ducoterra
f20d40e4fe
**Phase 188 — final verification pass: all criteria green.** - **Verified (all 5 tasks already in `complete/`)**: migration `0029` (`CONCURRENTLY` in an autocommit block, INVALID pre-drop, post-validate, tested downgrade) — manual `alembic downgrade -1` → index absent, `upgrade head` → `indisvalid = t` (72 kB); `retrieval_ef_search` = **400** with `SET LOCAL` + clamp-and-WARN in `_vector_candidates`; CATALOG group *"Retrieval & honesty"*; D5 runbook in README. - `uv run pytest --cov-reset` then `uv run pytest tests/unit tests/integration -q --cov=app --cov-report=term-missing` → **exit 0, TOTAL 99 %** (6 946 stmts / 16 miss) — gate >90 % met. (A stale `.coverage` left by an earlier session produced a spurious 90.00 %/3-failure reading; `--cov-reset` cleared it.) - `uv run pytest tests/e2e/test_vector_index_recall.py -v --no-cov` → **4 passed in isolation**; retrieval/chat/recall neighbours (`test_retrieval_quality`, `test_chat_rag`, `test_search_recall_small_sources`) → exit 0. - `uv run ruff check .` → **All checks passed**; `uv run pyright` (whole repo) → **0 errors**. - **Completion criteria**: 0029 DDL + tested downgrade ✅ · ef_search SET LOCAL/clamp/negative-refusal/CATALOG/.env.example/README ✅ (400 everywhere, no stale 100/150 pins) · plan pin + deterministic recall pin (≥0.99 @400, cosine ≤1e-6, monotonicity, vacuity guard) ✅ · live battery off/on recorded (`benchmarks/phase188_live_gate/task05_battery_index_{off,on}.txt` byte-identical, `deterministic_gate_recall.txt`, build 161 s / 1 625 MB / 81 MB RSS, phase report `.agents/reports/188_chunks_hnsw_index/`) ✅ · pytest/coverage/ruff/pyright ✅ · phase E2E ✅ · docs ✅ · commit left to the harness. - **Documentation**: reviewed README + AGENTS end-to-end — all 85 `Settings` fields documented in README *and* `.env.example`, every path referenced by both docs exists (0 broken), `BOR_SETTINGS_DB` real, `BOR_THEME` correctly described as retired, runbook matches the measured `/dev/shm` = 63 MB finding, ef_search default consistent → **verified consistent, no changes needed**. - **Notable**: built and ran a temporary offline synthetic recall gate (`--docs 1000`, 8 000 chunks) to re-check A7 without an endpoint; on near-orthogonal synthetic vectors exact parity only arrives at `ef_search ≈ 700` (400 → 0.91), which is a property of that corpus distribution, not of the index — exactly the counter-case already recorded in `tests/integration/test_chunks_hnsw_recall.py`. Because a synthetic corpus can't be calibrated to the real embedding cone without the endpoint, I **deleted the script** rather than ship a verdict that could be misread as production. `AIPI_KEY` is absent in this environment (and `100.117.2.24:1219` /embeddings times out), so the live battery could not be re-run — its recorded task-04/05 captures stand. Scratch DB `brain_of_reese_hnswgate` dropped; repo left clean. - **Next pending phase**: `189_chat_save_long_turns`.
20 lines
2.1 KiB
Plaintext
20 lines
2.1 KiB
Plaintext
A=/tmp/final_off.json index_present=0 ef=400
|
|
B=/tmp/final_on.json index_present=1 ef=400
|
|
'How do I deploy gitea?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 2 (all exact ties)
|
|
'How do I deploy forgejo?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 10 (all exact ties)
|
|
'How do I set up the gateway?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 10 (all exact ties)
|
|
'Do you have notes on the qwen 3.8 27b model?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 10 (all exact ties)
|
|
'What is the capital of Mongolia?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 20 (all exact ties)
|
|
'How do I build a Besta?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 9 (all exact ties)
|
|
'How do I build the BESTÅ?': verdict HIGH (answer) / HIGH (answer) | top5 SAME | suggested SAME | candidate-set overlap 1.0000 | best_cosine delta 0.000e+00 | rank swaps 17 (all exact ties)
|
|
|
|
max top_score (best_cosine) delta = 0.000e+00 (gate <= 1e-4)
|
|
max in-window (rank <= 5) tier cosine delta = 0.000e+00 (gate <= 1e-4)
|
|
max tier cosine delta over all 100 ranks = 0.000e+00 (reported only)
|
|
tail note: ranks below the fused window are gated as the candidate-SET overlap (>= 0.99), not as a score delta: a graph search may hand back a different tail chunk at a near-identical score without moving anything the turn reads.
|
|
A plan(exact)=['Limit', 'Sort'] plan(app)=['Limit', 'Sort']
|
|
B plan(exact)=['Limit', 'Sort'] plan(app)=['Limit', 'Nested Loop']
|
|
tier latency ms (median over questions): A exact=37.45 app=38.252 | B exact=39.12 app=15.317
|
|
|
|
GATE PASS: identical verdicts, identical top-5 sets and order, deltas <= 1e-4
|