dict):
+            continue
+        node_id = str(item.get("id") or "").strip()
+        if not node_id:
+            continue
+        labels[node_id] = str(item.get("label") or node_id)
+        notes[node_id] = _note_for(item, vault_paths)
+
+    for item in raw_nodes:
+        if len(result.nodes) >= max_nodes:
+            result.warnings.append(f"graph node limit {max_nodes} reached; remaining nodes skipped")
+            break
+        if not isinstance(item, dict):
+            continue
+        node_id = str(item.get("id") or "").strip()
+        if not node_id:
+            result.rejections.append(
+                _rejection("graph.json", platform, "graph node has no id")
+            )
+            continue
+        source_file = globs.normalise(item.get("source_file") or "")
+        note = notes.get(node_id)
+        if not note:
+            result.rejections.append(
+                _rejection(
+                    source_file or f"node:{node_id}", platform,
+                    "no resolvable source note", "graph node does not point at a file in the vault",
+                )
+            )
+            continue
+        if globs.normalise(source_file) != note:
+            # A node can be derived from one note while naming another file; the
+            # note is the traceable origin either way.
+            pass
+        if exclusions.classify(note, None)[0] == "image" or exclusions.extension(note) in exclusions.MEDIA_EXTENSIONS:
+            result.images.append(
+                {
+                    "path": note,
+                    "label": str(item.get("label") or node_id),
+                    "origin_url": file_url(base, repo, ref, note),
+                    "platform": platform,
+                    "note": note,
+                }
+            )
+            continue
+
+        node_area = areas.area_for(note, (area,), fallback=area)
+        label = str(item.get("label") or node_id)
+        content_hash = _hash(
+            {"kind": "node", "id": node_id, "label": label, "note": note,
+             "community": item.get("community"), "type": item.get("node_type")}
+        )
+        text = (
+            f"{label}. Knowledge graph node"
+            + (f" in community {item.get('community')}" if item.get("community") is not None else "")
+            + f", from the note {note}."
+        )
+        result.nodes.append(
+            {
+                "id": node_id,
+                "point_id": _point_id("node", node_id, content_hash),
+                "graph_id": node_id,
+                "label": label,
+                "norm_label": str(item.get("norm_label") or _norm(label)),
+                "node_type": str(item.get("node_type") or item.get("file_type") or "entity"),
+                "community": item.get("community"),
+                "source_note": note,
+                "origin_url": file_url(base, repo, ref, note),
+                "area": node_area,
+                "tier": 2,
+                "platform": platform,
+                "last_modified": last_modified,
+                "content_hash": content_hash,
+                "text": text,
+            }
+        )
+
+    admitted_nodes = {row["graph_id"]: row for row in result.nodes}
+    for item in raw_links:
+        if len(result.edges) >= max_edges:
+            result.warnings.append(f"graph edge limit {max_edges} reached; remaining edges skipped")
+            break
+        if not isinstance(item, dict):
+            continue
+        source = str(item.get("source") or "").strip()
+        target = str(item.get("target") or "").strip()
+        if not source or not target:
+            result.rejections.append(_rejection("graph.json", platform, "graph edge has no endpoints"))
+            continue
+        note = notes.get(source) or _note_for(item, vault_paths)
+        if not note:
+            result.rejections.append(
+                _rejection(
+                    f"edge:{source}->{target}", platform,
+                    "no resolvable source note", "graph edge does not point at a file in the vault",
+                )
+            )
+            continue
+        source_label = labels.get(source, source)
+        target_label = labels.get(target, target)
+        relation = str(item.get("relation") or item.get("rel") or "related_to")
+        area_name = areas.area_for(note, (area,), fallback=area)
+        content_hash = _hash(
+            {"kind": "edge", "source": source, "target": target,
+             "relation": relation, "note": note}
+        )
+        text = f"{source_label} {relation} {target_label}, from the note {note}."
+        result.edges.append(
+            {
+                "id": _edge_id(source, relation, target, content_hash),
+                "point_id": _point_id("edge", f"{source}>{relation}>{target}", content_hash),
+                "source_id": source,
+                "target_id": target,
+                "relation": relation,
+                "confidence": str(item.get("confidence") or "INFERRED"),
+                "confidence_score": item.get("confidence_score"),
+                "source_note": note,
+                "origin_url": file_url(base, repo, ref, note),
+                "area": area_name,
+                "tier": 2,
+                "platform": platform,
+                "last_modified": last_modified,
+                "content_hash": content_hash,
+                "text": text,
+            }
+        )
+
+    # Edges whose endpoints were not admitted are still traceable and are kept;
+    # an answer can name the relation and its note even when the node itself had
+    # no origin. `admitted_node_ids` is exposed for verify, which reports the gap.
+    result.admitted_node_ids = set(admitted_nodes)
+    result.orphan_edges = [
+        edge for edge in result.edges
+        if edge["source_id"] not in admitted_nodes or edge["target_id"] not in admitted_nodes
+    ]
+    return result
+
+
+def point_payload(row: dict, kind: str) -> dict:
+    """One Qdrant point payload for a node or edge. Same provenance as the row."""
+    payload = {
+        "platform": row["platform"],
+        "tier": row["tier"],
+        "source_type": "graph_node" if kind == "node" else "graph_edge",
+        "area": row["area"],
+        "origin_url": row["origin_url"],
+        "path": row["source_note"],
+        "source_note": row["source_note"],
+        "title": row.get("label") or row.get("relation") or row["source_note"],
+        "confidence": row.get("confidence") or "INFERRED",
+        "last_modified": row["last_modified"],
+        "content_hash": row["content_hash"],
+        "text": row["text"],
+        "kind": kind,
+        "graph_id": row.get("graph_id") or row.get("id"),
+    }
+    if kind == "node":
+        payload["community"] = row.get("community")
+        payload["node_type"] = row.get("node_type")
+    else:
+        payload["relation"] = row.get("relation")
+        payload["source_id"] = row.get("source_id")
+        payload["target_id"] = row.get("target_id")
+        payload["confidence_score"] = row.get("confidence_score")
+    return payload
+
+
+def query(cfg, store, text: str, area: str = None, limit: int = 10) -> dict:
+    """Graph query: keyword hits from SQLite, vector hits from Qdrant when up.
+
+    Returning both is deliberate: the keyword side always works, and the vector
+    side proves the graph collection is live. Each row carries `source_note`.
+    """
+    from ..qdrant import Qdrant, QdrantError
+
+    report = {"query": text, "area": area, "keyword": [], "vector": [], "ok": True}
+    report["keyword"] = store.graph_search(text, area=area, limit=limit)
+    if not report["keyword"]:
+        report["ok"] = False
+        report["reason"] = "no graph rows matched the query"
+    qdrant = Qdrant(cfg.qdrant_url, timeout=cfg.timeout)
+    if not qdrant.health() or not qdrant.collection_exists(cfg.graph_collection):
+        report["vector"] = []
+        report["vector_note"] = f"graph collection {cfg.graph_collection} is not available"
+        return report
+    from .. import embeddings
+
+    provider = embeddings.build_provider(cfg)
+    try:
+        vector = provider.embed([text])[0]
+        hits = qdrant.search(cfg.graph_collection, vector, limit=limit, area=area)
+    except (QdrantError, Exception) as exc:  # noqa: BLE001 - a graph query must not crash a shell
+        report["vector_note"] = f"graph vector search unavailable: {exc}"
+        return report
+    report["vector"] = [
+        {
+            "score": round(hit.get("score", 0.0), 4),
+            "kind": (hit.get("payload") or {}).get("kind"),
+            "label": (hit.get("payload") or {}).get("title"),
+            "source_note": (hit.get("payload") or {}).get("source_note"),
+            "origin_url": (hit.get("payload") or {}).get("origin_url"),
+        }
+        for hit in hits
+    ]
+    return report
+
+
+# -- helpers -------------------------------------------------------------
+
+
+def _note_for(item: dict, vault_paths: set) -> str:
+    source = globs.normalise(item.get("source_file") or "")
+    if not source:
+        return ""
+    if source in vault_paths:
+        return source
+    # Graphify sometimes prefixes `./` or uses an absolute scan root. A suffix
+    # match is only accepted when it is unique, so a note is never guessed.
+    candidates = [path for path in vault_paths if path.endswith("/" + source)]
+    if len(candidates) == 1:
+        return candidates[0]
+    return ""
+
+
+def _rejection(path: str, platform: str, reason: str, detail: str = ""):
+    rejection = exclusions.Rejection(path, platform, reason, detail)
+    return rejection.to_dict()
+
+
+def _hash(value) -> str:
+    body = json.dumps(value, sort_keys=True, separators=(",", ":")).encode("utf-8")
+    return hashlib.sha256(body).hexdigest()
+
+
+def _point_id(kind: str, key: str, content_hash: str) -> str:
+    digest = hashlib.sha256(f"graph\x00{kind}\x00{key}\x00{content_hash}".encode("utf-8")).hexdigest()
+    return str(uuid.UUID(hex=digest[:32]))
+
+
+def _edge_id(source: str, relation: str, target: str, content_hash: str) -> str:
+    return hashlib.sha256(
+        f"edge\x00{source}\x00{relation}\x00{target}\x00{content_hash}".encode("utf-8")
+    ).hexdigest()[:32]
+
+
+def _norm(label: str) -> str:
+    return re.sub(r"[^0-9a-z]+", "_", (label or "").lower()).strip("_")

corpus/store.py

@@ -0,0 +1,479 @@
+"""SQLite: the keyword index and the incremental state.
+
+Two jobs, one file (`RESUME_CORPUS_DB`, default `data/corpus/corpus.db`):
+
+* **`chunk_fts`** — an FTS5 index over chunk text with the same filter columns
+  the vectors carry. This is the keyword half of hybrid retrieval (decision
+  D10). Phase 2 indexes *pages* for the site's own search box; this indexes
+  *chunks*, because retrieval hands the model a passage, not a page.
+* **`manifest`** — one row per source document: its content hash, its chunk
+  count and the embedding model/dimension it was embedded with. Incremental mode
+  is a hash comparison against this table, and nothing else.
+
+`meta` records the collection contract (provider, model, dimension, chunker
+version) so a provider change is detected instead of silently corrupting the
+index.
+
+Standard library `sqlite3` only. The same file is opened read-only by the Phase
+5 retrieval service.
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import re
+import sqlite3
+import time
+import unicodedata
+
+SCHEMA_VERSION = "2"
+
+SCHEMA = """
+CREATE TABLE IF NOT EXISTS chunks (
+    id               TEXT PRIMARY KEY,
+    source_path      TEXT NOT NULL,
+    source_id        TEXT NOT NULL,
+    chunk_index      INTEGER NOT NULL,
+    area             TEXT NOT NULL,
+    level            INTEGER NOT NULL,
+    source_type      TEXT NOT NULL,
+    confidence       TEXT NOT NULL,
+    period           TEXT NOT NULL,
+    last_verified    TEXT NOT NULL,
+    title            TEXT NOT NULL,
+    claim            TEXT NOT NULL,
+    url              TEXT NOT NULL,
+    content_hash     TEXT NOT NULL,
+    embedding_model  TEXT NOT NULL,
+    embedding_dim    INTEGER NOT NULL,
+    chunker_version  TEXT NOT NULL,
+    chunk_tokens     INTEGER NOT NULL,
+    platform         TEXT NOT NULL,
+    tier             INTEGER NOT NULL,
+    origin_url       TEXT NOT NULL,
+    path             TEXT NOT NULL,
+    last_modified    TEXT NOT NULL,
+    public           INTEGER NOT NULL,
+    artefacts        TEXT NOT NULL,
+    tech             TEXT NOT NULL,
+    tags             TEXT NOT NULL,
+    text             TEXT NOT NULL
+);
+CREATE INDEX IF NOT EXISTS chunks_source ON chunks(source_path);
+CREATE INDEX IF NOT EXISTS chunks_area ON chunks(area);
+
+CREATE VIRTUAL TABLE IF NOT EXISTS chunk_fts USING fts5(
+    id UNINDEXED,
+    source_path UNINDEXED,
+    title,
+    terms,
+    text,
+    area UNINDEXED,
+    confidence UNINDEXED,
+    platform UNINDEXED,
+    tier UNINDEXED,
+    tokenize = 'porter unicode61 remove_diacritics 2'
+);
+
+CREATE TABLE IF NOT EXISTS manifest (
+    source_path      TEXT PRIMARY KEY,
+    content_hash     TEXT NOT NULL,
+    chunk_count      INTEGER NOT NULL,
+    embedding_model  TEXT NOT NULL,
+    embedding_dim    INTEGER NOT NULL,
+    chunker_version  TEXT NOT NULL,
+    last_verified    TEXT NOT NULL,
+    updated_at       TEXT NOT NULL
+);
+
+CREATE TABLE IF NOT EXISTS meta (
+    key   TEXT PRIMARY KEY,
+    value TEXT NOT NULL
+);
+
+-- The knowledge graph is not chunked (docs/phase-4-corpus.md §4). Nodes and
+-- edges live in their own tables and their own Qdrant collection, joined to note
+-- chunks by source_note. Every row carries the note it came from.
+CREATE TABLE IF NOT EXISTS graph_nodes (
+    id             TEXT PRIMARY KEY,
+    label          TEXT NOT NULL,
+    norm_label     TEXT NOT NULL,
+    node_type      TEXT NOT NULL,
+    community      INTEGER,
+    source_note    TEXT NOT NULL,
+    origin_url     TEXT NOT NULL,
+    area           TEXT NOT NULL,
+    tier           INTEGER NOT NULL,
+    platform       TEXT NOT NULL,
+    last_modified  TEXT NOT NULL,
+    content_hash   TEXT NOT NULL,
+    text           TEXT NOT NULL
+);
+CREATE INDEX IF NOT EXISTS graph_nodes_note ON graph_nodes(source_note);
+
+CREATE TABLE IF NOT EXISTS graph_edges (
+    id                TEXT PRIMARY KEY,
+    source_id         TEXT NOT NULL,
+    target_id         TEXT NOT NULL,
+    relation          TEXT NOT NULL,
+    confidence        TEXT NOT NULL,
+    confidence_score  REAL,
+    source_note       TEXT NOT NULL,
+    origin_url        TEXT NOT NULL,
+    area              TEXT NOT NULL,
+    tier              INTEGER NOT NULL,
+    platform          TEXT NOT NULL,
+    last_modified     TEXT NOT NULL,
+    content_hash      TEXT NOT NULL,
+    text              TEXT NOT NULL
+);
+CREATE INDEX IF NOT EXISTS graph_edges_source ON graph_edges(source_id);
+CREATE INDEX IF NOT EXISTS graph_edges_target ON graph_edges(target_id);
+CREATE INDEX IF NOT EXISTS graph_edges_note ON graph_edges(source_note);
+
+CREATE VIRTUAL TABLE IF NOT EXISTS graph_fts USING fts5(
+    id UNINDEXED,
+    kind UNINDEXED,
+    label,
+    text,
+    area UNINDEXED,
+    tokenize = 'porter unicode61 remove_diacritics 2'
+);
+"""
+
+# Column order used by every INSERT in this module.
+_CHUNK_COLUMNS = (
+    "id", "source_path", "source_id", "chunk_index", "area", "level", "source_type",
+    "confidence", "period", "last_verified", "title", "claim", "url", "content_hash",
+    "embedding_model", "embedding_dim", "chunker_version", "chunk_tokens",
+    "platform", "tier", "origin_url", "path", "last_modified",
+    "public", "artefacts", "tech", "tags", "text",
+)
+
+_JSON_FIELDS = ("artefacts", "tech", "tags")
+
+
+class StoreError(RuntimeError):
+    """The corpus database is unreadable or inconsistent."""
+
+
+class Store:
+    def __init__(self, path: str) -> None:
+        self.path = path
+        os.makedirs(os.path.dirname(os.path.abspath(path)) or ".", exist_ok=True)
+        self.db = sqlite3.connect(path)
+        self.db.row_factory = sqlite3.Row
+        self.db.execute("PRAGMA journal_mode = WAL")
+        self.db.execute("PRAGMA synchronous = NORMAL")
+        self.db.executescript(SCHEMA)
+        stored = self.get_meta("schema_version")
+        # A database from an older payload shape has no provenance columns. It is
+        # derived data, so the fix is a rebuild, but it must never be silently
+        # mixed with the new shape: `rebuild_schema` is called by a rebuild only.
+        self.schema_stale = stored not in (None, SCHEMA_VERSION)
+        if not self.schema_stale:
+            self.set_meta("schema_version", SCHEMA_VERSION)
+
+    # -- lifecycle --------------------------------------------------------
+
+    def close(self) -> None:
+        self.db.close()
+
+    def __enter__(self) -> "Store":
+        return self
+
+    def __exit__(self, *_exc) -> None:
+        self.close()
+
+    def reset(self) -> None:
+        """Empty every table. Used by a from-scratch rebuild."""
+        for table in ("chunk_fts", "chunks", "manifest", "meta", "graph_fts", "graph_nodes", "graph_edges"):
+            self.db.execute(f"DELETE FROM {table}")
+        self.db.commit()
+
+    def rebuild_schema(self) -> None:
+        """Drop and recreate every table. Only a rebuild may call this."""
+        for table in (
+            "chunk_fts", "chunks", "manifest", "meta",
+            "graph_fts", "graph_nodes", "graph_edges",
+        ):
+            self.db.execute(f"DROP TABLE IF EXISTS {table}")
+        self.db.commit()
+        self.db.executescript(SCHEMA)
+        self.schema_stale = False
+        self.set_meta("schema_version", SCHEMA_VERSION)
+
+    # -- writing ----------------------------------------------------------
+
+    def write_document(self, doc, payloads: list) -> None:
+        """Replace all rows for one document: chunks, FTS rows and manifest."""
+        self.delete_document(doc.path, keep_manifest=True)
+        for payload in payloads:
+            values = []
+            for column in _CHUNK_COLUMNS:
+                value = payload[column]
+                if column in _JSON_FIELDS:
+                    value = json.dumps(value, sort_keys=True)
+                values.append(value)
+            cursor = self.db.execute(
+                f"INSERT INTO chunks ({', '.join(_CHUNK_COLUMNS)}) "
+                f"VALUES ({', '.join('?' for _ in _CHUNK_COLUMNS)})",
+                values,
+            )
+            # `terms` is the keyword surface: the words a visitor or a recruiter
+            # would type that live in front-matter rather than the body — tech
+            # names, tags, the one-line claim. Without it, a search for "zigbee"
+            # misses a page whose body says "mesh". It is indexed, never retrieved.
+            terms = " ".join(
+                str(value)
+                for value in (
+                    payload["claim"],
+                    payload["title"],
+                    " ".join(payload["tech"]),
+                    " ".join(payload["tags"]),
+                )
+                if value
+            )
+            self.db.execute(
+                "INSERT INTO chunk_fts "
+                "(rowid, id, source_path, title, terms, text, area, confidence, platform, tier) "
+                "VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+                (
+                    cursor.lastrowid,
+                    payload["id"],
+                    payload["source_path"],
+                    payload["title"],
+                    terms,
+                    payload["text"],
+                    payload["area"],
+                    payload["confidence"],
+                    payload["platform"],
+                    payload["tier"],
+                ),
+            )
+        self.db.execute(
+            "INSERT OR REPLACE INTO manifest "
+            "(source_path, content_hash, chunk_count, embedding_model, embedding_dim, "
+            " chunker_version, last_verified, updated_at) VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
+            (
+                doc.path,
+                doc.digest,
+                len(payloads),
+                payloads[0]["embedding_model"] if payloads else "",
+                payloads[0]["embedding_dim"] if payloads else 0,
+                payloads[0]["chunker_version"] if payloads else "",
+                doc.last_verified,
+                time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
+            ),
+        )
+        self.db.commit()
+
+    def delete_document(self, source_path: str, keep_manifest: bool = False) -> int:
+        rowids = [
+            row["rowid"]
+            for row in self.db.execute("SELECT rowid FROM chunks WHERE source_path = ?", (source_path,))
+        ]
+        for rowid in rowids:
+            self.db.execute("DELETE FROM chunk_fts WHERE rowid = ?", (rowid,))
+        self.db.execute("DELETE FROM chunks WHERE source_path = ?", (source_path,))
+        if not keep_manifest:
+            self.db.execute("DELETE FROM manifest WHERE source_path = ?", (source_path,))
+        self.db.commit()
+        return len(rowids)
+
+    # -- reading ----------------------------------------------------------
+
+    def manifest(self) -> dict:
+        return {
+            row["source_path"]: dict(row)
+            for row in self.db.execute("SELECT * FROM manifest")
+        }
+
+    def chunk_count(self) -> int:
+        return int(self.db.execute("SELECT count(*) FROM chunks").fetchone()[0])
+
+    def source_count(self) -> int:
+        return int(self.db.execute("SELECT count(DISTINCT source_path) FROM chunks").fetchone()[0])
+
+    def chunk_ids(self) -> list:
+        return [row["id"] for row in self.db.execute("SELECT id FROM chunks ORDER BY id")]
+
+    def chunk_payloads(self, source_path: str = None) -> list:
+        """Every stored payload, optionally for one document. Used by verify()."""
+        sql = "SELECT * FROM chunks"
+        args = ()
+        if source_path:
+            sql += " WHERE source_path = ?"
+            args = (source_path,)
+        sql += " ORDER BY source_path, chunk_index"
+        out = []
+        for row in self.db.execute(sql, args):
+            payload = dict(row)
+            for field in _JSON_FIELDS:
+                payload[field] = json.loads(payload[field])
+            out.append(payload)
+        return out
+
+    def keyword_search(
+        self, query: str, area: str = None, confidence: str = None,
+        platform: str = None, tier: int = None, limit: int = 10,
+    ) -> list:
+        """BM25 over chunk text, optionally filtered before ranking."""
+        match = match_expression(query)
+        if not match:
+            return []
+        sql = [
+            "SELECT c.id, c.url, c.title, c.area, c.confidence, c.source_path,",
+            "       c.chunk_index, snippet(chunk_fts, 4, '<mark>', '</mark>', '…', 18) AS snippet,",
+            "       bm25(chunk_fts) AS score",
+            "FROM chunk_fts JOIN chunks c ON c.rowid = chunk_fts.rowid",
+            "WHERE chunk_fts MATCH ?",
+        ]
+        args = [match]
+        if area:
+            sql.append("AND c.area = ?")
+            args.append(area)
+        if confidence:
+            sql.append("AND c.confidence = ?")
+            args.append(confidence)
+        if platform:
+            sql.append("AND c.platform = ?")
+            args.append(platform)
+        if tier is not None:
+            sql.append("AND c.tier = ?")
+            args.append(int(tier))
+        sql.append("ORDER BY bm25(chunk_fts) LIMIT ?")
+        args.append(int(limit))
+        rows = self.db.execute(" ".join(sql), args)
+        return [dict(row) for row in rows]
+
+    def pool_texts(self) -> list:
+        """Every chunk text in the pool. The credential scan reads this."""
+        return [
+            (row["source_path"], row["chunk_index"], row["text"])
+            for row in self.db.execute("SELECT source_path, chunk_index, text FROM chunks")
+        ]
+
+    # -- graph ------------------------------------------------------------
+
+    def write_graph(self, nodes: list, edges: list) -> None:
+        """Replace the whole graph. Nodes and edges are rebuilt together."""
+        self.db.execute("DELETE FROM graph_fts")
+        self.db.execute("DELETE FROM graph_nodes")
+        self.db.execute("DELETE FROM graph_edges")
+        for node in nodes:
+            self.db.execute(
+                "INSERT INTO graph_nodes (id, label, norm_label, node_type, community, "
+                "source_note, origin_url, area, tier, platform, last_modified, content_hash, text) "
+                "VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+                (
+                    node["id"], node["label"], node.get("norm_label") or _norm(node["label"]),
+                    node.get("node_type") or "entity", node.get("community"),
+                    node["source_note"], node["origin_url"], node["area"], node["tier"],
+                    node["platform"], node["last_modified"], node["content_hash"], node["text"],
+                ),
+            )
+            self.db.execute(
+                "INSERT INTO graph_fts (id, kind, label, text, area) VALUES (?, 'node', ?, ?, ?)",
+                (node["id"], node["label"], node["text"], node["area"]),
+            )
+        for edge in edges:
+            self.db.execute(
+                "INSERT INTO graph_edges (id, source_id, target_id, relation, confidence, "
+                "confidence_score, source_note, origin_url, area, tier, platform, "
+                "last_modified, content_hash, text) "
+                "VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+                (
+                    edge["id"], edge["source_id"], edge["target_id"], edge["relation"],
+                    edge["confidence"], edge.get("confidence_score"), edge["source_note"],
+                    edge["origin_url"], edge["area"], edge["tier"], edge["platform"],
+                    edge["last_modified"], edge["content_hash"], edge["text"],
+                ),
+            )
+            self.db.execute(
+                "INSERT INTO graph_fts (id, kind, label, text, area) VALUES (?, 'edge', ?, ?, ?)",
+                (edge["id"], edge["relation"], edge["text"], edge["area"]),
+            )
+        self.db.commit()
+
+    def graph_counts(self) -> dict:
+        return {
+            "nodes": int(self.db.execute("SELECT count(*) FROM graph_nodes").fetchone()[0]),
+            "edges": int(self.db.execute("SELECT count(*) FROM graph_edges").fetchone()[0]),
+        }
+
+    def graph_node(self, node_id: str) -> dict:
+        row = self.db.execute("SELECT * FROM graph_nodes WHERE id = ?", (node_id,)).fetchone()
+        return dict(row) if row else {}
+
+    def graph_neighbours(self, node_id: str, limit: int = 50) -> list:
+        rows = self.db.execute(
+            "SELECT e.*, sn.label AS source_label, tn.label AS target_label "
+            "FROM graph_edges e "
+            "LEFT JOIN graph_nodes sn ON sn.id = e.source_id "
+            "LEFT JOIN graph_nodes tn ON tn.id = e.target_id "
+            "WHERE e.source_id = ? OR e.target_id = ? LIMIT ?",
+            (node_id, node_id, int(limit)),
+        )
+        return [dict(row) for row in rows]
+
+    def graph_search(self, query: str, area: str = None, limit: int = 10) -> list:
+        """Keyword search over node labels and edge relations, with provenance."""
+        match = match_expression(query)
+        if not match:
+            return []
+        sql = [
+            "SELECT f.id, f.kind, f.label, f.text, f.area,",
+            "       COALESCE(n.source_note, e.source_note) AS source_note,",
+            "       COALESCE(n.origin_url, e.origin_url) AS origin_url,",
+            "       n.community, e.relation, e.source_id, e.target_id",
+            "FROM graph_fts f",
+            "LEFT JOIN graph_nodes n ON f.kind = 'node' AND n.id = f.id",
+            "LEFT JOIN graph_edges e ON f.kind = 'edge' AND e.id = f.id",
+            "WHERE graph_fts MATCH ?",
+        ]
+        args = [match]
+        if area:
+            sql.append("AND f.area = ?")
+            args.append(area)
+        sql.append("ORDER BY bm25(graph_fts) LIMIT ?")
+        args.append(int(limit))
+        return [dict(row) for row in self.db.execute(" ".join(sql), args)]
+
+    # -- meta -------------------------------------------------------------
+
+    def set_meta(self, key: str, value) -> None:
+        self.db.execute(
+            "INSERT OR REPLACE INTO meta (key, value) VALUES (?, ?)",
+            (key, json.dumps(value, sort_keys=True)),
+        )
+        self.db.commit()
+
+    def get_meta(self, key: str, default=None):
+        row = self.db.execute("SELECT value FROM meta WHERE key = ?", (key,)).fetchone()
+        if row is None:
+            return default
+        return json.loads(row["value"])
+
+    def meta(self) -> dict:
+        return {row["key"]: json.loads(row["value"]) for row in self.db.execute("SELECT key, value FROM meta")}
+
+
+def _norm(label: str) -> str:
+    return re.sub(r"[^0-9a-z]+", "_", (label or "").lower()).strip("_")
+
+
+def match_expression(query: str) -> str:
+    """Free text -> a safe FTS5 expression: bare words, each a prefix, all required.
+
+    Mirrors `MatchExpression` in internal/search/search.go so the site's search
+    box and the retrieval path treat a visitor's words identically. FTS5
+    operators in the input are never honoured.
+    """
+    terms = []
+    for raw in re.split(r"[^0-9a-zA-Z\-]+", unicodedata.normalize("NFKC", query or "").lower()):
+        term = raw.strip("-")
+        if term:
+            terms.append('"' + term.replace('"', "") + '"*')
+    return " AND ".join(terms)

corpus/tests/fake_qdrant.py

@@ -0,0 +1,112 @@
+"""An in-memory stand-in for Qdrant, used by the pipeline tests.
+
+It implements the routes the pipeline actually calls and nothing else, and it is
+deliberately dumb: brute-force cosine over a dict. If the pipeline passes
+against this, the pipeline's own logic is right; the real server is exercised
+separately in `docs/phase-4-notes.md` (rebuild against a live container).
+"""
+
+from __future__ import annotations
+
+import math
+
+from corpus.qdrant import Qdrant, QdrantError, QdrantNotFound
+
+
+class FakeQdrant(Qdrant):
+    def __init__(self, dim: int):
+        super().__init__(url="fake://qdrant", timeout=1.0, attempts=1)
+        self.dim = dim
+        self.collections = {}  # name -> {id: {"vector": [...], "payload": {...}}}
+        self.calls = []
+
+    # -- transport --------------------------------------------------------
+
+    def _request(self, method: str, path: str, body=None) -> dict:
+        self.calls.append((method, path, body))
+        query = ""
+        if "?" in path:
+            path, query = path.split("?", 1)
+        parts = [part for part in path.split("/") if part]
+
+        if not parts:
+            return {"version": "fake-1.0"}
+
+        if parts[0] != "collections":
+            raise QdrantNotFound(f"no route {path}")
+        if len(parts) == 2:
+            name = parts[1]
+            if method == "PUT":
+                self.collections.setdefault(name, {})
+                return {"result": True}
+            if method == "DELETE":
+                if name not in self.collections:
+                    raise QdrantNotFound("collection not found")
+                del self.collections[name]
+                return {"result": True}
+            if name not in self.collections:
+                raise QdrantNotFound("collection not found")
+            return {
+                "result": {
+                    "points_count": len(self.collections[name]),
+                    "config": {"params": {"vectors": {"size": self.dim, "distance": "Cosine"}}},
+                }
+            }
+        name = parts[1]
+        if name not in self.collections:
+            raise QdrantNotFound("collection not found")
+        points = self.collections[name]
+
+        if parts[2:4] == ["points", "count"]:
+            return {"result": {"count": len(points)}}
+        if len(parts) == 3 and parts[2] == "points" and method == "PUT":
+            for point in (body or {}).get("points", []):
+                if len(point["vector"]) != self.dim:
+                    raise QdrantError("wrong vector dimension")
+                points[point["id"]] = {"vector": point["vector"], "payload": point["payload"]}
+            return {"result": {"status": "completed"}}
+        if parts[2:4] == ["points", "delete"]:
+            must = (((body or {}).get("filter") or {}).get("must")) or []
+            for point_id in list(points):
+                if _matches(points[point_id]["payload"], must):
+                    del points[point_id]
+            return {"result": {"status": "completed"}}
+        if parts[2:4] == ["points", "query"]:
+            vector = body["query"]
+            must = (((body or {}).get("filter") or {}).get("must")) or []
+            scored = []
+            for point_id, point in points.items():
+                if not _matches(point["payload"], must):
+                    continue
+                scored.append(
+                    {
+                        "id": point_id,
+                        "score": _cosine(vector, point["vector"]),
+                        "payload": point["payload"],
+                    }
+                )
+            scored.sort(key=lambda row: row["score"], reverse=True)
+            return {"result": {"points": scored[: int(body.get("limit", 10))]}}
+        if parts[2:4] == ["points", "scroll"]:
+            limit = int((body or {}).get("limit", 100))
+            rows = [
+                {"id": point_id, "payload": {"source_path": point["payload"].get("source_path")}}
+                for point_id, point in list(points.items())[:limit]
+            ]
+            return {"result": {"points": rows, "next_page_offset": None}}
+        raise QdrantNotFound(f"no route {path}")
+
+
+def _matches(payload: dict, conditions: list) -> bool:
+    for condition in conditions:
+        key = condition.get("key")
+        expected = (condition.get("match") or {}).get("value")
+        if payload.get(key) != expected:
+            return False
+    return True
+
+
+def _cosine(left: list, right: list) -> float:
+    dot = sum(a * b for a, b in zip(left, right))
+    norm = math.sqrt(sum(a * a for a in left)) * math.sqrt(sum(b * b for b in right))
+    return dot / norm if norm else 0.0

corpus/tests/support.py

@@ -0,0 +1,188 @@
+"""Shared fixtures for the corpus tests: a throwaway content tree and a fake Qdrant."""
+
+from __future__ import annotations
+
+import contextlib
+import os
+import shutil
+import tempfile
+from unittest import mock
+
+from corpus.config import Config
+from corpus.tests.fake_qdrant import FakeQdrant
+
+REPO_ROOT = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+DOC = """---
+id: {id}
+area: {area}
+level: {level}
+title: "{title}"
+claim: "{claim}"
+confidence: {confidence}
+period: "2023–"
+tech: [Go, SQLite]
+metrics: [{{label: endpoints, value: 8}}]
+artefacts: []
+public: true
+last_verified: {last_verified}
+tags: [{tags}]
+---
+
+# {title}
+
+{body}
+"""
+
+FENCE = """A paragraph before the code block explains the constraint.
+
+```go
+package main
+
+func main() {{
+	println("chunk boundary must never land inside this fence")
+}}
+```
+
+A paragraph after the code block.
+"""
+
+TABLE = """Measured results, one row per endpoint:
+
+| endpoint | latency | loss |
+|---|---|---|
+| alpha | 12ms | 0.0% |
+| beta | 14ms | 0.1% |
+| gamma | 9ms | 0.0% |
+
+And a closing paragraph.
+"""
+
+PROSE = "\n\n".join(
+    f"Paragraph {index} describes a delivery constraint in corporate language, "
+    "with enough words that the chunker has to think about where to break, and "
+    "never inside a fenced block, a table or a front-matter block."
+    for index in range(1, 40)
+)
+
+DEFAULT_DOCS = {
+    "ai/ai-overview.md": dict(
+        id="ai-overview", area="ai", level=1, title="Agent Systems",
+        claim="Harnesses, memory and routing.", confidence="running",
+        last_verified="2026-10-01", tags="agents, llm", body=PROSE,
+    ),
+    "edge/l3/edge-zigbee.md": dict(
+        id="edge-zigbee", area="edge", level=3, title="Low Power Mesh Sensing",
+        claim="A constrained mesh that survives power loss.", confidence="delivered",
+        last_verified="2026-09-15", tags="zigbee, mesh", body=FENCE,
+    ),
+    "platform/l3/platform-containers.md": dict(
+        id="platform-containers", area="platform", level=3, title="Container Estate",
+        claim="Fifty containers with measured resource limits.", confidence="in-progress",
+        last_verified="2026-08-01", tags="docker", body=TABLE,
+    ),
+}
+
+
+def temp_repo(docs: dict = None) -> str:
+    """Create a throwaway repository root with tools/ and content/."""
+    root = tempfile.mkdtemp(prefix="corpus-test-")
+    os.makedirs(os.path.join(root, "tools"), exist_ok=True)
+    shutil.copy(
+        os.path.join(REPO_ROOT, "tools", "validate_content.py"),
+        os.path.join(root, "tools", "validate_content.py"),
+    )
+    write_docs(root, docs or DEFAULT_DOCS)
+    return root
+
+
+def write_docs(root: str, docs: dict) -> None:
+    for rel, fields in docs.items():
+        path = os.path.join(root, "content", rel)
+        os.makedirs(os.path.dirname(path), exist_ok=True)
+        with open(path, "w", encoding="utf-8") as handle:
+            handle.write(DOC.format(**fields))
+
+
+def write_doc(root: str, rel: str, **fields) -> None:
+    path = os.path.join(root, "content", rel)
+    os.makedirs(os.path.dirname(path), exist_ok=True)
+    with open(path, "w", encoding="utf-8") as handle:
+        handle.write(DOC.format(**fields))
+    # A fresh interpreter's import cache is fine; the loader re-reads the file.
+
+
+def make_config(root: str, **overrides) -> Config:
+    values = dict(
+        provider="local-hashing",
+        model="local-hashing-v1",
+        dim=32,
+        qdrant_url="fake://qdrant",
+        collection="test_corpus",
+        graph_collection="test_graph",
+        offline=True,
+    )
+    values.update(overrides)
+    return Config.from_env(root=root, **values)
+
+
+@contextlib.contextmanager
+def fake_qdrant(cfg, fake: FakeQdrant = None):
+    """Point every module that builds a Qdrant client at one in-memory fake."""
+    fake = fake or FakeQdrant(cfg.dim)
+    with mock.patch("corpus.pipeline.Qdrant", return_value=fake), mock.patch(
+        "corpus.freshness.Qdrant", return_value=fake
+    ):
+        yield fake
+
+
+class StubServer:
+    """A throwaway HTTP server on a loopback ephemeral port, for provider tests.
+
+    `responder(method, path, body)` returns `(status, payload)`; every request is
+    recorded with its parsed body so a test can assert the wire shape without a
+    real provider or a real Qdrant.
+    """
+
+    def __init__(self, responder):
+        import json
+        import threading
+        from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
+
+        self.requests = []
+        self._responder = responder
+        outer = self
+
+        class Handler(BaseHTTPRequestHandler):
+            def _handle(self):
+                length = int(self.headers.get("Content-Length") or 0)
+                raw = self.rfile.read(length).decode("utf-8") if length else ""
+                body = json.loads(raw) if raw.strip() else None
+                outer.requests.append(
+                    {"method": self.command, "path": self.path, "body": body,
+                     "headers": dict(self.headers)}
+                )
+                status, payload = outer._responder(self.command, self.path, body)
+                data = payload.encode("utf-8") if isinstance(payload, str) else json.dumps(payload).encode("utf-8")
+                self.send_response(status)
+                self.send_header("Content-Type", "application/json")
+                self.send_header("Content-Length", str(len(data)))
+                self.end_headers()
+                self.wfile.write(data)
+
+            do_GET = do_POST = do_PUT = do_DELETE = _handle
+
+            def log_message(self, *_args):
+                pass
+
+        self.server = ThreadingHTTPServer(("127.0.0.1", 0), Handler)
+        self.url = f"http://127.0.0.1:{self.server.server_address[1]}"
+        self._thread = threading.Thread(target=self.server.serve_forever, daemon=True)
+
+    def __enter__(self):
+        self._thread.start()
+        return self
+
+    def __exit__(self, *_exc):
+        self.server.shutdown()
+        self.server.server_close()

corpus/tests/test_chunker.py

@@ -0,0 +1,145 @@
+"""Chunking rules: size, overlap, determinism, and the hard one — no split inside
+a fence, a table or front-matter (docs/phase-4-corpus.md DoD item 3)."""
+
+from __future__ import annotations
+
+import os
+import unittest
+
+from corpus import chunker, loader
+from corpus.config import Config
+from corpus.tests import support
+
+REPO_ROOT = support.REPO_ROOT
+
+
+def content_documents():
+    cfg = Config.from_env(root=REPO_ROOT, provider="local-hashing", dim=32)
+    return loader.load_documents(cfg)
+
+
+class BlockParsingTest(unittest.TestCase):
+    def test_fence_is_one_block(self):
+        blocks = chunker.parse_blocks(support.FENCE)
+        kinds = [block.kind for block in blocks]
+        self.assertEqual(kinds, ["para", "fence", "para"])
+        self.assertIn("chunk boundary must never land", blocks[1].text)
+
+    def test_table_is_one_block(self):
+        blocks = chunker.parse_blocks(support.TABLE)
+        table = [block for block in blocks if block.kind == "table"]
+        self.assertEqual(len(table), 1)
+        self.assertEqual(len(table[0].text.splitlines()), 5)  # header, rule, 3 rows
+
+    def test_front_matter_in_body_is_refused(self):
+        with self.assertRaises(chunker.ChunkError):
+            chunker.parse_blocks("---\nid: x\n---\n\n# Heading\n")
+
+    def test_unterminated_fence_runs_to_end(self):
+        blocks = chunker.parse_blocks("intro\n\n```go\nfunc main() {}\n")
+        self.assertEqual(blocks[-1].kind, "fence")
+        self.assertTrue(blocks[-1].text.rstrip().endswith("}"))
+
+    def test_headings_start_blocks(self):
+        blocks = chunker.parse_blocks("# One\n\ntext\n\n## Two\n\nmore\n")
+        self.assertEqual([block.kind for block in blocks], ["heading", "para", "heading", "para"])
+
+    def test_list_items_are_separate_blocks(self):
+        blocks = chunker.parse_blocks("- one\n- two\n  continued\n- three\n")
+        self.assertEqual([block.kind for block in blocks], ["list", "list", "list"])
+        self.assertIn("continued", blocks[1].text)
+
+
+class ChunkingTest(unittest.TestCase):
+    def test_small_body_is_one_chunk(self):
+        chunks = chunker.chunk_document("# Title\n\nA short body.\n")
+        self.assertEqual(len(chunks), 1)
+        self.assertEqual(chunks[0].index, 0)
+        self.assertFalse(chunks[0].oversize)
+
+    def test_long_body_respects_maximum_and_overlaps(self):
+        chunks = chunker.chunk_document(support.PROSE)
+        self.assertGreater(len(chunks), 1)
+        for chunk in chunks:
+            self.assertLessEqual(chunk.tokens, chunker.TOKEN_MAX, f"chunk {chunk.index} too large")
+        # Overlap: consecutive chunks share at least one block.
+        for previous, current in zip(chunks, chunks[1:]):
+            self.assertTrue(
+                set(id(block) for block in previous.blocks) & set(id(block) for block in current.blocks),
+                f"chunks {previous.index} and {current.index} do not overlap",
+            )
+
+    def test_chunk_indices_are_sequential(self):
+        chunks = chunker.chunk_document(support.PROSE)
+        self.assertEqual([chunk.index for chunk in chunks], list(range(len(chunks))))
+
+    def test_deterministic(self):
+        first = chunker.chunk_document(support.PROSE)
+        second = chunker.chunk_document(support.PROSE)
+        self.assertEqual([chunk.text for chunk in first], [chunk.text for chunk in second])
+
+    def test_code_fence_is_never_split(self):
+        body = support.PROSE + "\n\n" + support.FENCE + "\n\n" + support.PROSE
+        chunks = chunker.chunk_document(body)
+        fences = [block for block in chunker.parse_blocks(body) if block.kind == "fence"]
+        self.assertTrue(fences)
+        for fence in fences:
+            holders = [chunk for chunk in chunks if fence in chunk.blocks]
+            self.assertTrue(holders, "the fenced block disappeared from every chunk")
+            for holder in holders:
+                self.assertIn(fence.text, holder.text, "a chunk holds part of a fenced block")
+
+    def test_oversize_single_block_is_flagged_not_split(self):
+        huge = "```go\n" + "\n".join(f"line{i} := {i}" for i in range(1200)) + "\n```\n"
+        chunks = chunker.chunk_document(huge)
+        self.assertEqual(len(chunks), 1)
+        self.assertTrue(chunks[0].oversize)
+        self.assertGreater(chunks[0].tokens, chunker.TOKEN_MAX)
+
+
+class RealContentTest(unittest.TestCase):
+    """The same rules, over every document in content/."""
+
+    def test_no_boundary_splits_a_protected_region(self):
+        for doc in content_documents():
+            regions = chunker.protected_regions(doc.body)
+            chunks = chunker.chunk_document(doc.body)
+            chunker.verify_no_split(chunks, regions)  # raises on failure
+
+    def test_every_chunk_is_whole_blocks(self):
+        for doc in content_documents():
+            expected = [block.text for block in chunker.parse_blocks(doc.body)]
+            seen = []
+            for chunk in chunker.chunk_document(doc.body):
+                for block in chunk.blocks:
+                    self.assertIn(block.text, chunk.text)
+                    if not seen or seen[-1] != block.text:
+                        seen.append(block.text)
+            # Every block appears, in order, exactly as parsed (overlap repeats a
+            # block at the start of the next chunk, which is expected).
+            position = 0
+            for text in expected:
+                while position < len(seen) and seen[position] != text:
+                    position += 1
+                self.assertLess(position, len(seen), f"{doc.path}: block was dropped: {text[:40]!r}")
+                position += 1
+
+    def test_chunks_are_within_the_size_band(self):
+        for doc in content_documents():
+            for chunk in chunker.chunk_document(doc.body):
+                if chunk.oversize:
+                    self.assertEqual(len(chunk.blocks), 1)
+                    continue
+                self.assertLessEqual(chunk.tokens, chunker.TOKEN_MAX, f"{doc.path} chunk {chunk.index}")
+
+    def test_content_tree_is_chunked_completely(self):
+        total = 0
+        for doc in content_documents():
+            chunks = chunker.chunk_document(doc.body)
+            self.assertTrue(chunks, f"{doc.path}: no chunks")
+            total += len(chunks)
+        self.assertGreater(total, 0)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_embeddings.py

@@ -0,0 +1,125 @@
+"""Providers: determinism offline, and an OpenAI-compatible wire shape through OmniRoute.
+
+There is no direct vendor SDK call anywhere (AGENTS.md §4.5), so the provider
+test asserts the HTTP request shape, not a client library's behaviour.
+"""
+
+from __future__ import annotations
+
+import math
+import unittest
+
+from corpus import embeddings
+from corpus.tests import support
+
+
+class LocalHashingTest(unittest.TestCase):
+    def test_deterministic_and_unit_length(self):
+        provider = embeddings.LocalHashingProvider(dim=64)
+        first = provider.embed(["zigbee mesh latency"])[0]
+        second = embeddings.LocalHashingProvider(dim=64).embed(["zigbee mesh latency"])[0]
+        self.assertEqual(first, second)
+        self.assertEqual(len(first), 64)
+        self.assertAlmostEqual(math.sqrt(sum(value * value for value in first)), 1.0, places=9)
+
+    def test_different_text_differs(self):
+        provider = embeddings.LocalHashingProvider(dim=64)
+        a, b = provider.embed(["zigbee", "lora"])
+        self.assertNotEqual(a, b)
+
+    def test_empty_text_still_returns_a_unit_vector(self):
+        provider = embeddings.LocalHashingProvider(dim=16)
+        vector = provider.embed([""])[0]
+        self.assertEqual(len(vector), 16)
+        self.assertAlmostEqual(math.sqrt(sum(value * value for value in vector)), 1.0, places=9)
+
+    def test_stats_count_calls_and_texts(self):
+        provider = embeddings.LocalHashingProvider(dim=16)
+        provider.embed(["a", "b", "c"])
+        provider.embed(["d"])
+        self.assertEqual(provider.stats(), {"calls": 2, "texts": 4})
+        provider.reset_stats()
+        self.assertEqual(provider.stats(), {"calls": 0, "texts": 0})
+
+    def test_rejects_a_zero_dimension(self):
+        with self.assertRaises(embeddings.EmbeddingError):
+            embeddings.LocalHashingProvider(dim=0)
+
+
+class OmniRouteTest(unittest.TestCase):
+    def responder_factory(self, dim=8, status_plan=None):
+        plan = list(status_plan or [])
+        state = {"n": 0}
+
+        def responder(method, path, body):
+            state["n"] += 1
+            if plan:
+                status = plan.pop(0)
+                if status != 200:
+                    return status, {"error": "planned failure"}
+            vectors = [
+                {"index": index, "embedding": [0.1] * dim}
+                for index, _text in enumerate(body.get("input") or [])
+            ]
+            return 200, {"model": body.get("model"), "data": vectors}
+
+        return responder
+
+    def test_posts_to_v1_embeddings_with_the_configured_model(self):
+        with support.StubServer(self.responder_factory()) as server:
+            provider = embeddings.OmniRouteProvider(
+                api_url=server.url, api_key="test-key", model="text-embedding-3-small", dim=8
+            )
+            vectors = provider.embed(["one", "two"])
+        self.assertEqual(len(vectors), 2)
+        request = server.requests[0]
+        self.assertEqual(request["method"], "POST")
+        self.assertEqual(request["path"], "/v1/embeddings")
+        self.assertEqual(request["body"]["model"], "text-embedding-3-small")
+        self.assertEqual(request["body"]["input"], ["one", "two"])
+        self.assertEqual(request["headers"]["Authorization"], "Bearer test-key")
+        self.assertEqual(provider.stats(), {"calls": 1, "texts": 2})
+
+    def test_batches_requests(self):
+        with support.StubServer(self.responder_factory()) as server:
+            provider = embeddings.OmniRouteProvider(
+                api_url=server.url, api_key="k", model="m", dim=8, batch_size=2
+            )
+            provider.embed(["a", "b", "c", "d", "e"])
+        self.assertEqual(len(server.requests), 3)
+        self.assertEqual(provider.stats(), {"calls": 3, "texts": 5})
+
+    def test_dimension_mismatch_is_refused(self):
+        with support.StubServer(self.responder_factory(dim=4)) as server:
+            provider = embeddings.OmniRouteProvider(
+                api_url=server.url, api_key="k", model="m", dim=8
+            )
+            with self.assertRaises(embeddings.EmbeddingError) as caught:
+                provider.embed(["one"])
+        self.assertIn("dimension", str(caught.exception))
+
+    def test_retries_then_succeeds(self):
+        with support.StubServer(self.responder_factory(status_plan=[429, 200])) as server:
+            provider = embeddings.OmniRouteProvider(
+                api_url=server.url, api_key="k", model="m", dim=8, backoff=0.01
+            )
+            vectors = provider.embed(["one"])
+        self.assertEqual(len(vectors), 1)
+        self.assertEqual(len(server.requests), 2)
+
+    def test_client_error_is_not_retried(self):
+        with support.StubServer(self.responder_factory(status_plan=[400])) as server:
+            provider = embeddings.OmniRouteProvider(
+                api_url=server.url, api_key="k", model="m", dim=8, backoff=0.01
+            )
+            with self.assertRaises(embeddings.EmbeddingError):
+                provider.embed(["one"])
+        self.assertEqual(len(server.requests), 1)
+
+    def test_missing_key_is_refused_before_any_call(self):
+        with self.assertRaises(embeddings.EmbeddingError):
+            embeddings.OmniRouteProvider(api_url="http://127.0.0.1:1", api_key="", model="m", dim=8)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_freshness.py

@@ -0,0 +1,110 @@
+"""Freshness audit: stale documents, index drift and contract drift."""
+
+from __future__ import annotations
+
+import datetime
+import unittest
+
+from corpus import freshness, pipeline, store as store_module
+from corpus.tests import support
+
+
+class AuditTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.cfg = support.make_config(self.root)
+        self.addCleanup(_cleanup, self.root)
+        with support.fake_qdrant(self.cfg):
+            pipeline.run(self.cfg, mode="rebuild")
+
+    def audit(self, today=None, cfg=None):
+        with support.fake_qdrant(cfg or self.cfg):
+            return freshness.audit(cfg or self.cfg, today=today)
+
+    def test_clean_soon_after_rebuild(self):
+        report = self.audit(today=datetime.date(2026, 10, 5))
+        self.assertTrue(report["ok"], report["findings"])
+        self.assertEqual(report["corpus"]["documents"], 3)
+
+    def test_reports_stale_documents(self):
+        report = self.audit(today=datetime.date(2027, 10, 5))
+        stale = {row["source_path"] for row in report["findings"]["stale"]}
+        self.assertEqual(
+            stale,
+            {
+                "content/ai/ai-overview.md",
+                "content/edge/l3/edge-zigbee.md",
+                "content/platform/l3/platform-containers.md",
+            },
+        )
+        self.assertFalse(report["ok"])
+        text = freshness.render_text(report)
+        self.assertIn("STALE", text)
+        self.assertIn("platform-containers", text)
+
+    def test_warns_before_a_document_goes_stale(self):
+        # platform-containers was verified 2026-08-01. Horizon 180 days, warning
+        # 30, so on 2027-01-20 it is 172 days old: due soon, not yet stale.
+        report = self.audit(today=datetime.date(2027, 1, 20))
+        due = {row["source_path"] for row in report["findings"]["due_soon"]}
+        self.assertEqual(due, {"content/platform/l3/platform-containers.md"})
+        self.assertEqual(report["findings"]["stale"], [])
+        self.assertTrue(report["ok"])
+
+    def test_reports_index_drift_when_a_source_changes(self):
+        support.write_doc(
+            self.root,
+            "ai/ai-overview.md",
+            id="ai-overview", area="ai", level=1, title="Agent Systems",
+            claim="Edited but not re-indexed.", confidence="running",
+            last_verified="2026-10-05", tags="agents", body="New words for the corpus.",
+        )
+        report = self.audit(today=datetime.date(2026, 10, 6))
+        problems = [row["problem"] for row in report["findings"]["index_drift"]]
+        self.assertIn("source changed since it was embedded", problems)
+        self.assertFalse(report["ok"])
+
+    def test_reports_contract_drift_when_the_model_changes(self):
+        other = support.make_config(self.root, model="different-model")
+        report = self.audit(today=datetime.date(2026, 10, 5), cfg=other)
+        self.assertTrue(report["findings"]["contract_drift"])
+
+    def test_report_is_json_serialisable(self):
+        import json
+
+        report = self.audit(today=datetime.date(2026, 10, 5))
+        json.dumps(report)
+
+
+class AlertTest(unittest.TestCase):
+    def test_no_url_means_no_alert(self):
+        report = {"ok": True, "findings": {"stale": []}, "corpus": {}, "horizon": {}, "summary": {}}
+        self.assertFalse(freshness.alert(report, ""))
+
+    def test_posts_to_apprise(self):
+        report = {
+            "generated_at": "2026-10-26T00:00:00Z",
+            "today": "2026-10-26",
+            "horizon": {"max_age_days": 180, "warn_days": 30},
+            "corpus": {
+                "documents": 1, "chunks": 1, "sources": 1, "collection": "c",
+                "collection_points": 1, "provider": "p", "model": "m", "dim": 8,
+                "chunker_version": "v",
+            },
+            "findings": {"stale": [], "due_soon": [], "index_drift": [], "contract_drift": []},
+            "summary": {"stale": 0},
+            "ok": True,
+        }
+        with support.StubServer(lambda method, path, body: (200, "ok")) as server:
+            self.assertTrue(freshness.alert(report, server.url))
+        self.assertIn("title", server.requests[0]["body"])
+
+
+def _cleanup(root):
+    import shutil
+
+    shutil.rmtree(root, ignore_errors=True)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_ingestion.py

@@ -0,0 +1,346 @@
+"""Ingestion from the estate: exclusions, provenance, incremental work, graph.
+
+These tests never touch the network. `FakeGitea` implements the four read calls
+the harvester uses; `FakeQdrant` implements the collection routes the pipeline
+uses. The real estate is exercised separately (docs/phase-4-notes.md), so a
+failure here is a logic failure, not a connectivity one.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import os
+import unittest
+from unittest import mock
+
+from corpus import exclusions, guides, loader, pipeline, store as store_module
+from corpus.sources import estate, graph as graph_module
+from corpus.tests import support
+
+
+def gitea_guide(**overrides) -> guides.Guide:
+    values = dict(
+        name="gitea", enabled=True, what="Self-hosted Git.", fits=["ai", "platform", "edge", "products"],
+        read=["readme", "tree", "selected_source", "commits", "languages"],
+        base="https://gitea.example", auth="GITEA_TOKEN",
+        include={"repos": "all", "paths": []},
+        exclude={"repos": [], "paths": []},
+        limits={"max_file_kb": 200, "max_files_per_repo": 10, "max_commits": 5, "max_tree_entries": 100},
+    )
+    values.update(overrides)
+    return guides.Guide(**values)
+
+
+class FakeGitea:
+    def __init__(self, repos=None, trees=None, files=None, commits=None, languages=None):
+        self.repos_list = repos or []
+        self.trees = trees or {}
+        self.files = files or {}
+        self.commit_list = commits or []
+        self.language_map = languages or {}
+        self.calls = []
+
+    def repos(self):
+        return list(self.repos_list)
+
+    def tree(self, full_name, ref):
+        self.calls.append(("tree", full_name, ref))
+        return list(self.trees.get(full_name, []))
+
+    def raw(self, full_name, ref, path):
+        self.calls.append(("raw", full_name, ref, path))
+        try:
+            return self.files[(full_name, path)]
+        except KeyError:
+            from corpus.sources.gitea import GiteaNotFound
+
+            raise GiteaNotFound(f"{full_name}/{path}") from None
+
+    def commits(self, full_name, ref, limit=20):
+        return list(self.commit_list)
+
+    def languages(self, full_name):
+        return dict(self.language_map)
+
+
+def repo(full_name="sam/photo-pipeline", **overrides):
+    values = dict(
+        full_name=full_name, name=full_name.split("/")[-1], description="Photo ingestion pipeline",
+        topics=["photo", "python"], default_branch="main", archived=False,
+        created_at="2024-01-01T00:00:00Z", updated_at="2026-10-14T05:00:00Z",
+        html_url=f"https://gitea.example/{full_name}",
+    )
+    values.update(overrides)
+    return values
+
+
+def blob(path, size=None):
+    return {"type": "blob", "path": path, "size": len(path) if size is None else size}
+
+
+class ExclusionTest(unittest.TestCase):
+    """The DoD probe: a file in an excluded path never appears in the pool."""
+
+    def setUp(self):
+        self.guide = gitea_guide()
+        self.result = estate.Harvest()
+        tree = [
+            blob("README.md"),
+            blob("node_modules/left-pad/probe.md"),
+            blob(".venv/lib/probe.py"),
+            blob(".git/config"),
+            blob(".trash/old.md"),
+            blob("docs/.silverbullet.auth.json"),
+            blob("src/main.py"),
+            blob("assets/photo.png"),
+        ]
+        files = {
+            ("sam/photo-pipeline", "README.md"): b"# Photo pipeline\n\nDeduplicates by hash.\n",
+            ("sam/photo-pipeline", "node_modules/left-pad/probe.md"): b"# PROBE\n",
+            ("sam/photo-pipeline", ".venv/lib/probe.py"): b"PROBE = 1\n",
+            ("sam/photo-pipeline", ".git/config"): b"[core]\n",
+            ("sam/photo-pipeline", ".trash/old.md"): b"# old\n",
+            ("sam/photo-pipeline", "docs/.silverbullet.auth.json"): b'{"secret_key": "x"}\n',
+            ("sam/photo-pipeline", "src/main.py"): b"def dedup():\n    return True\n",
+        }
+        self.client = FakeGitea(
+            repos=[repo()],
+            trees={"sam/photo-pipeline": tree},
+            files=files,
+            commits=[{"commit": {"message": "dedup by hash", "author": {"date": "2026-10-14T05:00:00Z"}}}],
+            languages={"Python": 900, "Shell": 100},
+        )
+        self.documents = estate._harvest_one_repo(
+            self.result, self.client, self.guide, repo(), "https://gitea.example"
+        )
+
+    def test_probe_in_excluded_paths_never_appears(self):
+        paths = {doc.repo_path for doc in self.documents}
+        for probe in (
+            "node_modules/left-pad/probe.md",
+            ".venv/lib/probe.py",
+            ".git/config",
+            ".trash/old.md",
+            "docs/.silverbullet.auth.json",
+        ):
+            self.assertNotIn(probe, paths)
+        self.assertIn("README.md", paths)
+        self.assertIn("src/main.py", paths)
+
+    def test_rejections_are_reported_with_a_reason(self):
+        reasons = {(row["path"], row["reason"]) for row in self.result.rejections}
+        self.assertIn(("node_modules/left-pad/probe.md", "excluded directory 'node_modules'"), reasons)
+        self.assertIn(("docs/.silverbullet.auth.json", "credential-shaped file name"), reasons)
+        self.assertEqual([image["path"] for image in self.result.images], ["assets/photo.png"])
+
+    def test_media_is_an_image_reference_not_text(self):
+        self.assertEqual([image["path"] for image in self.result.images], ["assets/photo.png"])
+        for doc in self.documents:
+            self.assertNotIn("assets/photo.png", doc.body)
+
+    def test_every_document_carries_provenance(self):
+        for doc in self.documents:
+            self.assertEqual(doc.platform, "gitea")
+            self.assertEqual(doc.tier, 2)
+            self.assertTrue(doc.origin_url.startswith("https://gitea.example/sam/photo-pipeline"))
+            self.assertTrue(doc.last_modified)
+            self.assertIn(doc.area, self.guide.fits)
+
+
+class CredentialContentTest(unittest.TestCase):
+    def test_a_valid_path_with_a_private_key_is_rejected(self):
+        result = estate.Harvest()
+        guide = gitea_guide()
+        client = FakeGitea(
+            repos=[repo()],
+            trees={"sam/photo-pipeline": [blob("README.md"), blob("src/leak.py")]},
+            files={
+                ("sam/photo-pipeline", "README.md"): b"# Photo pipeline\n",
+                ("sam/photo-pipeline", "src/leak.py"): (
+                    ("KEY = '" + "-----BEGIN " + "PRIVATE KEY-----'\n").encode("utf-8")
+                ),
+            },
+            commits=[],
+            languages={"Python": 10},
+        )
+        estate._harvest_one_repo(result, client, guide, repo(), "https://gitea.example")
+        reasons = {row["reason"]: row for row in result.rejections}
+        self.assertIn("credential-shaped content", reasons)
+        self.assertEqual(reasons["credential-shaped content"]["path"], "src/leak.py")
+
+
+class PayloadProvenanceTest(unittest.TestCase):
+    def test_missing_provenance_field_fails_the_payload(self):
+        root = support.temp_repo()
+        self.addCleanup(lambda: __import__("shutil").rmtree(root, ignore_errors=True))
+        docs = loader.load_documents(support.make_config(root))
+        chunk = __import__("corpus.chunker", fromlist=["chunk_document"]).chunk_document(docs[0].body)[0]
+        payload = __import__("corpus.payloads", fromlist=["build_payload"]).build_payload(
+            docs[0], chunk, support.make_config(root)
+        )
+        payload["id"] = pipeline.chunk_id(docs[0].path, 0, docs[0].digest)
+        for field in ("platform", "tier", "origin_url", "path", "last_modified"):
+            broken = dict(payload)
+            broken.pop(field)
+            with self.assertRaises(Exception):
+                __import__("corpus.payloads", fromlist=["validate_payload"]).validate_payload(broken)
+
+
+GRAPH = {
+    "built_at_commit": "abc123",
+    "nodes": [
+        {"id": "photo_ingest", "label": "Photo ingest", "source_file": "300 areas/Photo.md",
+         "community": 7, "node_type": "entity"},
+        {"id": "settings", "label": "settings.json", "source_file": "settings.json",
+         "community": 118, "node_type": "code"},
+        {"id": "image_node", "label": "diagram", "source_file": "300 areas/diagram.png",
+         "community": 7, "node_type": "image"},
+    ],
+    "links": [
+        {"source": "photo_ingest", "target": "settings", "relation": "configured_by",
+         "source_file": "300 areas/Photo.md", "confidence": "EXTRACTED", "confidence_score": 0.9},
+        {"source": "settings", "target": "photo_ingest", "relation": "configures",
+         "source_file": "settings.json", "confidence": "INFERRED"},
+    ],
+}
+
+
+class GraphTest(unittest.TestCase):
+    def parse(self):
+        return graph_module.parse(
+            json.dumps(GRAPH).encode("utf-8"),
+            platform="graphify", base="https://gitea.example", repo="sam/obsidian-vault",
+            ref="main", last_modified="2026-10-14T05:00:00Z", area="ai",
+            vault_paths={"300 areas/Photo.md", "300 areas/diagram.png"},
+        )
+
+    def test_nodes_and_edges_carry_their_source_note(self):
+        result = self.parse()
+        self.assertEqual(len(result.nodes), 1)
+        self.assertEqual(result.nodes[0]["id"], "photo_ingest")
+        self.assertEqual(result.nodes[0]["source_note"], "300 areas/Photo.md")
+        self.assertIn("/sam/obsidian-vault/src/branch/main/300%20areas/Photo.md", result.nodes[0]["origin_url"])
+        self.assertTrue(result.edges)
+        for edge in result.edges:
+            self.assertTrue(edge["source_note"])
+
+    def test_a_node_without_a_resolvable_note_is_refused(self):
+        result = self.parse()
+        rejected = {row["path"] for row in result.rejections}
+        self.assertIn("settings.json", rejected)
+        self.assertNotIn("settings", {row["id"] for row in result.nodes})
+
+    def test_media_becomes_an_image_reference(self):
+        result = self.parse()
+        self.assertEqual([image["path"] for image in result.images], ["300 areas/diagram.png"])
+        self.assertNotIn("image_node", {row["id"] for row in result.nodes})
+
+    def test_graph_query_returns_the_note(self):
+        root = support.temp_repo()
+        self.addCleanup(lambda: __import__("shutil").rmtree(root, ignore_errors=True))
+        cfg = support.make_config(root)
+        harvest = estate.Harvest()
+        harvest.documents = loader.load_documents(cfg)
+        harvest.graph = self.parse()
+        harvest.sources["graphify"] = {"platform": "graphify", "enabled": True, "documents": 0}
+        fake = support.FakeQdrant(cfg.dim)
+        with mock.patch("corpus.sources.estate.harvest", return_value=harvest):
+            with support.fake_qdrant(cfg, fake):
+                pipeline.run(cfg, mode="rebuild")
+                with store_module.Store(cfg.db_path) as store:
+                    rows = store.graph_search("photo")
+                    self.assertTrue(rows)
+                    self.assertTrue(all(row["source_note"] for row in rows))
+                    self.assertEqual(store.graph_node("photo_ingest")["source_note"], "300 areas/Photo.md")
+                    self.assertTrue(store.graph_neighbours("photo_ingest"))
+
+
+class IncrementalEstateTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.addCleanup(lambda: __import__("shutil").rmtree(self.root, ignore_errors=True))
+        self.cfg = support.make_config(self.root)
+        self.fake = support.FakeQdrant(self.cfg.dim)
+        self.harvest = estate.Harvest()
+        self.harvest.documents = loader.load_documents(self.cfg)
+        self.harvest.documents.append(self.harvested("a" * 64, "First version of the repository."))
+        self.harvest.sources["gitea"] = {"platform": "gitea", "enabled": True, "documents": 1}
+
+    def harvested(self, digest, body):
+        return loader.Document(
+            path="gitea:sam/photo-pipeline/README.md", source_id="sam/photo-pipeline/README.md",
+            area="platform", level=3, title="sam/photo-pipeline — README.md",
+            claim=body, confidence="historic", period="2024–2026", tech=[], metrics=[], artefacts=[],
+            public=True, last_verified="2026-10-14", tags=["gitea"], body=body, digest=digest,
+            url="https://gitea.example/sam/photo-pipeline/src/branch/main/README.md",
+            platform="gitea", tier=2, origin_url="https://gitea.example/sam/photo-pipeline/src/branch/main/README.md",
+            last_modified="2026-10-14T05:00:00Z", source_type="doc", kind="markdown",
+            repo_path="README.md",
+        )
+
+    def run_pipeline(self, mode="sync"):
+        with mock.patch("corpus.sources.estate.harvest", return_value=self.harvest):
+            with support.fake_qdrant(self.cfg, self.fake):
+                return pipeline.run(self.cfg, mode=mode)
+
+    def test_rebuild_and_second_run_do_not_duplicate(self):
+        first = self.run_pipeline("rebuild")
+        points = len(self.fake.collections[self.cfg.collection])
+        self.assertEqual(first.platforms["gitea"]["documents"], 1)
+        second = self.run_pipeline("sync")
+        self.assertEqual(len(self.fake.collections[self.cfg.collection]), points)
+        self.assertEqual(second.added, 0)
+        self.assertEqual(second.updated, 0)
+        self.assertEqual(second.texts_embedded, 0)
+
+    def test_only_the_changed_harvested_document_is_reembedded(self):
+        self.run_pipeline("rebuild")
+        self.harvest.documents[-1] = self.harvested("b" * 64, "Second version, one line changed.")
+        stats = self.run_pipeline("sync")
+        self.assertEqual(stats.updated, 1)
+        expected = len(__import__("corpus.chunker", fromlist=["chunk_document"]).chunk_document(
+            "Second version, one line changed."
+        ))
+        self.assertEqual(stats.texts_embedded, expected)
+        self.assertEqual(stats.embed_calls, 1)
+        self.assertEqual(stats.unchanged, 3)
+
+    def test_pool_scan_finds_a_planted_secret(self):
+        self.harvest.documents[-1] = self.harvested(
+            "c" * 64,
+            "AWS_ACCESS_KEY_ID=" + "AKIA" + "ABCDEFGHIJKLMNOP" + "\nThis line is a planted credential.",
+        )
+        with self.assertRaises(pipeline.PipelineError) as caught:
+            self.run_pipeline("rebuild")
+        self.assertIn("credential-shaped content", str(caught.exception))
+
+
+class CodeChunkerTest(unittest.TestCase):
+    def test_a_function_is_never_cut(self):
+        from corpus import chunker
+
+        body = "\n\n".join(
+            f"def function_{index}():\n    value = {{'index': {index}}}\n    return value"
+            for index in range(120)
+        )
+        chunks = chunker.chunk_code(body)
+        self.assertGreater(len(chunks), 1)
+        for chunk in chunks:
+            depth = 0
+            for line in chunk.text.splitlines():
+                depth = chunker._code_depth(line, depth)
+            self.assertEqual(depth, 0, f"chunk {chunk.index} ends inside a block")
+            self.assertIn("def function_", chunk.text)
+
+    def test_oversize_single_function_is_flagged_not_cut(self):
+        from corpus import chunker
+
+        body = "def huge():\n" + "\n".join(f"    x{index} = {index}" for index in range(2000))
+        chunks = chunker.chunk_code(body)
+        self.assertEqual(len(chunks), 1)
+        self.assertTrue(chunks[0].oversize)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_payloads.py

@@ -0,0 +1,100 @@
+"""Payload rules: every field present, enums enforced, missing field fails the run."""
+
+from __future__ import annotations
+
+import unittest
+
+from corpus import chunker, loader, payloads
+from corpus.config import Config
+from corpus.pipeline import chunk_id
+from corpus.tests import support
+
+
+def first_payload():
+    cfg = Config.from_env(root=support.REPO_ROOT, provider="local-hashing", dim=32)
+    doc = loader.load_documents(cfg)[0]
+    chunk = chunker.chunk_document(doc.body)[0]
+    payload = payloads.build_payload(doc, chunk, cfg)
+    payload["id"] = chunk_id(doc.path, chunk.index, doc.digest)
+    return cfg, doc, payload
+
+
+class BuildTest(unittest.TestCase):
+    def test_payload_has_every_required_field(self):
+        cfg, doc, payload = first_payload()
+        payloads.validate_payload(payload)
+        for field in payloads.REQUIRED_FIELDS:
+            self.assertIn(field, payload)
+        self.assertEqual(payload["area"], doc.area)
+        self.assertEqual(payload["confidence"], doc.confidence)
+        self.assertEqual(payload["last_verified"], doc.last_verified)
+        self.assertEqual(payload["embedding_model"], cfg.model)
+        self.assertEqual(payload["embedding_dim"], cfg.dim)
+
+    def test_every_real_document_produces_valid_payloads(self):
+        cfg = Config.from_env(root=support.REPO_ROOT, provider="local-hashing", dim=32)
+        for doc in loader.load_documents(cfg):
+            for chunk in chunker.chunk_document(doc.body):
+                payload = payloads.build_payload(doc, chunk, cfg)
+                payload["id"] = chunk_id(doc.path, chunk.index, doc.digest)
+                payloads.validate_payload(payload, where=doc.path)
+
+    def test_source_type_is_doc_by_default(self):
+        _cfg, doc, payload = first_payload()
+        self.assertEqual(payload["source_type"], "doc")
+
+    def test_source_type_from_tags(self):
+        doc = loader.Document(
+            path="content/ai/adr-x.md", source_id="adr-x", area="ai", level=3, title="ADR",
+            claim="c", confidence="delivered", period="2024", tech=[], metrics=[], artefacts=[],
+            public=True, last_verified="2026-01-01", tags=["adr"], body="", digest="0" * 64,
+            url="/ai/a/adr-x/",
+        )
+        self.assertEqual(payloads.source_type_for(doc), "adr")
+
+
+class ValidationTest(unittest.TestCase):
+    def test_missing_field_raises(self):
+        _cfg, _doc, payload = first_payload()
+        for field in ("area", "level", "source_type", "confidence", "period", "artefacts", "last_verified", "text"):
+            broken = dict(payload)
+            del broken[field]
+            with self.assertRaises(payloads.PayloadError, msg=f"missing {field} passed validation"):
+                payloads.validate_payload(broken)
+
+    def test_missing_provenance_field_raises(self):
+        _cfg, _doc, payload = first_payload()
+        for field in payloads.PROVENANCE_FIELDS:
+            broken = dict(payload)
+            del broken[field]
+            with self.assertRaises(payloads.PayloadError, msg=f"missing {field} passed validation"):
+                payloads.validate_payload(broken)
+
+    def test_bad_enums_raise(self):
+        _cfg, _doc, payload = first_payload()
+        for field, value in (
+            ("area", "sales"),
+            ("level", 4),
+            ("source_type", "blog"),
+            ("confidence", "vibes"),
+        ):
+            broken = dict(payload)
+            broken[field] = value
+            with self.assertRaises(payloads.PayloadError, msg=f"{field}={value!r} passed validation"):
+                payloads.validate_payload(broken)
+
+    def test_empty_text_raises(self):
+        _cfg, _doc, payload = first_payload()
+        payload["text"] = "   "
+        with self.assertRaises(payloads.PayloadError):
+            payloads.validate_payload(payload)
+
+    def test_bad_date_raises(self):
+        _cfg, _doc, payload = first_payload()
+        payload["last_verified"] = "last october"
+        with self.assertRaises(payloads.PayloadError):
+            payloads.validate_payload(payload)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_pipeline.py

@@ -0,0 +1,179 @@
+"""The pipeline: deterministic ids, no duplicate vectors on re-run, incremental mode,
+incomplete payloads never reach the store (docs/phase-4-corpus.md DoD 1-6)."""
+
+from __future__ import annotations
+
+import os
+import unittest
+from unittest import mock
+
+from corpus import chunker, loader, pipeline, store as store_module
+from corpus.tests import support
+
+
+class RebuildTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.cfg = support.make_config(self.root)
+        self.addCleanup(_cleanup, self.root)
+        # One fake for the whole test: a second run must see the first run's points.
+        self.fake = support.FakeQdrant(self.cfg.dim)
+
+    def run_pipeline(self, mode="sync", cfg=None):
+        cfg = cfg or self.cfg
+        with support.fake_qdrant(cfg, self.fake):
+            return pipeline.run(cfg, mode=mode)
+
+    def test_rebuild_from_empty_volume(self):
+        stats = self.run_pipeline("rebuild")
+        self.assertEqual(stats.documents, 3)
+        self.assertEqual(stats.added, 3)
+        self.assertGreater(stats.store_chunks, 0)
+        self.assertEqual(stats.store_chunks, stats.collection_points)
+        self.assertEqual(stats.collection_points, len(self.fake.collections[self.cfg.collection]))
+        self.assertGreater(stats.embed_calls, 0)
+
+    def test_running_twice_does_not_duplicate_vectors(self):
+        first = self.run_pipeline("rebuild")
+        points_after_first = len(self.fake.collections[self.cfg.collection])
+        second = self.run_pipeline("sync")
+        points_after_second = len(self.fake.collections[self.cfg.collection])
+        self.assertEqual(points_after_first, points_after_second)
+        self.assertEqual(first.store_chunks, second.store_chunks)
+        self.assertEqual(second.added, 0)
+        self.assertEqual(second.updated, 0)
+        self.assertEqual(second.chunks_written, 0)
+
+    def test_incremental_reembeds_only_changed_documents(self):
+        self.run_pipeline("rebuild")
+
+        # Nothing changed: no embedding calls at all.
+        unchanged = self.run_pipeline("sync")
+        self.assertEqual(unchanged.embed_calls, 0)
+        self.assertEqual(unchanged.texts_embedded, 0)
+
+        # One document changes: exactly one embedding call, for that document.
+        support.write_doc(
+            self.root,
+            "edge/l3/edge-zigbee.md",
+            id="edge-zigbee", area="edge", level=3, title="Low Power Mesh Sensing",
+            claim="A constrained mesh that survives power loss.", confidence="delivered",
+            last_verified="2026-09-15", tags="zigbee, mesh",
+            body="The mesh now also reports battery voltage on every node.",
+        )
+        changed = self.run_pipeline("sync")
+        self.assertEqual(changed.added + changed.updated, 1)
+        self.assertEqual(changed.embed_calls, 1)
+        expected_chunks = len(
+            chunker.chunk_document(
+                loader.load_documents(support.make_config(self.root))[1].body
+            )
+        )
+        self.assertEqual(changed.texts_embedded, expected_chunks)
+        self.assertEqual(changed.unchanged, 2)
+
+    def test_removed_document_is_removed_from_both_indexes(self):
+        self.run_pipeline("rebuild")
+        before = self.run_pipeline("sync").store_chunks
+        os.remove(os.path.join(self.root, "content", "platform", "l3", "platform-containers.md"))
+        after = self.run_pipeline("sync")
+        self.assertEqual(after.removed, 1)
+        self.assertLess(after.store_chunks, before)
+        self.assertEqual(after.store_chunks, after.collection_points)
+        with store_module.Store(self.cfg.db_path) as store:
+            self.assertEqual(store.keyword_search("endpoint latency loss"), [])
+
+    def test_ids_are_deterministic(self):
+        first = pipeline.chunk_id("content/ai/ai-overview.md", 0, "a" * 64)
+        second = pipeline.chunk_id("content/ai/ai-overview.md", 0, "a" * 64)
+        other = pipeline.chunk_id("content/ai/ai-overview.md", 1, "a" * 64)
+        self.assertEqual(first, second)
+        self.assertNotEqual(first, other)
+        self.assertEqual(len(first), 36)
+
+    def test_invalid_payload_aborts_before_any_write(self):
+        real_build = pipeline.payloads.build_payload
+
+        def broken(doc, chunk, cfg):
+            payload = real_build(doc, chunk, cfg)
+            payload.pop("confidence")
+            return payload
+
+        with mock.patch("corpus.payloads.build_payload", side_effect=broken):
+            with self.assertRaises(pipeline.PipelineError):
+                self.run_pipeline("rebuild")
+        self.assertEqual(self.fake.collections.get(self.cfg.collection, {}), {})
+
+    def test_provider_change_is_refused_without_rebuild(self):
+        self.run_pipeline("rebuild")
+        other = support.make_config(self.root, model="local-hashing-v2")
+        with self.assertRaises(pipeline.PipelineError) as caught:
+            self.run_pipeline("sync", cfg=other)
+        self.assertIn("governed migration", str(caught.exception))
+
+    def test_contract_is_recorded(self):
+        self.run_pipeline("rebuild")
+        with store_module.Store(self.cfg.db_path) as store:
+            contract = store.get_meta("collection")
+        self.assertEqual(contract["model"], self.cfg.model)
+        self.assertEqual(contract["dim"], self.cfg.dim)
+        self.assertEqual(contract["chunker_version"], self.cfg.chunker_version)
+        self.assertTrue(os.path.exists(self.cfg.meta_path))
+
+
+class VerifyTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.cfg = support.make_config(self.root)
+        self.addCleanup(_cleanup, self.root)
+
+    def test_clean_after_rebuild(self):
+        with support.fake_qdrant(self.cfg):
+            pipeline.run(self.cfg, mode="rebuild")
+            self.assertEqual(pipeline.verify(self.cfg), [])
+
+    def test_detects_source_changed_without_sync(self):
+        with support.fake_qdrant(self.cfg):
+            pipeline.run(self.cfg, mode="rebuild")
+            support.write_doc(
+                self.root,
+                "ai/ai-overview.md",
+                id="ai-overview", area="ai", level=1, title="Agent Systems",
+                claim="Changed.", confidence="running", last_verified="2026-10-02",
+                tags="agents", body="A different body entirely.",
+            )
+            problems = pipeline.verify(self.cfg)
+        self.assertTrue(any("changed since it was embedded" in problem for problem in problems), problems)
+
+    def test_detects_missing_document(self):
+        with support.fake_qdrant(self.cfg):
+            pipeline.run(self.cfg, mode="rebuild")
+            os.remove(os.path.join(self.root, "content", "ai", "ai-overview.md"))
+            problems = pipeline.verify(self.cfg)
+        self.assertTrue(any("no longer in content/" in problem for problem in problems), problems)
+
+
+class AreaFilterTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.cfg = support.make_config(self.root)
+        self.addCleanup(_cleanup, self.root)
+
+    def test_filter_narrows_the_result_set(self):
+        with support.fake_qdrant(self.cfg):
+            pipeline.run(self.cfg, mode="rebuild")
+            result = pipeline.demo_area_filter(self.cfg, "container estate limits", area="edge")
+        self.assertTrue(result["unfiltered"])
+        self.assertTrue(result["filtered"])
+        self.assertEqual({row["area"] for row in result["filtered"]}, {"edge"})
+        self.assertGreater(len({row["area"] for row in result["unfiltered"]}), 1)
+
+
+def _cleanup(root):
+    import shutil
+
+    shutil.rmtree(root, ignore_errors=True)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_qdrant_client.py

@@ -0,0 +1,104 @@
+"""Qdrant client: the wire shape the pipeline depends on, and the fallback route."""
+
+from __future__ import annotations
+
+import unittest
+
+from corpus import qdrant as qdrant_module
+from corpus.tests import support
+
+
+class ClientTest(unittest.TestCase):
+    def test_create_collection_sends_size_and_distance(self):
+        with support.StubServer(lambda method, path, body: (200, {"result": True})) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            client.create_collection("resume_corpus", 1536)
+        request = server.requests[0]
+        self.assertEqual(request["method"], "PUT")
+        self.assertEqual(request["path"], "/collections/resume_corpus")
+        body = request["body"]
+        self.assertEqual(body["vectors"]["size"], 1536)
+        self.assertEqual(body["vectors"]["distance"], "Cosine")
+        # A memory-bound host: payloads and the HNSW graph live on disk.
+        self.assertTrue(body["on_disk_payload"])
+        self.assertTrue(body["hnsw_config"]["on_disk"])
+        self.assertFalse(body["vectors"]["on_disk"])
+        self.assertGreater(body["optimizers_config"]["memmap_threshold"], 0)
+
+    def test_upsert_waits_and_batches(self):
+        with support.StubServer(lambda method, path, body: (200, {"result": {}})) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            points = [{"id": f"id-{index}", "vector": [0.0], "payload": {}} for index in range(5)]
+            sent = client.upsert("c", points, batch=2)
+        self.assertEqual(sent, 5)
+        self.assertEqual(len(server.requests), 3)
+        self.assertTrue(all("wait=true" in request["path"] for request in server.requests))
+        self.assertEqual(sum(len(r["body"]["points"]) for r in server.requests), 5)
+
+    def test_delete_by_source_sends_a_source_filter(self):
+        with support.StubServer(lambda method, path, body: (200, {"result": {}})) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            client.delete_by_source("c", "content/edge/l3/edge-zigbee.md")
+        request = server.requests[0]
+        self.assertEqual(request["method"], "POST")
+        self.assertIn("/collections/c/points/delete", request["path"])
+        self.assertEqual(
+            request["body"],
+            {"filter": {"must": [{"key": "source_path", "match": {"value": "content/edge/l3/edge-zigbee.md"}}]}},
+        )
+
+    def test_search_filters_before_searching(self):
+        responder = lambda method, path, body: (200, {"result": {"points": []}})  # noqa: E731
+        with support.StubServer(responder) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            client.search("c", [0.1, 0.2], limit=3, area="edge", confidence="delivered")
+        body = server.requests[0]["body"]
+        self.assertEqual(server.requests[0]["path"], "/collections/c/points/query")
+        self.assertEqual(body["limit"], 3)
+        self.assertEqual(
+            body["filter"],
+            {
+                "must": [
+                    {"key": "area", "match": {"value": "edge"}},
+                    {"key": "confidence", "match": {"value": "delivered"}},
+                ]
+            },
+        )
+
+    def test_search_falls_back_to_the_legacy_route(self):
+        def responder(method, path, body):
+            if path.endswith("/points/query"):
+                return 404, {"status": "not found"}
+            return 200, {"result": [{"id": "p1", "score": 0.9, "payload": {"area": "ai"}}]}
+
+        with support.StubServer(responder) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            hits = client.search("c", [0.1])
+        self.assertEqual([hit["id"] for hit in hits], ["p1"])
+        self.assertEqual([r["path"] for r in server.requests][-1], "/collections/c/points/search")
+
+    def test_point_count_is_exact(self):
+        responder = lambda method, path, body: (200, {"result": {"count": 42}})  # noqa: E731
+        with support.StubServer(responder) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=1)
+            self.assertEqual(client.point_count("c"), 42)
+        self.assertEqual(server.requests[0]["body"], {"exact": True})
+
+    def test_health_is_false_when_unreachable(self):
+        client = qdrant_module.Qdrant("http://127.0.0.1:1", timeout=0.2, attempts=1)
+        self.assertFalse(client.health())
+
+    def test_upsert_retries_a_server_error(self):
+        plan = [500, 200]
+
+        def responder(method, path, body):
+            return plan.pop(0), {"result": {}}
+
+        with support.StubServer(responder) as server:
+            client = qdrant_module.Qdrant(server.url, attempts=2)
+            client.upsert("c", [{"id": "x", "vector": [0.0], "payload": {}}])
+        self.assertEqual(len(server.requests), 2)
+
+
+if __name__ == "__main__":
+    unittest.main()

corpus/tests/test_store.py

@@ -0,0 +1,106 @@
+"""FTS5 chunk index, manifest and meta: the keyword half of hybrid retrieval."""
+
+from __future__ import annotations
+
+import unittest
+
+from corpus import chunker, loader, payloads, store as store_module
+from corpus.config import Config
+from corpus.pipeline import chunk_id
+from corpus.tests import support
+
+
+def build_store(path, cfg, documents):
+    store = store_module.Store(path)
+    for doc in documents:
+        payload_list = []
+        for chunk in chunker.chunk_document(doc.body):
+            payload = payloads.build_payload(doc, chunk, cfg)
+            payload["id"] = chunk_id(doc.path, chunk.index, doc.digest)
+            payload_list.append(payload)
+        store.write_document(doc, payload_list)
+    return store
+
+
+class StoreTest(unittest.TestCase):
+    def setUp(self):
+        self.root = support.temp_repo()
+        self.cfg = support.make_config(self.root)
+        self.addCleanup(_cleanup, self.root)
+        self.documents = loader.load_documents(self.cfg)
+        self.store = build_store(self.cfg.db_path, self.cfg, self.documents)
+        self.addCleanup(self.store.close)
+
+    def test_counts_match_the_documents(self):
+        self.assertEqual(self.store.source_count(), len(self.documents))
+        expected = sum(len(chunker.chunk_document(doc.body)) for doc in self.documents)
+        self.assertEqual(self.store.chunk_count(), expected)
+
+    def test_keyword_search_finds_exact_tokens(self):
+        hits = self.store.keyword_search("zigbee")
+        self.assertTrue(hits, "no hit for a token that is literally in the document")
+        self.assertEqual(hits[0]["area"], "edge")
+
+    def test_keyword_search_respects_area_filter(self):
+        all_hits = self.store.keyword_search("latency")
+        edge_hits = self.store.keyword_search("latency", area="edge")
+        self.assertTrue(all_hits)
+        for hit in edge_hits:
+            self.assertEqual(hit["area"], "edge")
+
+    def test_keyword_search_returns_snippets_and_scores(self):
+        hits = self.store.keyword_search("fence")
+        self.assertTrue(hits)
+        self.assertIn("score", hits[0])
+        self.assertIn("snippet", hits[0])
+
+    def test_operator_injection_is_not_honoured(self):
+        # FTS5 operators must be treated as words, never as query syntax.
+        hits = self.store.keyword_search('edge OR "NEAR"(a b)')
+        self.assertIsInstance(hits, list)
+
+    def test_manifest_records_the_content_hash(self):
+        manifest = self.store.manifest()
+        self.assertEqual(set(manifest), {doc.path for doc in self.documents})
+        for doc in self.documents:
+            self.assertEqual(manifest[doc.path]["content_hash"], doc.digest)
+            self.assertGreater(manifest[doc.path]["chunk_count"], 0)
+
+    def test_delete_removes_rows_from_both_tables(self):
+        target = self.documents[1]
+        self.store.delete_document(target.path)
+        self.assertNotIn(target.path, self.store.manifest())
+        for row in self.store.chunk_payloads():
+            self.assertNotEqual(row["source_path"], target.path)
+
+    def test_meta_roundtrip(self):
+        self.store.set_meta("collection", {"model": "x", "dim": 8})
+        self.assertEqual(self.store.get_meta("collection"), {"model": "x", "dim": 8})
+        self.assertIn("collection", self.store.meta())
+
+    def test_payloads_survive_the_roundtrip(self):
+        rows = self.store.chunk_payloads()
+        self.assertTrue(rows)
+        for row in rows:
+            payloads.validate_payload(row)
+
+
+class MatchExpressionTest(unittest.TestCase):
+    def test_terms_become_prefix_queries(self):
+        self.assertEqual(store_module.match_expression("Zigbee mesh"), '"zigbee"* AND "mesh"*')
+
+    def test_empty_query_is_empty(self):
+        self.assertEqual(store_module.match_expression("   "), "")
+
+    def test_punctuation_is_dropped(self):
+        self.assertEqual(store_module.match_expression("-edge- LoRaWAN, 2023"), '"edge"* AND "lorawan"* AND "2023"*')
+
+
+def _cleanup(root):
+    import shutil
+
+    shutil.rmtree(root, ignore_errors=True)
+
+
+if __name__ == "__main__":
+    unittest.main()

deploy/compose.yml

@@ -30,11 +30,13 @@ services:
     restart: unless-stopped
     mem_limit: 120m
     memswap_limit: 120m
-    # The app listens on 0.0.0.0 inside the container namespace; the host-side
-    # binding is loopback, so nothing is reachable from the LAN except through
-    # docker-proxy on 127.0.0.1:8096 — which is what Caddy proxies to.
+    # Published on this host's LAN address, not loopback. Caddy lives on .35 and
+    # proxies to 192.168.20.13:8096, exactly as it does for offsite (:8095),
+    # photo-filter (:8092), garage-ui (:3909) and prefect (:4200). The port is
+    # reachable inside the house only; Caddy on .35 remains the sole public
+    # listener, which is the intent of AGENTS.md section 4.
     ports:
-      - "127.0.0.1:8096:8096"
+      - "192.168.20.13:8096:8096"
     read_only: true
     security_opt:
       - no-new-privileges:true

deploy/qdrant.yml

@@ -0,0 +1,66 @@
+# Qdrant for the corpus (docs/phase-4-corpus.md, reference/stack.md).
+#
+#   mem_limit == memswap_limit == 400m   the cap in the memory budget, and the
+#                                        container cannot reach host swap
+#   ports on 127.0.0.1 only              Caddy on .35 is the only public listener
+#   its own compose project              stopping or recreating it cannot disturb
+#                                        the Mode A site (test A1 still passes:
+#                                        the site never calls Qdrant)
+#
+# Run on the runtime host (.13):
+#
+#   docker compose -f deploy/qdrant.yml up -d
+#   curl -s http://127.0.0.1:6333/        # must answer with a version
+#
+# Then, in the corpus environment (RESUME_QDRANT_URL is already the default):
+#
+#   python3 -m corpus.cli rebuild
+#
+# Deleting the volume is always recoverable: `rebuild` re-derives every vector
+# from content/.
+
+name: ai-resume-index
+
+services:
+  qdrant:
+    image: qdrant/qdrant:v1.20.0
+    container_name: ai-resume-qdrant
+    restart: unless-stopped
+    mem_limit: 400m
+    memswap_limit: 400m
+    ports:
+      - "127.0.0.1:6333:6333"   # REST — the pipeline and the query service use this
+      - "127.0.0.1:6334:6334"   # gRPC — reserved for the Phase 5 query service
+    volumes:
+      - qdrant_storage:/qdrant/storage
+    environment:
+      # The corpus is ~1k chunks today and the measurement in
+      # reference/stack.md records the ceiling. Keep the HNSW defaults; do not
+      # tune for speed on a memory-bound host.
+      QDRANT__TELEMETRY_DISABLED: "true"
+      QDRANT__LOG_LEVEL: "WARN"
+    logging:
+      driver: json-file
+      options:
+        max-size: 10m
+        max-file: "3"
+    healthcheck:
+      test: ["CMD-SHELL", "bash -c '</dev/tcp/127.0.0.1/6333' || exit 1"]
+      interval: 30s
+      timeout: 3s
+      retries: 3
+      start_period: 10s
+
+# The host's docker daemon has every default address pool allocated by other
+# projects, so an implicit compose network cannot be created here. Pin a /28 of
+# RFC1918 space that no other bridge uses; it is separate from the site's
+# 10.89.0.0/28 so stopping one project cannot disturb the other.
+networks:
+  default:
+    ipam:
+      config:
+        - subnet: 10.89.0.16/28
+          gateway: 10.89.0.17
+
+volumes:
+  qdrant_storage:

docs/phase-4-corpus.md

@@ -1,81 +1,157 @@
-# Phase 4 — Corpus and retrieval
+# Phase 4 — Ingestion from sources
 
-**Goal:** `content/` becomes a retrievable corpus with filters that work before the vector search, and a freshness audit that cannot silently rot.
+**Goal:** a fireable tool that reads the estate — Gitea, the notes vault, the knowledge graph — and keeps a searchable pool current. The pool is not built from `content/` alone; `content/` is the top tier of it.
 
-**Depends on:** Phase 1 (content), Phase 2 (FTS5 index exists).
+**Depends on:** Phase 1 (content and schema), Phase 2 (FTS5 index).
+
+**Why this changed.** An earlier version of this document said the corpus would be built from `content/` only. That was wrong. A hand-curated set cannot answer "how did the photo pipeline deduplicate" or "what did the garden controller do before the rewrite". The estate can. Curated claims answer questions about Sam; the estate answers questions about the work.
 
 ---
 
-## Work
+## 1. Sources
 
-### 4.1 Pipeline
+| Source | Reached by | What to take |
+|---|---|---|
+| **Gitea**, all repositories | API over HTTPS | `README.md`, file tree (paths only), selected source files, recent commit messages, language statistics, repo description and topics |
+| **`sam/obsidian-vault`** | Same Gitea API — the vault is a repo | Markdown body, front-matter title and tags, folder path |
+| **Graphify output** | `graph.json` committed to the vault repo | Nodes (label, source_file, community), edges (relation, confidence) |
+| **`content/`** | Local repo | The curated claims. **Tier 1** — outranks everything above on conflict |
 
-```
-content/  ──► validate ──► chunk ──► enrich ──► embed ──► upsert Qdrant
-                                                              │
-                                    SQLite FTS5 (from Phase 2) ┘
+Gitea is the only network dependency. The vault and the graph arrive through it, so there is no cross-machine copy and no SSH from `.13` to `.27`.
+
+### What not to read
+
+Media never becomes text. Images, video and audio are recorded as `image` node references, tagged with their path, and nothing more.
+
+---
+
+## 2. Platform guides
+
+One file per platform in `platforms/<name>.yml`. They are configuration **and** public copy — the site shows them so a visitor can see what the assistant reads.
+
+```yaml
+name: gitea
+enabled: true
+what: "Self-hosted Git. Source for every current project."
+fits: [ai, platform, edge]
+base: https://gitea.lab.audasmedia.com.au
+auth: GITEA_TOKEN
+include:
+  repos: all
+exclude:
+  paths: ["**/node_modules/**", "**/.venv/**", "**/.git/**", "**/*.env*",
+          "**/.silverbullet.auth.json", "**/*.pem", "**/*.key"]
+  repos: []
+read: [readme, tree, selected_source, commits, languages]
+limits: { max_documents: 4000, max_file_kb: 200 }
 ```
 
-- **Chunking:** 400–800 tokens, ~12% overlap, **code-aware** — never split inside a fenced block, table or front-matter.
-- **Deterministic IDs:** `sha256(source_path + chunk_index + content_hash)`. Re-running must update, not duplicate.
-- **Incremental:** hash the source set; only changed documents are re-chunked and re-embedded.
-- **One command** rebuilds from scratch. Deleting the Qdrant volume is always recoverable.
+```yaml
+name: obsidian
+enabled: true
+what: "Personal knowledge base. High-level notes on every system, in Sam's words."
+fits: [about, ai, platform, edge]
+base: sam/obsidian-vault
+include:
+  paths: ["200 projects/**", "300 areas/**", "400 Personal Family/**",
+          "100 Table of Contents/**", "010 inbox/**"]
+exclude:
+  paths: [".obsidian/**", ".trash/**", ".venv/**", "templates/**",
+          "**/*.excalidraw", "**/*.png", "**/*.jpg"]
+limits: { max_documents: 4000, max_file_kb: 200 }
+```
 
-### 4.2 Payload schema (the retrieval filter)
+```yaml
+name: graphify
+enabled: true
+what: "Knowledge graph over the vault. Answers 'what connects X to Y' rather than 'what mentions X'."
+fits: [ai]
+base: sam/obsidian-vault
+read: [graph_json]
+```
+
+`fits` drives the payload `area`, which is what lets the query filter before it searches.
+
+---
+
+## 3. Metadata on every chunk
 
 ```json
 {
-  "area": "ai|platform|edge|products|history|about",
-  "level": 1,
-  "source_type": "doc|code|adr|metric",
-  "confidence": "delivered|running|in-progress|abandoned|historic",
-  "period": "2023–",
-  "artefacts": ["..."],
-  "last_verified": "2026-10-20",
+  "platform": "gitea | obsidian | graphify | curated",
+  "tier": 1,
+  "source_type": "doc | code | commit | graph_node | claim",
+  "area": "ai | platform | edge | products | history | about",
+  "origin_url": "https://gitea.../sam/photo-pipeline/src/x.py",
+  "path": "src/x.py",
+  "title": "photo-pipeline / dedup",
+  "confidence": "delivered | running | in-progress | abandoned | historic",
+  "last_modified": "2026-10-14T05:00:00Z",
+  "content_hash": "sha256:...",
+  "public": true,
   "text": "..."
 }
 ```
 
-`area` and `confidence` are the two that matter: they let the query path filter *before* searching, and they let the answer qualify unfinished work.
+`tier: 1` is curated `content/`. Everything harvested is `tier: 2`. The query layer uses this to prefer the curated claim when the two disagree.
 
-### 4.3 Embeddings
-- Batch, low volume, incremental.
-- **Record the embedding model and dimension per collection.** A provider change is a governed migration, not a silent corruption.
-- Provider decision pending Sam — see `reference/stack.md`. Build provider-agnostic behind one interface.
-
-### 4.4 Scheduling
-- Prefect deployment on the existing Prefect server; **worker only**, do not deploy a second server.
-- Trigger: git push to `content/`, plus a nightly audit.
-- Nightly audit reports documents past their `last_verified` horizon and alerts via the existing Apprise service.
+`origin_url` is mandatory for harvested chunks. A chunk with no origin is not admitted — an answer that cannot be traced is worse than no answer.
 
 ---
 
-## Definition of Done
+## 4. Pipeline
 
-- [ ] One command rebuilds the corpus from an empty volume.
-- [ ] Running it twice produces no duplicate vectors (count is stable).
-- [ ] No chunk boundary lands inside a code fence, table or front-matter block.
-- [ ] Every chunk carries all payload fields; a missing field fails the run rather than defaulting.
-- [ ] Payload filtering demonstrably narrows results — demonstrate the same query with and without an `area` filter.
-- [ ] Incremental mode re-embeds only changed documents; prove it with the provider call count.
-- [ ] Freshness audit runs and reports stale documents.
-- [ ] Qdrant memory measured at 500 / 5,000 / 20,000 chunks and recorded in `reference/stack.md`, within its cap.
+```
+guides  →  harvest (per platform)  →  normalise to text  →  chunk  →  enrich  →  embed  →  upsert
+                                        │
+                                        └─ reject: excluded path, no origin, binary, oversize
+```
+
+- **Chunking** is code-aware: 400–800 tokens, ~12% overlap, never split inside a fenced block, table, front-matter block or function. Use the language grammar where available; fall back to paragraph splitting.
+- **Deterministic IDs:** `sha256(origin_url + chunk_index + content_hash)`. Re-running updates, never duplicates.
+- **Incremental:** compare `content_hash` against what the pool already holds; embed only what changed. A full rebuild must always be possible and is one command.
+- **Graph** is not chunked. Nodes and edges go into their own collection, joined to note chunks by `path`, so a graph answer can name the note it came from.
 
 ---
 
-## Out of scope
+## 5. Scheduling
 
-- Answering questions. Phase 4 ends at "retrieval returns the right chunks".
-- Query graphs, prompts, guardrails — Phase 5 and 6.
-- Ingesting anything outside `content/`. Four sources maximum, and this phase uses one: the repo.
+- On demand: one command.
+- Scheduled: a Prefect deployment on the existing server. Worker only; do not deploy a second Prefect server.
+- Reports what changed each run: documents added, updated, removed, and any excluded-path rejections.
+- A source that has not changed in its expected window raises an alert through the existing Apprise service. Silent staleness is the failure mode that matters.
 
 ---
 
-## Risks
+## 6. Definition of Done
+
+- [ ] One command rebuilds the pool from empty.
+- [ ] A second run produces no duplicate vectors.
+- [ ] A probe document placed in an excluded path never appears in the pool, and the run reports the rejection.
+- [ ] A document with no resolvable origin is refused, not defaulted.
+- [ ] Every chunk carries all fields in section 3.
+- [ ] Incremental mode re-embeds only changed documents, proven by embedding call count.
+- [ ] Graph nodes and edges are queryable and each resolves to a source note.
+- [ ] The scheduled run reports deltas and alerts on staleness.
+- [ ] Qdrant stays inside its `mem_limit`; footprint recorded at 1k, 10k and 40k chunks.
+- [ ] No credential, `.env`, key file or family-personal binary is present in the pool — verified by search, not by assertion.
+
+---
+
+## 7. Out of scope
+
+- Answering questions. This phase ends at "retrieval returns the right chunks with their origins".
+- The query graph, prompts and guardrails — Phase 5.
+- Building the Graphify graph. It already exists and is rebuilt on `.27` with its own skill. This phase consumes its output.
+
+---
+
+## 8. Risks
 
 | Risk | Mitigation |
 |---|---|
-| Corpus grows beyond RAM | Incremental by default; measure at three corpus sizes; cap Qdrant with `mem_limit` |
-| Duplicate vectors on re-run | Deterministic IDs are mandatory, not optional |
-| Silent staleness | Nightly audit + Apprise alert; a stale index is otherwise an invisible quality failure |
-| Provider rate limits during bulk embed | Batch with retry/backoff; ingest is low-volume and incremental |
+| Volume — 50 repos plus 280 notes exceeds the memory cap | Cap per platform, skip large files, measure at three corpus sizes |
+| Scraped text contradicts a curated claim | `tier` field; the query prefers tier 1 and says so |
+| A secret is harvested | Exclusion list plus a post-run scan of the pool for credential patterns, failing the run if found |
+| Staleness is invisible | Scheduled run plus alert |
+| The estate is unavailable when a recruiter asks | The pool is local; only ingestion needs Gitea. Serving never does |

docs/phase-4-notes.md

@@ -0,0 +1,304 @@
+# Phase 4 notes — corpus and retrieval
+
+Operator manual for the corpus pipeline, and the evidence for the Phase 4
+Definition of Done in `docs/phase-4-corpus.md`. Every number below was measured
+on 2026-10-26 against a live Qdrant 1.20.0 container with a 400 MiB cap.
+
+---
+
+## 1. What exists now
+
+| Path | What it is |
+|---|---|
+| `corpus/` | The pipeline: one Python package, standard library only |
+| `corpus/cli.py` | `python3 -m corpus.cli {rebuild,sync,audit,verify,demo-filter,hybrid,measure,info}` |
+| `corpus/chunker.py` | Code-aware chunking (fenced block, table and front-matter are atomic) |
+| `corpus/embeddings.py` | The only place a vector is made; two providers behind one interface |
+| `corpus/qdrant.py` | Qdrant REST client (no `qdrant-client` dependency) |
+| `corpus/store.py` | SQLite: FTS5 over chunks, the incremental manifest, the collection contract |
+| `corpus/pipeline.py` | validate → chunk → enrich → embed → upsert, in that order |
+| `corpus/freshness.py` | Nightly audit: stale documents, index drift, contract drift |
+| `corpus/measure.py` | Memory measurement at 500 / 5,000 / 20,000 chunks |
+| `corpus/tests/` | 76 unit tests, no container, no network |
+| `flows/corpus.py` | Prefect flows: `corpus-sync`, `corpus-rebuild`, `corpus-audit` |
+| `deploy/qdrant.yml` | Qdrant container: `mem_limit` = `memswap_limit` = 400 m, loopback only |
+
+Two indexes are written from one pass over `content/`, so they cannot drift:
+
+* **vectors** in Qdrant (`resume_corpus`), one point per chunk;
+* **keywords** in SQLite (`data/corpus/corpus.db`), one FTS5 row per chunk.
+
+Retrieval uses both (`corpus.cli hybrid`, reciprocal rank fusion). Vectors miss
+exact tokens such as repo names, `LoRaWAN`, `MQTT` and dates; FTS5 alone misses
+paraphrase. Decision D10 requires the pair.
+
+---
+
+## 2. Commands
+
+```sh
+make corpus          # rebuild from content/ (drops the collection first)
+make corpus-sync     # incremental: only changed documents are re-embedded
+make corpus-audit    # freshness + drift audit; exit 1 on findings
+make corpus-verify   # index vs content/ vs Qdrant, consistency report
+make corpus-demo     # the same query with and without an area filter
+make corpus-test     # 76 unit tests
+```
+
+Direct form, when you need flags:
+
+```sh
+python3 -m corpus.cli rebuild --qdrant-url http://127.0.0.1:6333 --dim 1024
+python3 -m corpus.cli audit --today 2027-01-20 --json --fail-on-findings
+python3 -m corpus.cli measure --sizes 500,5000,20000 --container ai-resume-qdrant
+```
+
+Configuration is environment-first. Nothing is read from a file in this repo.
+
+| Variable | Default | Meaning |
+|---|---|---|
+| `RESUME_QDRANT_URL` | `http://127.0.0.1:6333` | Qdrant REST endpoint |
+| `RESUME_COLLECTION` | `resume_corpus` | collection name |
+| `RESUME_EMBED_PROVIDER` | `local-hashing` | `local-hashing` or `omniroute` |
+| `RESUME_EMBED_MODEL` | provider default | recorded per collection |
+| `RESUME_EMBED_DIM` | provider default | recorded per collection |
+| `RESUME_CORPUS_DB` | `data/corpus/corpus.db` | FTS5 + manifest |
+| `RESUME_MAX_AGE_DAYS` | `180` | freshness horizon |
+| `RESUME_WARN_DAYS` | `30` | how early "due soon" starts |
+| `RESUME_APPRISE_URL` | unset | existing Apprise endpoint for audit alerts |
+| `OMNIROUTE_API_KEY` | unset | only for the `omniroute` provider; never in the repo |
+
+---
+
+## 3. Pipeline
+
+```
+content/ ─► make validate ─► chunk ─► enrich ─► embed ─► upsert Qdrant
+                                        │
+                                        └─► SQLite FTS5 + manifest
+```
+
+* **Validate first.** The pipeline runs the Phase 1 validator
+  (`tools/validate_content.py`). A broken content tree stops the run before a
+  single vector is written.
+* **Chunking.** Target 600 tokens, hard ceiling 800, ~12 % overlap. The body is
+  parsed into atomic blocks (fenced code, table, heading, list item, quote,
+  paragraph, rule) and chunks are packed from whole blocks, so a boundary
+  cannot fall inside a fence, a table or front-matter. A list item or a table
+  that exceeds 800 tokens becomes one oversize chunk and is flagged, never cut.
+  Tokens are counted as words plus punctuation, which is a conservative
+  under-estimate of BPE tokens.
+* **Payload before vectors.** Every chunk payload is built and validated for
+  every changed document *before* one point is written. A missing or invalid
+  field aborts the run with the document and chunk named.
+* **Deterministic ids.** `sha256(source_path + chunk_index + content_hash)`,
+  rendered as a UUID because Qdrant accepts only integers or UUIDs. Re-running
+  updates points; a document whose chunk count shrinks has its old points
+  deleted by `source_path` filter, so orphans cannot accumulate.
+* **Incremental.** One hash comparison against the manifest per document. The
+  document is re-embedded when its bytes changed, or when the embedding model,
+  dimension or chunker version changed.
+* **Consistency check.** At the end of every run the point count in Qdrant must
+  equal the chunk count in SQLite, or the run fails.
+
+Corpus today: **40 documents → 51 chunks**, token range 100–707 (median 225),
+8 protected regions (fenced blocks and tables), 0 boundary violations.
+
+---
+
+## 4. Payload
+
+`docs/phase-4-corpus.md` §4.2 fields, plus provenance so an answer can cite:
+
+```json
+{
+  "area": "edge", "level": 3, "source_type": "doc", "confidence": "running",
+  "period": "2020–", "artefacts": [], "last_verified": "2026-10-26",
+  "text": "# Artefact — Distributed audio …",
+  "id": "014d792f-28d9-e0a0-610e-7ffc04eea3ad",
+  "source_path": "content/edge/l3/edge-snapcast.md", "source_id": "edge-snapcast",
+  "chunk_index": 0, "title": "…", "claim": "…", "url": "/edge/a/edge-snapcast/",
+  "content_hash": "12fc4fa9…", "embedding_model": "local-hashing-v1",
+  "embedding_dim": 1024, "chunker_version": "blocks-v1", "chunk_tokens": 191,
+  "public": true, "tech": ["Snapcast", "Mopidy", "…"], "tags": ["edge", "audio"]
+}
+```
+
+`area` and `confidence` are the two that matter. The query path filters on them
+*before* it searches, and an answer can qualify unfinished work instead of
+presenting it as delivered.
+
+---
+
+## 5. Freshness audit
+
+```sh
+make corpus-audit                  # human table, exit 1 on findings
+python3 -m corpus.cli audit --json # machine-readable, for Prefect
+```
+
+Findings, in order of seriousness:
+
+| Finding | Meaning |
+|---|---|
+| `stale` | `last_verified` older than `RESUME_MAX_AGE_DAYS` (180). The claim may no longer be true |
+| `due_soon` | inside the warning window (30 days by default). A warning, not a failure |
+| `index_drift` | the source changed, disappeared or was never indexed since it was embedded |
+| `contract_drift` | the collection was built with a different model, dimension or chunker |
+
+`ok` is false when there is a `stale`, `index_drift` or `contract_drift`
+finding. `due_soon` alone keeps the audit green.
+
+`--alert` posts the report to `RESUME_APPRISE_URL` when the audit is not ok.
+The Prefect flow does this by default.
+
+---
+
+## 6. Scheduling (Prefect, existing server, worker only)
+
+`flows/corpus.py` defines three flows. It imports Prefect when present and runs
+as plain Python when it is not, so the CLI never depends on the orchestrator.
+
+On `.13`, which already runs the Prefect server — **do not start a second one**:
+
+```sh
+# once, per deployment
+export RESUME_QDRANT_URL=http://127.0.0.1:6333
+export RESUME_EMBED_PROVIDER=omniroute RESUME_EMBED_MODEL=<approved model> RESUME_EMBED_DIM=<dim>
+prefect deploy flows/corpus.py:sync_corpus    --name corpus-sync
+prefect deploy flows/corpus.py:freshness_audit --name corpus-audit --cron "0 4 * * *"
+```
+
+Triggers required by `docs/phase-4-corpus.md` §4.4:
+
+* **on push touching `content/`** — a deployment trigger on the repository, or
+  `make corpus-sync` from the same post-receive hook that runs `deploy.sh`;
+* **nightly audit** — cron `0 4 * * *`, which alerts through Apprise.
+
+This deployment is **not applied yet**: it needs the runtime host and the
+provider decision in §8. Both are Sam's. The flows themselves are ready.
+
+---
+
+## 7. Qdrant memory at three corpus sizes
+
+Measured with `python3 -m corpus.cli measure`, Qdrant 1.20.0 in a container
+capped at 400 MiB (`mem_limit` = `memswap_limit`), 1024-dimensional vectors,
+real payloads from `content/`, throttled ingest in batches of 250.
+
+| Chunks | Points | Settled RSS | Peak during ingest | Growth | Bytes/chunk | Ingest time |
+|---|---|---|---|---|---|---|
+| 0 | 0 | 88.6 MiB | — | — | — | — |
+| 500 | 500 | 128.3 MiB | 148.5 MiB | +39.7 MiB | 83 KiB | 1.1 s |
+| 5,000 | 5,000 | 137.7 MiB | 305.0 MiB | +49.1 MiB | 10 KiB | 10.9 s |
+| 20,000 | 20,000 | 150.9 MiB | 292.3 MiB | +62.3 MiB | 3 KiB | 43.7 s |
+
+All three sizes are **within the 400 MiB cap**. The peak is the number that
+matters: the cgroup limit is tripped during a bulk write, not after it.
+
+### The finding that changed the design
+
+With Qdrant's **default** settings the container was OOM-killed at exactly the
+cap between 5,000 and 20,000 chunks (`anon-rss: 404204kB`, `exitCode 137`,
+`constraint=CONSTRAINT_MEMCG`). The default keeps the HNSW graph and the payload
+in RAM and holds several segments live during ingest.
+
+The collection is now created memory-frugal, and the pipeline is the only code
+that creates it (`corpus/qdrant.py`):
+
+```json
+{
+  "vectors": {"size": 1024, "distance": "Cosine", "on_disk": false},
+  "hnsw_config": {"m": 8, "ef_construct": 64, "on_disk": true},
+  "on_disk_payload": true,
+  "optimizers_config": {"default_segment_number": 2, "memmap_threshold": 5000, "indexing_threshold": 20000}
+}
+```
+
+Vectors stay in RAM because search reads them on every query. The HNSW graph and
+the payloads go to disk, where the kernel can reclaim those pages instead of
+killing the process. `memmap_threshold` 5,000 makes segments that size mmap;
+`indexing_threshold` 20,000 stops a bulk ingest from building the graph while it
+is still writing. The cost is build time and a little search latency, both
+irrelevant for a corpus of this size and a public site with no query burst.
+
+Practical ceiling: at 1024 dimensions the cap holds 20,000 chunks. If the corpus
+grows past that, set `RESUME_QDRANT_VECTORS_ON_DISK=1`, or reduce the dimension,
+or raise the cap — in that order of preference. Vector memory scales linearly
+with dimension: 1,536-dimension vectors need 50 % more than recorded here.
+
+---
+
+## 8. Embedding provider — still Sam's decision
+
+`reference/stack.md` lists the open item: Gemini free tier, OpenRouter, or
+local. Phase 4 does not wait for it:
+
+* **`omniroute`** is implemented and tested. OpenAI-compatible
+  `/v1/embeddings`, batched, retry with backoff on 429 and 5xx, key from the
+  environment only, no provider SDK, no direct vendor call (AGENTS.md §4.5).
+* **`local-hashing`** is the default for tests, for offline rebuilds and for the
+  memory measurement. It is deterministic feature hashing, **not a semantic
+  model**: it is a placeholder that lets the whole pipeline run without a
+  provider decision, and it must not be used for answers.
+
+The provider, model, dimension and chunker version are stored in the collection
+contract (`data/corpus/collection.json` and the `meta` table). A run whose
+configuration differs is refused:
+
+```
+the collection contract does not match this configuration — refusing to mix
+vectors in one collection (docs/phase-4-corpus.md §4.3).
+```
+
+### Changing the provider (governed migration)
+
+1. Get Sam's decision and record it in `reference/decisions.md`.
+2. Set `RESUME_EMBED_PROVIDER`, `RESUME_EMBED_MODEL`, `RESUME_EMBED_DIM`, and
+   `OMNIROUTE_API_KEY` in `~/.config/environment.d/`.
+3. `make corpus` — a rebuild, because a provider change invalidates every
+   vector. It is one command and takes seconds at this corpus size.
+4. `make corpus-verify` and `make corpus-audit`.
+
+---
+
+## 9. Recovery
+
+Deleting the Qdrant volume or `data/corpus/` is **always recoverable**:
+
+```sh
+docker compose -f deploy/qdrant.yml down -v
+docker compose -f deploy/qdrant.yml up -d
+make corpus          # re-derives every vector and FTS row from content/
+make corpus-verify
+```
+
+There is no state that exists only in the index. `content/` is the source of
+truth; the corpus is a derived artefact and lives in `data/`, which is
+gitignored.
+
+---
+
+## 10. Definition of Done — evidence
+
+| DoD item | Evidence |
+|---|---|
+| One command rebuilds from an empty volume | `make corpus` on an empty Qdrant volume: 40 documents → 51 chunks, 51 points, 1.0 s |
+| Running it twice produces no duplicate vectors | rebuild ×2 → 51 points both times; `sync` with no changes → 0 embedding calls, 0 points sent |
+| No chunk boundary inside a fence, table or front-matter | `corpus/tests/test_chunker.py`; 51/51 stored chunks reproduce from `content/`, 8 protected regions, 0 violations |
+| Every chunk carries all payload fields; missing field fails the run | `corpus/tests/test_payloads.py` (each field removed in turn); 51/51 stored payloads validate; a broken payload aborts before any write (`test_invalid_payload_aborts_before_any_write`) |
+| Payload filtering demonstrably narrows results | `make corpus-demo` — query "distributed audio endpoints": unfiltered returns edge *and* ai documents; `--area edge` returns edge only, `--area ai` returns ai only |
+| Incremental re-embeds only changed documents | edit one of 40 documents → `updated=1 unchanged=39 embedding calls=1 texts=1`; re-run → `calls=0` |
+| Freshness audit runs and reports stale documents | `python3 -m corpus.cli audit` today: 0 findings. `--today 2027-10-05`: 40 stale documents, one row per document with age in days. `--today 2026-11-01`: 0 findings |
+| Qdrant memory at 500 / 5,000 / 20,000 chunks, within cap | §7 table; 128 / 138 / 151 MiB settled, 149 / 305 / 292 MiB peak, cap 400 MiB |
+
+---
+
+## 11. Not done here
+
+* Qdrant is **not deployed to `.13`** and the Prefect deployment is not applied.
+  `deploy/qdrant.yml` is ready; the runtime host is Phase 3's (`AUD-143`) and
+  Sam's to apply.
+* No provider is chosen (§8).
+* Answering questions, the query graph, the streaming console and guardrails are
+  Phase 5 and Phase 6. Phase 4 ends at "retrieval returns the right chunks".

docs/phase-5-ai-overlay.md

@@ -80,3 +80,43 @@ One hand-written TypeScript island, ES modules, no framework, no bundler.
 | Panel refactor needed | Phase 2 §2.4 contract; if it fails, fix the panel contract, do not fork the UI |
 | Streaming stalls behind Caddy | Explicit flush settings; p90 measurement, not a spot check |
 | Answers drift from the site | Citation verification is a node, not a hope |
+
+---
+
+## 5.6 Citations point at the original
+
+An answer cites where it came from, not a curated page standing in for it.
+
+| Source | Cited as |
+|---|---|
+| Gitea repo file | full URL to the file in Gitea |
+| Vault note | full URL to the note in `sam/obsidian-vault` |
+| Graph answer | the note title the node came from, plus its URL |
+| Curated claim | the site page |
+
+A citation whose URL cannot be opened is dropped, not rendered. A claim with no
+traceable origin is not returned at all.
+
+## 5.7 Curated outranks harvested
+
+`content/` is tier 1; everything harvested is tier 2. When the two disagree, the
+tier 1 claim wins and the answer says so plainly: the curated description is
+current; the older text is background.
+
+This is the rule that stops a stale README from quietly overwriting Sam's own
+description of his work.
+
+## 5.8 The live source readout
+
+While the answer streams, a panel shows what is being read, named as the visitor
+would recognise it:
+
+```
+reading · gitea/photo-pipeline/README.md
+        · obsidian/210 AI Resume/Graphify.md
+        · graph · community 42 · vault infrastructure
+```
+
+This is a feature, not a debug view. It answers the question every recruiter has
+and does not ask: is this making things up? Show the working — it is also the most
+convincing demonstration of the system on the page.

flows/corpus.py

@@ -0,0 +1,127 @@
+"""Prefect entry points for the corpus pipeline.
+
+`docs/phase-4-corpus.md` §4.4: *"Prefect deployment on the existing Prefect
+server; worker only, do not deploy a second server. Trigger: git push to
+content/, plus a nightly audit."*
+
+Prefect is optional at runtime. If it is not installed, these functions are
+plain callables and the same code runs straight from the CLI
+(`python3 -m corpus.cli sync`), which is what `make corpus` does. Nothing about
+the pipeline depends on the orchestrator, and the orchestrator depends on
+nothing but the pipeline.
+
+The deployment commands (run on `.13`, which already has the Prefect server) are
+in `docs/phase-4-notes.md` §6. This file is only the work definition.
+"""
+
+from __future__ import annotations
+
+import sys
+
+try:  # pragma: no cover - exercised only where Prefect is installed
+    from prefect import flow as _flow
+    from prefect import get_run_logger as _get_run_logger
+
+    PREFECT_AVAILABLE = True
+except ImportError:  # pragma: no cover
+    PREFECT_AVAILABLE = False
+
+    def _flow(*_args, **_kwargs):
+        def decorator(func):
+            return func
+
+        return decorator
+
+    def _get_run_logger():
+        class _Logger:
+            def info(self, message, *args, **kwargs):
+                print(message, file=sys.stderr)
+
+            warning = error = info
+
+        return _Logger()
+
+
+from corpus import freshness, pipeline  # noqa: E402  (after the Prefect shim)
+from corpus.config import Config  # noqa: E402
+
+
+@_flow(name="corpus-sync", log_prints=True)
+def sync_corpus() -> dict:
+    """Incremental refresh of the whole estate. Runs on a push or on demand."""
+    logger = _get_run_logger()
+    cfg = Config.from_env()
+    stats = pipeline.run(cfg, mode="sync", log=logger.info)
+    return stats.to_dict()
+
+
+@_flow(name="corpus-scheduled-ingest", log_prints=True)
+def scheduled_ingest(alert: bool = True) -> dict:
+    """Nightly: harvest, report what changed, then audit and alert on staleness.
+
+    This is the scheduled run the DoD asks for: it reports documents added,
+    updated and removed per platform, and it raises an Apprise alert when a
+    source is missing, unavailable or has stopped changing.
+    """
+    logger = _get_run_logger()
+    cfg = Config.from_env()
+    stats = pipeline.run(cfg, mode="sync", log=logger.info)
+    logger.info(stats.summary())
+    report = freshness.audit(cfg, log=logger.info)
+    logger.info(freshness.render_text(report))
+    if alert and not report["ok"]:
+        try:
+            freshness.alert(report, cfg.apprise_url)
+        except Exception as exc:  # noqa: BLE001 - the alert must not fail the run
+            logger.warning(f"Apprise alert failed: {exc}")
+    return {
+        "changed": {
+            "documents": stats.documents,
+            "added": stats.added,
+            "updated": stats.updated,
+            "removed": stats.removed,
+            "unchanged": stats.unchanged,
+            "embedding_calls": stats.embed_calls,
+            "texts_embedded": stats.texts_embedded,
+            "platforms": stats.platforms,
+            "rejections": stats.rejection_summary,
+            "graph_rebuilt": stats.graph_rebuilt,
+        },
+        "audit": report,
+    }
+
+
+@_flow(name="corpus-rebuild", log_prints=True)
+def rebuild_corpus() -> dict:
+    """From-scratch rebuild. The recovery path: deleting the volume is always safe."""
+    logger = _get_run_logger()
+    cfg = Config.from_env()
+    stats = pipeline.run(cfg, mode="rebuild", log=logger.info)
+    return stats.to_dict()
+
+
+@_flow(name="corpus-audit", log_prints=True)
+def freshness_audit(alert: bool = True) -> dict:
+    """Nightly audit. Reports stale documents and index drift, then alerts."""
+    logger = _get_run_logger()
+    cfg = Config.from_env()
+    report = freshness.audit(cfg, log=logger.info)
+    logger.info(freshness.render_text(report))
+    if alert and not report["ok"]:
+        try:
+            freshness.alert(report, cfg.apprise_url)
+        except Exception as exc:  # noqa: BLE001 - the alert must not fail the audit
+            logger.warning(f"Apprise alert failed: {exc}")
+    return report
+
+
+if __name__ == "__main__":  # pragma: no cover - convenience for a manual run
+    command = sys.argv[1] if len(sys.argv) > 1 else "sync"
+    if command == "audit":
+        print(freshness.render_text(freshness_audit(alert=False)))
+    elif command == "rebuild":
+        print(rebuild_corpus())
+    elif command == "scheduled":
+        print(scheduled_ingest(alert=False))
+    else:
+        print(sync_corpus())

platforms/gitea.yml

@@ -0,0 +1,46 @@
+# Platform guide — Gitea.
+#
+# This file is both configuration and public copy. The site shows it so a
+# visitor can see exactly what the assistant reads. Keep the wording honest.
+name: gitea
+enabled: true
+what: >-
+  Self-hosted Git. The source of truth for every current project: README files,
+  file trees, selected source files, recent commit messages and language
+  statistics. This is where the work itself lives.
+fits: [ai, platform, edge, products]
+base: https://gitea.lab.audasmedia.com.au
+auth: GITEA_TOKEN
+include:
+  repos: all
+  # No path include means: every text file the repository rules admit.
+  paths: []
+exclude:
+  repos:
+    # The vault is read by the obsidian guide, so its notes are not harvested
+    # twice under a different platform.
+    - sam/obsidian-vault
+  paths:
+    - "**/node_modules/**"
+    - "**/.venv/**"
+    - "**/venv/**"
+    - "**/.git/**"
+    - "**/.trash/**"
+    - "**/.obsidian/**"
+    - "**/.pi/**"
+    - "**/__pycache__/**"
+    - "**/graphify-out/**"
+    - "**/graphify-out-temp/**"
+    - "**/vendor/**"
+    - "**/third_party/**"
+    - "**/*.min.js"
+    - "**/*.min.css"
+    - "**/*.lock"
+read: [readme, tree, selected_source, commits, languages]
+public: true
+limits:
+  max_documents: 4000
+  max_file_kb: 200
+  max_files_per_repo: 40
+  max_commits: 20
+  max_tree_entries: 3000

platforms/graphify.yml

@@ -0,0 +1,23 @@
+# Platform guide — Graphify knowledge graph.
+#
+# This file is both configuration and public copy.
+name: graphify
+enabled: true
+what: >-
+  A knowledge graph over the notes. It answers "what connects X to Y" rather
+  than "what mentions X": nodes are the things in the estate, edges are the
+  relations between them, and every one points back at the note it came from.
+fits: [ai]
+base: sam/obsidian-vault
+read: [graph_json]
+public: true
+limits:
+  max_documents: 4000
+  # graph.json is the only file this platform reads. These are the places it may
+  # be committed; the first match wins.
+  graph_paths:
+    - graphify-out/graph.json
+    - graph.json
+    - Graph/graph.json
+  max_nodes: 20000
+  max_edges: 40000

platforms/obsidian.yml

@@ -0,0 +1,44 @@
+# Platform guide — Obsidian vault (sam/obsidian-vault, read through Gitea).
+#
+# This file is both configuration and public copy.
+#
+# Privacy note: only project, area and index notes are read. The personal and
+# family folders (000 daily, 400 Personal Family, 500 archive) and the templates
+# are deliberately out of scope. They are not work evidence and they contain
+# third-party personal detail. See docs/content-redaction.md.
+name: obsidian
+enabled: true
+what: >-
+  A personal knowledge base. High-level notes on every system, in Sam's own
+  words: what a system is for, what changed, and what went wrong. Project and
+  area notes only; personal, family and daily notes are not read.
+fits: [about, ai, platform, edge, products]
+base: sam/obsidian-vault
+include:
+  paths:
+    - "010 inbox/**"
+    - "100 Table of Contents/**"
+    - "200 projects/**"
+    - "300 areas/**"
+exclude:
+  paths:
+    - ".obsidian/**"
+    - ".trash/**"
+    - ".venv/**"
+    - "templates/**"
+    - "000 daily/**"
+    - "400 Personal Family/**"
+    - "500 archive/**"
+    - "**/*.excalidraw"
+    - "**/*.png"
+    - "**/*.jpg"
+    - "**/*.jpeg"
+    - "**/*.gif"
+    - "**/*.webp"
+    - "**/*.pdf"
+    - "**/*Bedroom Lights*"
+read: [notes]
+public: true
+limits:
+  max_documents: 4000
+  max_file_kb: 200

reference/stack.md

@@ -36,6 +36,25 @@
 | Guardrail (if local) | 300 m |
 | **Total ceiling** | **1.0 GiB / 4 containers** |
 
+### Qdrant memory, measured (2026-10-26, v1.20.0, 400 m cap, 1024 dims)
+
+| Chunks | Settled RSS | Peak during ingest |
+|---|---|---|
+| 500 | 128 MiB | 149 MiB |
+| 5,000 | 138 MiB | 305 MiB |
+| 20,000 | 151 MiB | 292 MiB |
+
+Within the cap at all three sizes, but only with the memory-frugal collection
+settings in `corpus/qdrant.py`: HNSW graph and payloads on disk
+(`hnsw_config.on_disk`, `on_disk_payload`), `m=8`, `ef_construct=64`,
+`memmap_threshold=5000`, `indexing_threshold=20000`. With Qdrant's defaults the
+container is OOM-killed between 5,000 and 20,000 chunks. The pipeline is the only
+code that creates the collection, so those settings cannot drift.
+
+Ceiling: 20,000 chunks at 1024 dimensions. Beyond that, set
+`RESUME_QDRANT_VECTORS_ON_DISK=1`, reduce the dimension, or raise the cap — in
+that order. Detail: `docs/phase-4-notes.md` §7.
+
 Every service sets `mem_limit` **and** `memswap_limit` to the same value so it cannot consume host swap. Bind to `127.0.0.1`; Caddy is the only public listener.
 
 ## Paths
@@ -56,6 +75,6 @@ Every service sets `mem_limit` **and** `memswap_limit` to the same value so it c
 
 | Item | Owner |
 |---|---|
-| Embedding provider (Gemini free tier vs OpenRouter vs local) | Sam — decide before Phase 4 |
+| Embedding provider (Gemini free tier vs OpenRouter vs local) | Sam — Phase 4 shipped provider-agnostic (`corpus/embeddings.py`); the `omniroute` adapter is ready, `local-hashing` is a placeholder, so the decision is still Sam's before Phase 5 answers questions |
 | Repo public/private after launch | Sam — decide after Phase 3 |
 | Phone number on the contact page | Sam |