Once the port is published on 192.168.20.13 the host no longer listens on
127.0.0.1:8096, so wait_healthy(), do_status() and the rollback path would all
get connection-refused and every deploy would roll back. They now probe
$PUBLISH_IP (RESUME_PUBLISH_IP, default 192.168.20.13). assert_shape() asserts
the published host IP is that address instead of loopback; the mem_limit ==
memswap_limit checks are unchanged. check_port stays address-agnostic on
`sport = :8096`.
Records the amendment in reference/stack.md (ports table + the bind rule) and
docs/phase-3-deploy.md (§3.2 bullet, flow diagram, DoD line), citing Sam's
decision of 2026-10-26, interaction 54aeb1a2. AGENTS.md is Sam's to change and
is untouched.
- docs/phase-4-corpus.md rewritten: Gitea (all repos), vault via Gitea, Graphify
graph, curated content as tier 1. Platform guides are config and public copy.
- docs/phase-5-ai-overlay.md: citations to originals, tier preference, live
source readout.
- Board: AUD-144 (index content/ only) retired; AUD-147 replaces it, blocked by
AUD-141 and AUD-142. AUD-145 relinked to AUD-143 and AUD-147.
Satisfies docs/phase-4-corpus.md. content/ stays the single source of truth:
one pass over it writes both indexes, so the static pages and the retrieval
corpus cannot drift.
- corpus/: stdlib-only pipeline. validate -> chunk -> enrich -> embed -> upsert.
Chunking is block-based, so a boundary cannot fall inside a fenced block, a
table or front-matter (verified structurally, not by inspection).
- Deterministic ids: sha256(source_path + chunk_index + content_hash), rendered
as a UUID. Re-running updates points; re-chunking deletes the old ones by
source filter, so duplicates and orphans are both impossible.
- Incremental by content hash against a SQLite manifest. A provider, dimension
or chunker change is refused instead of mixed: it needs a rebuild.
- Payloads per section 4.2 plus provenance, validated before any vector is
written; a missing field fails the run and names the document and chunk.
- SQLite FTS5 over chunks (the keyword half of hybrid retrieval, decision D10)
with area/confidence filters and a `terms` column so front-matter tech and
tags are findable.
- Freshness audit: stale documents, due soon, index drift, contract drift;
optional Apprise alert; exit code for Prefect.
- Embeddings behind one interface: omniroute (OpenAI-compatible, retries, key
from the environment only) and local-hashing (offline placeholder). The
provider decision is still Sam's; the pipeline does not depend on it.
- deploy/qdrant.yml: mem_limit = memswap_limit = 400m, loopback only.
- docs/phase-4-notes.md: operator manual, DoD evidence, measured memory.
- 76 unit tests, no container and no network.
Measured, Qdrant 1.20.0 under a 400 m cap at 1024 dimensions: 500 / 5,000 /
20,000 chunks settle at 128 / 138 / 151 MiB, peaking at 149 / 305 / 292 MiB
during ingest. Qdrant's default settings are OOM-killed between 5,000 and
20,000, so the collection is created with HNSW and payloads on disk; recorded in
reference/stack.md with the ceiling and the escalation order.
Note: the `make corpus*` targets were swept into the Phase 3 deploy commit
24590cc by concurrent work in this tree; the Makefile content is correct, and
this commit contains the pipeline they call.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Satisfies the Phase 1 Definition of Done:
- content/ validates cleanly against the frozen schema. `make validate`
reports OK: 40 document(s) validated, 0 violations.
- Six areas exist, each with an L1 entry: ai-overview, platform-overview,
edge-overview, products-overview, history-overview, about-overview.
- All six areas have an L2 entry: ai-systems, platform-estate, edge-fleet,
products-business, history-engagements, about-narrative.
- 28 L3 artefact stubs under content/<area>/l3/. Every artefacts[].url in
every document resolves to one of them; the validator proves there are no
dangling references (and a deliberately broken copy exits 1).
- Every metric value traces to a source line in the evidence notes or the
approved metric set in reference/content-map.md (per-metric citations are
recorded in the AUD-131 completion comment).
- Every document carries confidence and last_verified=2026-10-26; the
validator rejects a missing or non-canonical value.
- docs/content-redaction.md records the redaction pass: identifiers, third
party personal data, credentials and infrastructure detail excluded, with
reasons, plus the two numbers that are reported rather than smoothed.
Honesty positions held: 3 businesses active (12 adoptions over time),
Where-Woof deployed and pre-market, failures kept in (edge-failures), no
promotion of in-progress work to delivered.
No site code, no templates, no CSS, no rendering (Phase 2).
Implements AGENTS.md section 7 front-matter schema as a single-file,
standard-library-only Python validator; no third-party YAML dependency
and no framework. Adds the `make validate` gate, a 20-case fixture set
and docs/content-schema.md.
DoD satisfied:
- make validate runs and exits 0 on an empty/clean content/ tree
(missing dir and empty dir both exit 0; fixture ok-empty covers it).
- Every violation class has a fixture under test/fixtures/fail-*; each
exits non-zero with one line per violation (file:line: field: rule).
`make test` runs all 20 cases: 20/20 pass.
- Dangling artefact reference is fatal (fail-dangling-artefact;
artefacts[].url must resolve to an existing in-repo L3 stub or asset).
- Unknown front-matter field is fatal (fail-unknown-field), including
unknown keys nested inside artefacts[]/metrics[].
- docs/content-schema.md committed; states the frozen schema, links to
AGENTS.md section 7, and records that any later schema change is a
breaking change requiring Sam's approval.
- Scope respected: no content authored, no rendering, no site code.
Files: Makefile, tools/validate_content.py, test/fixtures/**,
docs/content-schema.md, .gitignore (__pycache__).
