 ducoterra
|
396e4d47fb
|
feat(rag): stream grounded RAG answers over SSE with source citations
Phase 03 (Story: Chat RAG Answer — happy path):
- app/rag/retriever.py: top-k cosine search + parent-doc selection with
per-doc dedupe and BOR_MAX_CONTEXT_CHARS cap ([…truncated…] marker)
- app/rag/prompts.py: locked persona + HIGH/DEFLECT prompt builders
- app/rag/llm.py: LLMError + chat_stream (turbo, temp 0.4, max 700, stream)
- app/api/chat.py: POST /api/chat SSE — delta* then done{deflected,
sources, suggestions}; query_log row + PLAN §9 per-turn log line;
structured error event on mid-stream failure, JSON 503 when DB down
- frontend: SSE reader, live bubble streaming, source chips -> /sources.html,
red role=alert banner, Send button state that always recovers
- fix(scaffold): [hidden] { display: none !important } — .kb-banner's
display:flex was overriding the hidden attribute (banner always visible)
- tests: unit (retriever/prompts/sse/llm) + integration (real Postgres RAG
turn, query_log, error + 503 paths, mid-turn failures) + Playwright story
suite (grounded answer, log row, raw SSE shape); smoke placeholder test
replaced with the real never-stale-button contract
|
2026-08-21 17:17:02 -04:00 |
|
|
ducoterra
|
99c48cbe06
|
feat(rag): index markdown KB — chunker, embed client, delta importer, Sources page
Phase 02 (story: import documents):
- fence-aware markdown chunker (heading sections, 200-char overlap,
heading anchor on every chunk, 1200-char hard cap, fence blocks
kept atomic and split under the cap)
- LLMClient over aipi (LiteLLM) reusing the openai client's httpx
transport to send a clean {model, input} payload — the openai SDK
injects encoding_format, which aipi's openai_like group rejects;
token-budget batching + halving retry for the endpoint's
~1024-token per-request input cap
- two-phase per-file upsert importer: sha256 delta (unchanged skip),
atomic commit, A9 exclusion walk, per-source prune, per-file error
tolerance (rollback + log + continue, non-zero CLI exit), adaptive
re-chunk at half target for URL-dense files the endpoint rejects
- scripts/import_docs CLI (repeatable --source, --prune, --limit,
defaults ~/Homelab + ~/Deployments)
- GET /api/docs with per-doc chunk counts; Sources page wired to the
real endpoint (stat cards, full-width a11y table, designed empty
state, DOM-built rows — no innerHTML)
- tests: 63 passed (chunker/llm/importer units, docs API + importer
integration), story E2E 3/3 (real endpoints, in-thread import);
app/ coverage 98%
- real KB imported: 672 docs / 8969 chunks in ~3m, idempotent
re-run (672 unchanged, 0 batches)
- harness: .agent/validate.sh now gates through uv (pytest +
coverage >90% + ruff + pyright) instead of system python3
|
2026-08-21 16:24:45 -04:00 |
|
|
ducoterra
|
022da8e2bc
|
feat: scaffold Brain of Reese — FastAPI RAG chat over Postgres 17 + pgvector
Foundation (phase 01, verified):
- FastAPI app: /api/health, /api/suggestions, /api/chat (placeholder),
static frontend served locally (no CDN)
- Postgres 17 + pgvector via db/Containerfile + compose.yaml
(podman compose up -d db), Alembic initial migration (documents,
chunks with vector(768), query_log)
- LLM client targeting https://aipi.reeseapps.com/v1 (turbo/embed);
scripts/llm_probe.py verified models + 768-dim embeddings live
- Conditional debugpy: imported only when DEBUGPY=1 (attach on demand,
:5678); logging config for clean single-line logs
- Frontend shell: mobile-first chat + Sources pages, tokens, a11y baselines
- Tests: 24 unit+integration (99% coverage on app/), ruff + pyright clean,
Playwright smoke E2E (3 tests) against a deterministic mock LLM
- Planning: .agent/PLAN.md (architecture + LOCKED decisions), AGENTS.md,
6 user stories, 7 phase files (one story / one phase / one Playwright
suite each)
|
2026-08-21 13:42:21 -04:00 |
|