130 lines
3.8 KiB
Python
130 lines
3.8 KiB
Python
"""Contract test for POST /search.
|
|
|
|
The hybrid search backend depends on live OpenSearch + Qdrant + embedder; we
|
|
patch :func:`app.indexing.hybrid_search.run_search` so the route can be
|
|
exercised with the real request/response schemas without bringing infra up.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import uuid
|
|
|
|
import pytest
|
|
from fastapi.testclient import TestClient
|
|
|
|
from app.api.schemas import (
|
|
Citation,
|
|
SearchHit,
|
|
SearchResponse,
|
|
)
|
|
from app.config import settings
|
|
from app.main import app
|
|
|
|
|
|
@pytest.fixture
|
|
def client() -> TestClient:
|
|
return TestClient(app)
|
|
|
|
|
|
def _stub_response(query: str) -> SearchResponse:
|
|
doc_id = uuid.uuid4()
|
|
chunk_id = uuid.uuid4()
|
|
return SearchResponse(
|
|
query=query,
|
|
mode="hybrid",
|
|
total_candidates=42,
|
|
reranked=True,
|
|
results=[
|
|
SearchHit(
|
|
rank=1,
|
|
score=0.91,
|
|
document_id=doc_id,
|
|
chunk_id=chunk_id,
|
|
original_file_name="GOST_21.501-93.pdf",
|
|
source_path="/data/input/standards/GOST_21.501-93.pdf",
|
|
page_number=12,
|
|
block_type="paragraph",
|
|
text=f"Highlighted text for {query}.",
|
|
citation=Citation(
|
|
pdf="GOST_21.501-93.pdf",
|
|
page=12,
|
|
block_id="b-12-0",
|
|
table_id=None,
|
|
figure_id=None,
|
|
),
|
|
quality_flags={"low_ocr_confidence": False, "needs_manual_review": False},
|
|
metadata={"section_heading": "Глава 2"},
|
|
)
|
|
],
|
|
)
|
|
|
|
|
|
def test_search_returns_hit_with_citation(client: TestClient, monkeypatch):
|
|
from app.api import routes_search as routes
|
|
from app.indexing import hybrid_search
|
|
|
|
def fake_run(req):
|
|
assert req.query
|
|
return _stub_response(req.query)
|
|
|
|
monkeypatch.setattr(hybrid_search, "run_search", fake_run)
|
|
# The route imports run_search lazily; patch the module-level binding too.
|
|
monkeypatch.setattr(routes, "run_search", fake_run, raising=False)
|
|
|
|
res = client.post(
|
|
f"{settings.app_api_prefix}/search",
|
|
json={
|
|
"query": "ГОСТ 21.501-93",
|
|
"limit": 10,
|
|
"filters": {
|
|
"document_id": None,
|
|
"source_path": None,
|
|
"block_type": None,
|
|
"min_ocr_confidence": None,
|
|
},
|
|
"search_mode": "hybrid",
|
|
},
|
|
)
|
|
|
|
assert res.status_code == 200, res.text
|
|
body = res.json()
|
|
assert body["query"] == "ГОСТ 21.501-93"
|
|
assert body["mode"] == "hybrid"
|
|
assert body["reranked"] is True
|
|
assert body["total_candidates"] == 42
|
|
assert len(body["results"]) == 1
|
|
|
|
hit = body["results"][0]
|
|
assert hit["rank"] == 1
|
|
assert hit["page_number"] == 12
|
|
assert hit["block_type"] == "paragraph"
|
|
assert hit["citation"]["pdf"] == "GOST_21.501-93.pdf"
|
|
assert hit["citation"]["page"] == 12
|
|
assert hit["citation"]["block_id"] == "b-12-0"
|
|
|
|
|
|
def test_search_rejects_empty_query(client: TestClient, monkeypatch):
|
|
"""Schema validation should reject empty query without hitting the backend."""
|
|
from app.indexing import hybrid_search
|
|
|
|
def must_not_run(_req): # noqa: ARG001
|
|
raise AssertionError("backend should not be called for invalid input")
|
|
|
|
monkeypatch.setattr(hybrid_search, "run_search", must_not_run)
|
|
|
|
res = client.post(
|
|
f"{settings.app_api_prefix}/search",
|
|
json={
|
|
"query": "",
|
|
"limit": 10,
|
|
"filters": {
|
|
"document_id": None,
|
|
"source_path": None,
|
|
"block_type": None,
|
|
"min_ocr_confidence": None,
|
|
},
|
|
"search_mode": "hybrid",
|
|
},
|
|
)
|
|
assert res.status_code == 422
|