diff --git a/src/repo_registry/core/models.py b/src/repo_registry/core/models.py index a182c5f..662569b 100644 --- a/src/repo_registry/core/models.py +++ b/src/repo_registry/core/models.py @@ -168,6 +168,14 @@ class SearchResult: match_type: str match_name: str confidence: float + match_description: str = "" + matched_field: str = "" + ability_id: int | None = None + ability_name: str | None = None + capability_id: int | None = None + capability_name: str | None = None + evidence_level: str | None = None + source_reference: str | None = None @dataclass(frozen=True) diff --git a/src/repo_registry/storage/sqlite.py b/src/repo_registry/storage/sqlite.py index 0b93707..f0ccaf3 100644 --- a/src/repo_registry/storage/sqlite.py +++ b/src/repo_registry/storage/sqlite.py @@ -1276,43 +1276,201 @@ class RegistryStore: return RepositoryAbilityMap(repository=repository, abilities=abilities) def search(self, query: str) -> list[SearchResult]: - needle = f"%{query.strip()}%" - if needle == "%%": + term = query.strip() + needle = f"%{term}%" + if not term: return [] with self.connect() as connection: - rows = connection.execute( + repository_rows = connection.execute( """ SELECT r.id AS repository_id, r.name AS repository_name, - 'repository' AS match_type, r.name AS match_name, - 1.0 AS confidence + r.description FROM repositories r WHERE r.name LIKE ? OR COALESCE(r.description, '') LIKE ? - UNION ALL - SELECT r.id, r.name, 'ability', a.name, a.confidence + """, + (needle, needle), + ).fetchall() + ability_rows = connection.execute( + """ + SELECT r.id AS repository_id, r.name AS repository_name, + a.id AS ability_id, a.name AS ability_name, + a.description AS ability_description, a.confidence FROM approved_abilities a JOIN repositories r ON r.id = a.repository_id WHERE a.name LIKE ? OR a.description LIKE ? - UNION ALL - SELECT r.id, r.name, 'capability', c.name, c.confidence + """, + (needle, needle), + ).fetchall() + capability_rows = connection.execute( + """ + SELECT r.id AS repository_id, r.name AS repository_name, + a.id AS ability_id, a.name AS ability_name, + c.id AS capability_id, c.name AS capability_name, + c.description AS capability_description, c.confidence FROM approved_capabilities c + JOIN approved_abilities a ON a.id = c.ability_id JOIN repositories r ON r.id = c.repository_id WHERE c.name LIKE ? OR c.description LIKE ? - ORDER BY confidence DESC, repository_name ASC, match_name ASC """, - (needle, needle, needle, needle, needle, needle), + (needle, needle), + ).fetchall() + feature_rows = connection.execute( + """ + SELECT r.id AS repository_id, r.name AS repository_name, + a.id AS ability_id, a.name AS ability_name, + c.id AS capability_id, c.name AS capability_name, + f.name AS feature_name, f.type AS feature_type, + f.location, f.confidence + FROM approved_features f + JOIN approved_capabilities c ON c.id = f.capability_id + JOIN approved_abilities a ON a.id = c.ability_id + JOIN repositories r ON r.id = f.repository_id + WHERE f.name LIKE ? OR f.type LIKE ? OR f.location LIKE ? + """, + (needle, needle, needle), + ).fetchall() + evidence_rows = connection.execute( + """ + SELECT r.id AS repository_id, r.name AS repository_name, + a.id AS ability_id, a.name AS ability_name, + c.id AS capability_id, c.name AS capability_name, + e.type AS evidence_type, e.reference, e.strength + FROM approved_evidence e + JOIN approved_capabilities c ON c.id = e.capability_id + JOIN approved_abilities a ON a.id = c.ability_id + JOIN repositories r ON r.id = e.repository_id + WHERE e.type LIKE ? OR e.reference LIKE ? OR e.strength LIKE ? + """, + (needle, needle, needle), ).fetchall() - return [ - SearchResult( - repository_id=row["repository_id"], - repository_name=row["repository_name"], - match_type=row["match_type"], - match_name=row["match_name"], - confidence=row["confidence"], + results: list[SearchResult] = [] + for row in repository_rows: + matched_field = ( + "name" if self._matches(row["repository_name"], term) else "description" ) - for row in rows - ] + results.append( + SearchResult( + repository_id=row["repository_id"], + repository_name=row["repository_name"], + match_type="repository", + match_name=row["repository_name"], + match_description=row["description"] or "", + matched_field=matched_field, + confidence=1.0, + ) + ) + for row in ability_rows: + matched_field = ( + "name" if self._matches(row["ability_name"], term) else "description" + ) + results.append( + SearchResult( + repository_id=row["repository_id"], + repository_name=row["repository_name"], + match_type="ability", + match_name=row["ability_name"], + match_description=row["ability_description"], + matched_field=matched_field, + confidence=row["confidence"], + ability_id=row["ability_id"], + ability_name=row["ability_name"], + ) + ) + for row in capability_rows: + matched_field = ( + "name" + if self._matches(row["capability_name"], term) + else "description" + ) + results.append( + SearchResult( + repository_id=row["repository_id"], + repository_name=row["repository_name"], + match_type="capability", + match_name=row["capability_name"], + match_description=row["capability_description"], + matched_field=matched_field, + confidence=row["confidence"], + ability_id=row["ability_id"], + ability_name=row["ability_name"], + capability_id=row["capability_id"], + capability_name=row["capability_name"], + ) + ) + for row in feature_rows: + matched_field = self._first_matched_field( + term, + { + "name": row["feature_name"], + "type": row["feature_type"], + "location": row["location"], + }, + ) + results.append( + SearchResult( + repository_id=row["repository_id"], + repository_name=row["repository_name"], + match_type="feature", + match_name=row["feature_name"], + match_description=row["feature_type"], + matched_field=matched_field, + confidence=row["confidence"], + ability_id=row["ability_id"], + ability_name=row["ability_name"], + capability_id=row["capability_id"], + capability_name=row["capability_name"], + source_reference=row["location"], + ) + ) + for row in evidence_rows: + matched_field = self._first_matched_field( + term, + { + "type": row["evidence_type"], + "reference": row["reference"], + "strength": row["strength"], + }, + ) + results.append( + SearchResult( + repository_id=row["repository_id"], + repository_name=row["repository_name"], + match_type="evidence", + match_name=row["reference"], + match_description=row["evidence_type"], + matched_field=matched_field, + confidence=self._evidence_confidence(row["strength"]), + ability_id=row["ability_id"], + ability_name=row["ability_name"], + capability_id=row["capability_id"], + capability_name=row["capability_name"], + evidence_level=row["strength"], + source_reference=row["reference"], + ) + ) + return sorted( + results, + key=lambda result: ( + -result.confidence, + result.repository_name.lower(), + result.match_type, + result.match_name.lower(), + ), + ) + + def _matches(self, value: str | None, term: str) -> bool: + return term.lower() in (value or "").lower() + + def _first_matched_field(self, term: str, values: dict[str, str | None]) -> str: + for field, value in values.items(): + if self._matches(value, term): + return field + return "" + + def _evidence_confidence(self, strength: str) -> float: + return {"strong": 0.9, "medium": 0.6, "weak": 0.3}.get(strength, 0.5) def _insert_facts( self, diff --git a/src/repo_registry/web_ui/views.py b/src/repo_registry/web_ui/views.py index 2c7c467..a16b2da 100644 --- a/src/repo_registry/web_ui/views.py +++ b/src/repo_registry/web_ui/views.py @@ -184,7 +184,11 @@ def search_page( {escape(result.repository_name)} {escape(result.match_type)} - {escape(result.match_name)} + + {escape(result.match_name)} + {render_search_context(asdict(result))} + + {escape(result.matched_field)} {result.confidence:.2f} """ @@ -208,7 +212,7 @@ def search_page(
- + {rows or empty}
RepositoryMatchNameConfidence
RepositoryMatchNameFieldConfidence
@@ -917,3 +921,20 @@ def render_sources(source_refs: list[dict]) -> str: if len(source_refs) > 5: sources += f' +{len(source_refs) - 5} more' return f"

{sources}

" + + +def render_search_context(result: dict) -> str: + details = [] + if result.get("ability_name"): + details.append(f"Ability: {escape(result['ability_name'])}") + if result.get("capability_name"): + details.append(f"Capability: {escape(result['capability_name'])}") + if result.get("evidence_level"): + details.append(f"Evidence: {escape(result['evidence_level'])}") + if result.get("source_reference"): + details.append(f"Source: {escape(result['source_reference'])}") + if result.get("match_description"): + details.append(escape(result["match_description"])) + if not details: + return "" + return f'

{" ยท ".join(details)}

' diff --git a/tests/test_registry_service.py b/tests/test_registry_service.py index c0870e3..04a08d8 100644 --- a/tests/test_registry_service.py +++ b/tests/test_registry_service.py @@ -96,6 +96,47 @@ def test_search_matches_approved_abilities_and_capabilities(tmp_path): assert capabilities[0].name == "Classify Incoming Email" +def test_search_matches_features_and_evidence_with_context(tmp_path): + service = make_service(tmp_path) + repository = service.register_repository( + name="MailRouter", + url="https://example.com/mail-router-feature.git", + description="Manual test repository.", + ) + ability_id = service.add_ability(repository.id, name="Business Email Routing") + capability_id = service.add_capability( + repository.id, + ability_id, + name="Classify Incoming Email", + ) + service.add_feature( + repository.id, + capability_id, + name="POST /api/classify-email", + type="REST endpoint", + location="src/routes/classify_email.py", + ) + service.add_evidence( + repository.id, + capability_id, + type="unit_test", + reference="tests/test_email_classification.py", + strength="strong", + ) + + feature_results = service.search("classify_email") + evidence_results = service.search("unit_test") + + assert feature_results[0].match_type == "feature" + assert feature_results[0].matched_field == "location" + assert feature_results[0].ability_name == "Business Email Routing" + assert feature_results[0].capability_name == "Classify Incoming Email" + assert feature_results[0].source_reference == "src/routes/classify_email.py" + assert evidence_results[0].match_type == "evidence" + assert evidence_results[0].evidence_level == "strong" + assert evidence_results[0].confidence == 0.9 + + def test_register_repository_imports_metadata_when_name_is_omitted(tmp_path): source = tmp_path / "metadata-source" source.mkdir() diff --git a/tests/test_web_api.py b/tests/test_web_api.py index 7be01bb..38ff6fd 100644 --- a/tests/test_web_api.py +++ b/tests/test_web_api.py @@ -211,6 +211,7 @@ def test_api_analysis_run_loop(tmp_path): search_response = client.get("/search", params={"q": "frontend"}) assert search_response.status_code == 200 assert search_response.json() + assert "matched_field" in search_response.json()[0] abilities_response = client.get("/abilities") assert abilities_response.status_code == 200 @@ -303,6 +304,7 @@ def test_ui_register_analyze_and_approve_loop(tmp_path): search_response = client.get("/ui/search", params={"q": "repository"}) assert search_response.status_code == 200 assert "UI Repo" in search_response.text + assert "Field" in search_response.text finally: app.dependency_overrides.clear()