b2379e3275
Module 5 phase 4 step 1 (docs/kb/modules/05-faza4-plan.md, §4): first user-facing HTTP entry point to the KB. FastAPI wrapping kb_retrieval.cascade_query/flat_query — GET /search (query_text -> embed via Ollama@SOLARIA -> cascade/flat -> envelope join -> JSON with per-source links) and GET /healthz. Search API only, no answer synthesis (phase 5) and no server-side dist filtering — the 0.45/0.55 colour thresholds are a frontend concern (plan §7, a later step). Hard startup invariant (plan §2 decision 2): refuses to start unless the configured EMBED_MODEL is present in both document_chunk.model and document_summary.embedding_model. Note the latter: document_summary.model is the LLM that *wrote* the summary (claude-haiku-4-5/gemma3:12b), not the embedder — checked live against kb-postgres@PIHA before writing this, see app/startup.py's docstring. Verified end-to-end with a live docker run: the invariant crash-loops on a mismatched EMBED_MODEL and passes through to a real /search hit against the live corpus with a correct model. Repo-only: no deploy, no npm/OIDC/DNS wiring (plan §8, later step), no local embed fallback (plan §5, later step) — Ollama@SOLARIA is called directly and a failure surfaces as 503, not a crash. Also: scripts/deploy/deploy.sh's gate now builds each service via `docker compose build` instead of a raw `docker build <svc_dir>`, so a service whose docker-compose.yml declares a repo-root build context (needed here to COPY packages/kb-retrieval/, the packages/ Dockerfile convention already documented in CLAUDE.md) resolves the same way in the gate as it does at real deploy time (deploy-node.sh's `docker compose ... up --build`). No behavior change for existing single-context services — verified against llm-gateway's compose file. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
57 lines
2.5 KiB
Python
57 lines
2.5 KiB
Python
"""Unit tests for the startup model invariant -- no real DB."""
|
|
from __future__ import annotations
|
|
|
|
import pathlib
|
|
import sys
|
|
|
|
import pytest
|
|
|
|
sys.path.insert(0, str(pathlib.Path(__file__).resolve().parents[1]))
|
|
|
|
from app.startup import ModelInvariantError, validate_embed_model # noqa: E402
|
|
|
|
|
|
class _FakeConn:
|
|
def __init__(self, chunk_models: set, summary_embedding_models: set):
|
|
self._chunk_models = chunk_models
|
|
self._summary_embedding_models = summary_embedding_models
|
|
|
|
async def fetch(self, query, *params):
|
|
if "FROM document_chunk" in query:
|
|
return [{"model": m} for m in self._chunk_models]
|
|
if "FROM document_summary" in query:
|
|
return [{"embedding_model": m} for m in self._summary_embedding_models]
|
|
raise AssertionError(f"unexpected query: {query}")
|
|
|
|
|
|
class TestValidateEmbedModel:
|
|
async def test_passes_when_model_present_in_both_tables(self):
|
|
conn = _FakeConn(chunk_models={"bge-m3"}, summary_embedding_models={"bge-m3"})
|
|
await validate_embed_model(conn, "bge-m3") # must not raise
|
|
|
|
async def test_fails_when_chunk_model_missing(self):
|
|
conn = _FakeConn(chunk_models={"some-other-model"}, summary_embedding_models={"bge-m3"})
|
|
with pytest.raises(ModelInvariantError, match="document_chunk"):
|
|
await validate_embed_model(conn, "bge-m3")
|
|
|
|
async def test_fails_when_chunk_table_empty(self):
|
|
conn = _FakeConn(chunk_models=set(), summary_embedding_models={"bge-m3"})
|
|
with pytest.raises(ModelInvariantError, match="document_chunk"):
|
|
await validate_embed_model(conn, "bge-m3")
|
|
|
|
async def test_fails_when_summary_embedding_model_missing(self):
|
|
conn = _FakeConn(chunk_models={"bge-m3"}, summary_embedding_models={"some-other-model"})
|
|
with pytest.raises(ModelInvariantError, match="document_summary"):
|
|
await validate_embed_model(conn, "bge-m3")
|
|
|
|
async def test_does_not_confuse_summary_writer_model_with_embedding_model(self):
|
|
# Regression guard: document_summary.model is the LLM that WROTE the summary
|
|
# (claude-haiku-4-5/gemma3:12b), never bge-m3 -- checking that column instead of
|
|
# embedding_model would make this invariant impossible to satisfy.
|
|
conn = _FakeConn(
|
|
chunk_models={"bge-m3"},
|
|
summary_embedding_models={"bge-m3"},
|
|
)
|
|
assert conn._summary_embedding_models == {"bge-m3"}
|
|
await validate_embed_model(conn, "bge-m3") # must not raise
|