test_musehub_search.py
python
| 1 | """Tests for MuseHub search endpoints. |
| 2 | |
| 3 | Covers semantic similarity search: |
| 4 | - GET /musehub/search/similar?commit={sha} returns ranked results |
| 5 | - Private repos are excluded from results (public_only enforced) |
| 6 | - 404 when commit SHA is not found |
| 7 | - 503 when Qdrant is unavailable |
| 8 | - Results are sorted by descending score |
| 9 | - Unauthenticated requests return 401 |
| 10 | |
| 11 | Covers cross-repo global search: |
| 12 | - test_global_search_page_renders — GET /musehub/ui/search returns 200 HTML |
| 13 | - test_global_search_results_grouped — JSON results are grouped by repo |
| 14 | - test_global_search_public_only — private repos are excluded |
| 15 | - test_global_search_json — JSON content-type returned |
| 16 | - test_global_search_empty_query_handled — graceful response for empty result set |
| 17 | - test_global_search_requires_auth — 401 without JWT |
| 18 | - test_global_search_keyword_mode — keyword mode matches across message terms |
| 19 | - test_global_search_pattern_mode — pattern mode uses SQL LIKE |
| 20 | - test_global_search_pagination — page/page_size params respected |
| 21 | |
| 22 | Covers in-repo search: |
| 23 | - test_search_page_renders — GET /musehub/ui/{repo_id}/search → 200 HTML |
| 24 | - test_search_keyword_mode — keyword search returns matching commits |
| 25 | - test_search_keyword_empty_query — empty keyword query returns empty matches |
| 26 | - test_search_musical_property — musical property filter works |
| 27 | - test_search_natural_language — ask mode returns matching commits |
| 28 | - test_search_pattern_message — pattern matches commit message |
| 29 | - test_search_pattern_branch — pattern matches branch name |
| 30 | - test_search_json_response — JSON search endpoint returns SearchResponse shape |
| 31 | - test_search_date_range_since — since filter excludes old commits |
| 32 | - test_search_date_range_until — until filter excludes future commits |
| 33 | - test_search_invalid_mode — invalid mode returns 422 |
| 34 | - test_search_unknown_repo — unknown repo_id returns 404 |
| 35 | - test_search_requires_auth — unauthenticated request returns 401 |
| 36 | - test_search_limit_respected — limit caps result count |
| 37 | |
| 38 | All tests use the shared ``client`` and ``auth_headers`` fixtures from |
| 39 | conftest.py. Qdrant calls are mocked — no live vector database required. |
| 40 | """ |
| 41 | from __future__ import annotations |
| 42 | |
| 43 | import uuid |
| 44 | from datetime import datetime, timezone |
| 45 | from unittest.mock import MagicMock, patch |
| 46 | |
| 47 | import pytest |
| 48 | from httpx import AsyncClient |
| 49 | from sqlalchemy.ext.asyncio import AsyncSession |
| 50 | |
| 51 | from musehub.db.musehub_models import MusehubCommit, MusehubObject, MusehubRepo |
| 52 | from musehub.main import app |
| 53 | from musehub.muse_cli.models import MuseCliCommit, MuseCliSnapshot |
| 54 | from musehub.services.musehub_qdrant import SimilarCommitResult, get_qdrant_client |
| 55 | |
| 56 | |
| 57 | # --------------------------------------------------------------------------- |
| 58 | # Helpers — similarity search |
| 59 | # --------------------------------------------------------------------------- |
| 60 | |
| 61 | |
| 62 | def _make_similar_result( |
| 63 | commit_id: str, |
| 64 | repo_id: str = "repo-pub", |
| 65 | score: float = 0.9, |
| 66 | branch: str = "main", |
| 67 | author: str = "composer", |
| 68 | ) -> SimilarCommitResult: |
| 69 | return SimilarCommitResult( |
| 70 | commit_id=commit_id, |
| 71 | repo_id=repo_id, |
| 72 | score=score, |
| 73 | branch=branch, |
| 74 | author=author, |
| 75 | ) |
| 76 | |
| 77 | |
| 78 | # --------------------------------------------------------------------------- |
| 79 | # Helpers — global search (uses MusehubCommit / MusehubRepo directly) |
| 80 | # --------------------------------------------------------------------------- |
| 81 | |
| 82 | |
| 83 | async def _make_repo( |
| 84 | db_session: AsyncSession, |
| 85 | *, |
| 86 | name: str = "test-repo", |
| 87 | visibility: str = "public", |
| 88 | owner: str = "test-owner", |
| 89 | ) -> str: |
| 90 | """Seed a MuseHub repo and return its repo_id.""" |
| 91 | import re as _re |
| 92 | slug = _re.sub(r"[^a-z0-9]+", "-", name.lower()).strip("-")[:64].strip("-") or "repo" |
| 93 | repo = MusehubRepo(name=name, owner="testuser", slug=slug, visibility=visibility, owner_user_id=owner) |
| 94 | db_session.add(repo) |
| 95 | await db_session.commit() |
| 96 | await db_session.refresh(repo) |
| 97 | return str(repo.repo_id) |
| 98 | |
| 99 | |
| 100 | async def _make_commit( |
| 101 | db_session: AsyncSession, |
| 102 | repo_id: str, |
| 103 | *, |
| 104 | commit_id: str, |
| 105 | message: str, |
| 106 | author: str = "alice", |
| 107 | branch: str = "main", |
| 108 | ) -> None: |
| 109 | """Seed a MusehubCommit for global search tests.""" |
| 110 | commit = MusehubCommit( |
| 111 | commit_id=commit_id, |
| 112 | repo_id=repo_id, |
| 113 | branch=branch, |
| 114 | parent_ids=[], |
| 115 | message=message, |
| 116 | author=author, |
| 117 | timestamp=datetime.now(tz=timezone.utc), |
| 118 | ) |
| 119 | db_session.add(commit) |
| 120 | await db_session.commit() |
| 121 | |
| 122 | |
| 123 | # --------------------------------------------------------------------------- |
| 124 | # Helpers — in-repo search (uses MuseCliCommit / MuseCliSnapshot) |
| 125 | # --------------------------------------------------------------------------- |
| 126 | |
| 127 | |
| 128 | async def _make_search_repo(db: AsyncSession) -> str: |
| 129 | """Seed a minimal MuseHub repo for in-repo search tests; return repo_id.""" |
| 130 | repo = MusehubRepo( |
| 131 | name="search-test-repo", |
| 132 | owner="testuser", |
| 133 | slug="search-test-repo", |
| 134 | visibility="private", |
| 135 | owner_user_id="test-owner", |
| 136 | ) |
| 137 | db.add(repo) |
| 138 | await db.commit() |
| 139 | await db.refresh(repo) |
| 140 | return str(repo.repo_id) |
| 141 | |
| 142 | |
| 143 | async def _make_snapshot(db: AsyncSession, snapshot_id: str) -> None: |
| 144 | """Seed a minimal snapshot so FK constraint on MuseCliCommit is satisfied.""" |
| 145 | snap = MuseCliSnapshot(snapshot_id=snapshot_id, manifest={}) |
| 146 | db.add(snap) |
| 147 | await db.flush() |
| 148 | |
| 149 | |
| 150 | async def _make_search_commit( |
| 151 | db: AsyncSession, |
| 152 | *, |
| 153 | repo_id: str, |
| 154 | message: str, |
| 155 | branch: str = "main", |
| 156 | author: str = "test-author", |
| 157 | committed_at: datetime | None = None, |
| 158 | ) -> MuseCliCommit: |
| 159 | """Seed a MuseCliCommit for in-repo search tests.""" |
| 160 | snap_id = "snap-" + str(uuid.uuid4()).replace("-", "")[:16] |
| 161 | await _make_snapshot(db, snap_id) |
| 162 | commit = MuseCliCommit( |
| 163 | commit_id=str(uuid.uuid4()).replace("-", ""), |
| 164 | repo_id=repo_id, |
| 165 | branch=branch, |
| 166 | snapshot_id=snap_id, |
| 167 | message=message, |
| 168 | author=author, |
| 169 | committed_at=committed_at or datetime.now(timezone.utc), |
| 170 | ) |
| 171 | db.add(commit) |
| 172 | await db.flush() |
| 173 | return commit |
| 174 | |
| 175 | |
| 176 | # --------------------------------------------------------------------------- |
| 177 | # Similarity search — authentication |
| 178 | # --------------------------------------------------------------------------- |
| 179 | |
| 180 | |
| 181 | @pytest.mark.anyio |
| 182 | async def test_similar_search_unknown_commit_returns_404_without_auth( |
| 183 | client: AsyncClient, |
| 184 | ) -> None: |
| 185 | """GET /musehub/search/similar returns 404 for an unknown commit without a token. |
| 186 | |
| 187 | Uses optional_token — the endpoint is public; a non-existent commit → 404. |
| 188 | """ |
| 189 | resp = await client.get("/api/v1/musehub/search/similar?commit=non-existent-commit-id") |
| 190 | assert resp.status_code == 404 |
| 191 | |
| 192 | |
| 193 | # --------------------------------------------------------------------------- |
| 194 | # Similarity search — 404 for unknown commit |
| 195 | # --------------------------------------------------------------------------- |
| 196 | |
| 197 | |
| 198 | @pytest.mark.anyio |
| 199 | async def test_similar_search_returns_404_for_unknown_commit( |
| 200 | client: AsyncClient, |
| 201 | auth_headers: dict[str, str], |
| 202 | ) -> None: |
| 203 | """404 is returned when the query commit SHA does not exist in the Hub.""" |
| 204 | resp = await client.get( |
| 205 | "/api/v1/musehub/search/similar?commit=nonexistent-sha", |
| 206 | headers=auth_headers, |
| 207 | ) |
| 208 | assert resp.status_code == 404 |
| 209 | assert "nonexistent-sha" in resp.json()["detail"] |
| 210 | |
| 211 | |
| 212 | # --------------------------------------------------------------------------- |
| 213 | # Similarity search — successful search |
| 214 | # --------------------------------------------------------------------------- |
| 215 | |
| 216 | |
| 217 | @pytest.mark.anyio |
| 218 | async def test_similar_search_returns_results( |
| 219 | client: AsyncClient, |
| 220 | auth_headers: dict[str, str], |
| 221 | ) -> None: |
| 222 | """Successful search returns ranked SimilarCommitResponse list.""" |
| 223 | create_resp = await client.post( |
| 224 | "/api/v1/musehub/repos", |
| 225 | json={"name": "jazz-ballad", "owner": "testuser", "visibility": "public"}, |
| 226 | headers=auth_headers, |
| 227 | ) |
| 228 | assert create_resp.status_code == 201 |
| 229 | repo_id = create_resp.json()["repoId"] |
| 230 | |
| 231 | commit_id = "abc-jazz-001" |
| 232 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 233 | push_resp = await client.post( |
| 234 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 235 | json={ |
| 236 | "branch": "main", |
| 237 | "headCommitId": commit_id, |
| 238 | "commits": [ |
| 239 | { |
| 240 | "commitId": commit_id, |
| 241 | "parentIds": [], |
| 242 | "message": "Jazz ballad in Db major at 72 BPM", |
| 243 | "timestamp": "2024-01-01T00:00:00Z", |
| 244 | } |
| 245 | ], |
| 246 | "objects": [], |
| 247 | "force": False, |
| 248 | }, |
| 249 | headers=auth_headers, |
| 250 | ) |
| 251 | assert push_resp.status_code == 200 |
| 252 | |
| 253 | mock_results = [ |
| 254 | _make_similar_result("similar-001", score=0.95), |
| 255 | _make_similar_result("similar-002", score=0.87), |
| 256 | ] |
| 257 | mock_qdrant = MagicMock() |
| 258 | mock_qdrant.search_similar.return_value = mock_results |
| 259 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 260 | try: |
| 261 | search_resp = await client.get( |
| 262 | f"/api/v1/musehub/search/similar?commit={commit_id}&limit=5", |
| 263 | headers=auth_headers, |
| 264 | ) |
| 265 | finally: |
| 266 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 267 | |
| 268 | assert search_resp.status_code == 200 |
| 269 | data = search_resp.json() |
| 270 | assert data["queryCommit"] == commit_id |
| 271 | assert len(data["results"]) == 2 |
| 272 | assert data["results"][0]["score"] >= data["results"][1]["score"] |
| 273 | |
| 274 | |
| 275 | @pytest.mark.anyio |
| 276 | async def test_similar_search_public_only_enforced( |
| 277 | client: AsyncClient, |
| 278 | auth_headers: dict[str, str], |
| 279 | ) -> None: |
| 280 | """search_similar is called with public_only=True — private commits excluded.""" |
| 281 | create_resp = await client.post( |
| 282 | "/api/v1/musehub/repos", |
| 283 | json={"name": "public-jazz", "owner": "testuser", "visibility": "public"}, |
| 284 | headers=auth_headers, |
| 285 | ) |
| 286 | assert create_resp.status_code == 201 |
| 287 | repo_id = create_resp.json()["repoId"] |
| 288 | commit_id = "pub-commit-001" |
| 289 | |
| 290 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 291 | await client.post( |
| 292 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 293 | json={ |
| 294 | "branch": "main", |
| 295 | "headCommitId": commit_id, |
| 296 | "commits": [ |
| 297 | { |
| 298 | "commitId": commit_id, |
| 299 | "parentIds": [], |
| 300 | "message": "C major 120 BPM", |
| 301 | "timestamp": "2024-01-01T00:00:00Z", |
| 302 | } |
| 303 | ], |
| 304 | "objects": [], |
| 305 | "force": False, |
| 306 | }, |
| 307 | headers=auth_headers, |
| 308 | ) |
| 309 | |
| 310 | mock_qdrant = MagicMock() |
| 311 | mock_qdrant.search_similar.return_value = [] |
| 312 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 313 | try: |
| 314 | resp = await client.get( |
| 315 | f"/api/v1/musehub/search/similar?commit={commit_id}", |
| 316 | headers=auth_headers, |
| 317 | ) |
| 318 | finally: |
| 319 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 320 | |
| 321 | assert resp.status_code == 200 |
| 322 | call_kwargs = mock_qdrant.search_similar.call_args.kwargs |
| 323 | assert call_kwargs.get("public_only") is True |
| 324 | |
| 325 | |
| 326 | @pytest.mark.anyio |
| 327 | async def test_similar_search_excludes_query_commit( |
| 328 | client: AsyncClient, |
| 329 | auth_headers: dict[str, str], |
| 330 | ) -> None: |
| 331 | """The query commit itself is passed as exclude_commit_id to avoid self-match.""" |
| 332 | create_resp = await client.post( |
| 333 | "/api/v1/musehub/repos", |
| 334 | json={"name": "self-exclude-test", "owner": "testuser", "visibility": "public"}, |
| 335 | headers=auth_headers, |
| 336 | ) |
| 337 | repo_id = create_resp.json()["repoId"] |
| 338 | commit_id = "self-excl-001" |
| 339 | |
| 340 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 341 | await client.post( |
| 342 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 343 | json={ |
| 344 | "branch": "main", |
| 345 | "headCommitId": commit_id, |
| 346 | "commits": [ |
| 347 | { |
| 348 | "commitId": commit_id, |
| 349 | "parentIds": [], |
| 350 | "message": "G minor 85 BPM", |
| 351 | "timestamp": "2024-01-01T00:00:00Z", |
| 352 | } |
| 353 | ], |
| 354 | "objects": [], |
| 355 | "force": False, |
| 356 | }, |
| 357 | headers=auth_headers, |
| 358 | ) |
| 359 | |
| 360 | mock_qdrant = MagicMock() |
| 361 | mock_qdrant.search_similar.return_value = [] |
| 362 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 363 | try: |
| 364 | await client.get( |
| 365 | f"/api/v1/musehub/search/similar?commit={commit_id}", |
| 366 | headers=auth_headers, |
| 367 | ) |
| 368 | finally: |
| 369 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 370 | |
| 371 | call_kwargs = mock_qdrant.search_similar.call_args.kwargs |
| 372 | assert call_kwargs.get("exclude_commit_id") == commit_id |
| 373 | |
| 374 | |
| 375 | @pytest.mark.anyio |
| 376 | async def test_similar_search_503_when_qdrant_unavailable( |
| 377 | client: AsyncClient, |
| 378 | auth_headers: dict[str, str], |
| 379 | ) -> None: |
| 380 | """503 is returned when Qdrant raises an exception.""" |
| 381 | create_resp = await client.post( |
| 382 | "/api/v1/musehub/repos", |
| 383 | json={"name": "qdrant-fail-test", "owner": "testuser", "visibility": "public"}, |
| 384 | headers=auth_headers, |
| 385 | ) |
| 386 | repo_id = create_resp.json()["repoId"] |
| 387 | commit_id = "qdrant-fail-001" |
| 388 | |
| 389 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 390 | await client.post( |
| 391 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 392 | json={ |
| 393 | "branch": "main", |
| 394 | "headCommitId": commit_id, |
| 395 | "commits": [ |
| 396 | { |
| 397 | "commitId": commit_id, |
| 398 | "parentIds": [], |
| 399 | "message": "A minor 95 BPM", |
| 400 | "timestamp": "2024-01-01T00:00:00Z", |
| 401 | } |
| 402 | ], |
| 403 | "objects": [], |
| 404 | "force": False, |
| 405 | }, |
| 406 | headers=auth_headers, |
| 407 | ) |
| 408 | |
| 409 | mock_qdrant = MagicMock() |
| 410 | mock_qdrant.search_similar.side_effect = ConnectionError("Qdrant down") |
| 411 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 412 | try: |
| 413 | resp = await client.get( |
| 414 | f"/api/v1/musehub/search/similar?commit={commit_id}", |
| 415 | headers=auth_headers, |
| 416 | ) |
| 417 | finally: |
| 418 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 419 | |
| 420 | assert resp.status_code == 503 |
| 421 | |
| 422 | |
| 423 | @pytest.mark.anyio |
| 424 | async def test_similar_search_limit_respected( |
| 425 | client: AsyncClient, |
| 426 | auth_headers: dict[str, str], |
| 427 | ) -> None: |
| 428 | """The limit query parameter is forwarded to Qdrant search_similar.""" |
| 429 | create_resp = await client.post( |
| 430 | "/api/v1/musehub/repos", |
| 431 | json={"name": "limit-test", "owner": "testuser", "visibility": "public"}, |
| 432 | headers=auth_headers, |
| 433 | ) |
| 434 | repo_id = create_resp.json()["repoId"] |
| 435 | commit_id = "limit-test-001" |
| 436 | |
| 437 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 438 | await client.post( |
| 439 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 440 | json={ |
| 441 | "branch": "main", |
| 442 | "headCommitId": commit_id, |
| 443 | "commits": [ |
| 444 | { |
| 445 | "commitId": commit_id, |
| 446 | "parentIds": [], |
| 447 | "message": "E major 140 BPM", |
| 448 | "timestamp": "2024-01-01T00:00:00Z", |
| 449 | } |
| 450 | ], |
| 451 | "objects": [], |
| 452 | "force": False, |
| 453 | }, |
| 454 | headers=auth_headers, |
| 455 | ) |
| 456 | |
| 457 | mock_qdrant = MagicMock() |
| 458 | mock_qdrant.search_similar.return_value = [] |
| 459 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 460 | try: |
| 461 | await client.get( |
| 462 | f"/api/v1/musehub/search/similar?commit={commit_id}&limit=3", |
| 463 | headers=auth_headers, |
| 464 | ) |
| 465 | finally: |
| 466 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 467 | |
| 468 | call_kwargs = mock_qdrant.search_similar.call_args.kwargs |
| 469 | assert call_kwargs.get("limit") == 3 |
| 470 | |
| 471 | |
| 472 | # --------------------------------------------------------------------------- |
| 473 | # Similarity search — DI override (regression) |
| 474 | # --------------------------------------------------------------------------- |
| 475 | |
| 476 | |
| 477 | @pytest.mark.anyio |
| 478 | async def test_similar_search_qdrant_injected_via_dependency_override( |
| 479 | client: AsyncClient, |
| 480 | auth_headers: dict[str, str], |
| 481 | ) -> None: |
| 482 | """Qdrant client is injected via FastAPI dependency_overrides, not module patching. |
| 483 | |
| 484 | Regression test: confirms that get_qdrant_client is a proper |
| 485 | FastAPI dependency so tests can override it without patching module internals. |
| 486 | """ |
| 487 | create_resp = await client.post( |
| 488 | "/api/v1/musehub/repos", |
| 489 | json={"name": "di-test-repo", "owner": "testuser", "visibility": "public"}, |
| 490 | headers=auth_headers, |
| 491 | ) |
| 492 | assert create_resp.status_code == 201 |
| 493 | repo_id = create_resp.json()["repoId"] |
| 494 | commit_id = "di-test-commit-001" |
| 495 | |
| 496 | with patch("musehub.api.routes.musehub.sync.embed_push_commits"): |
| 497 | await client.post( |
| 498 | f"/api/v1/musehub/repos/{repo_id}/push", |
| 499 | json={ |
| 500 | "branch": "main", |
| 501 | "headCommitId": commit_id, |
| 502 | "commits": [ |
| 503 | { |
| 504 | "commitId": commit_id, |
| 505 | "parentIds": [], |
| 506 | "message": "DI refactor test commit in Cmaj", |
| 507 | "timestamp": "2024-01-01T00:00:00Z", |
| 508 | } |
| 509 | ], |
| 510 | "objects": [], |
| 511 | "force": False, |
| 512 | }, |
| 513 | headers=auth_headers, |
| 514 | ) |
| 515 | |
| 516 | mock_qdrant = MagicMock() |
| 517 | mock_qdrant.search_similar.return_value = [_make_similar_result("di-result-001", score=0.88)] |
| 518 | app.dependency_overrides[get_qdrant_client] = lambda: mock_qdrant |
| 519 | try: |
| 520 | resp = await client.get( |
| 521 | f"/api/v1/musehub/search/similar?commit={commit_id}&limit=5", |
| 522 | headers=auth_headers, |
| 523 | ) |
| 524 | finally: |
| 525 | app.dependency_overrides.pop(get_qdrant_client, None) |
| 526 | |
| 527 | assert resp.status_code == 200 |
| 528 | data = resp.json() |
| 529 | assert data["queryCommit"] == commit_id |
| 530 | assert len(data["results"]) == 1 |
| 531 | assert data["results"][0]["commitId"] == "di-result-001" |
| 532 | assert mock_qdrant.search_similar.called |
| 533 | |
| 534 | |
| 535 | # --------------------------------------------------------------------------- |
| 536 | # Global search — UI page |
| 537 | # --------------------------------------------------------------------------- |
| 538 | |
| 539 | |
| 540 | @pytest.mark.anyio |
| 541 | async def test_global_search_page_renders( |
| 542 | client: AsyncClient, |
| 543 | db_session: AsyncSession, |
| 544 | ) -> None: |
| 545 | """GET /musehub/ui/search returns 200 HTML with a search form (no auth required).""" |
| 546 | response = await client.get("/musehub/ui/search") |
| 547 | assert response.status_code == 200 |
| 548 | assert "text/html" in response.headers["content-type"] |
| 549 | body = response.text |
| 550 | assert "Global Search" in body |
| 551 | assert "Muse Hub" in body |
| 552 | assert "q-input" in body |
| 553 | assert "mode-sel" in body |
| 554 | |
| 555 | |
| 556 | @pytest.mark.anyio |
| 557 | async def test_global_search_page_pre_fills_query( |
| 558 | client: AsyncClient, |
| 559 | db_session: AsyncSession, |
| 560 | ) -> None: |
| 561 | """GET /musehub/ui/search?q=jazz pre-fills the search form with 'jazz'.""" |
| 562 | response = await client.get("/musehub/ui/search?q=jazz&mode=keyword") |
| 563 | assert response.status_code == 200 |
| 564 | body = response.text |
| 565 | assert "jazz" in body |
| 566 | |
| 567 | |
| 568 | # --------------------------------------------------------------------------- |
| 569 | # Global search — JSON API |
| 570 | # --------------------------------------------------------------------------- |
| 571 | |
| 572 | |
| 573 | @pytest.mark.anyio |
| 574 | async def test_global_search_accessible_without_auth( |
| 575 | client: AsyncClient, |
| 576 | db_session: AsyncSession, |
| 577 | ) -> None: |
| 578 | """GET /api/v1/musehub/search returns 200 without a JWT. |
| 579 | |
| 580 | Global search is a public endpoint — uses optional_token, so unauthenticated |
| 581 | requests are allowed and return results for public repos. |
| 582 | """ |
| 583 | response = await client.get("/api/v1/musehub/search?q=jazz") |
| 584 | assert response.status_code == 200 |
| 585 | |
| 586 | |
| 587 | @pytest.mark.anyio |
| 588 | async def test_global_search_json( |
| 589 | client: AsyncClient, |
| 590 | db_session: AsyncSession, |
| 591 | auth_headers: dict[str, str], |
| 592 | ) -> None: |
| 593 | """GET /api/v1/musehub/search returns JSON with correct content-type.""" |
| 594 | response = await client.get( |
| 595 | "/api/v1/musehub/search?q=jazz", |
| 596 | headers=auth_headers, |
| 597 | ) |
| 598 | assert response.status_code == 200 |
| 599 | assert "application/json" in response.headers["content-type"] |
| 600 | data = response.json() |
| 601 | assert "groups" in data |
| 602 | assert "query" in data |
| 603 | assert data["query"] == "jazz" |
| 604 | |
| 605 | |
| 606 | @pytest.mark.anyio |
| 607 | async def test_global_search_public_only( |
| 608 | client: AsyncClient, |
| 609 | db_session: AsyncSession, |
| 610 | auth_headers: dict[str, str], |
| 611 | ) -> None: |
| 612 | """Private repos must not appear in global search results.""" |
| 613 | public_id = await _make_repo(db_session, name="public-beats", visibility="public") |
| 614 | private_id = await _make_repo(db_session, name="secret-beats", visibility="private") |
| 615 | |
| 616 | await _make_commit( |
| 617 | db_session, public_id, commit_id="pub001abc", message="jazz groove session" |
| 618 | ) |
| 619 | await _make_commit( |
| 620 | db_session, private_id, commit_id="priv001abc", message="jazz private session" |
| 621 | ) |
| 622 | |
| 623 | response = await client.get( |
| 624 | "/api/v1/musehub/search?q=jazz", |
| 625 | headers=auth_headers, |
| 626 | ) |
| 627 | assert response.status_code == 200 |
| 628 | data = response.json() |
| 629 | repo_ids_in_results = {g["repoId"] for g in data["groups"]} |
| 630 | assert public_id in repo_ids_in_results |
| 631 | assert private_id not in repo_ids_in_results |
| 632 | |
| 633 | |
| 634 | @pytest.mark.anyio |
| 635 | async def test_global_search_results_grouped( |
| 636 | client: AsyncClient, |
| 637 | db_session: AsyncSession, |
| 638 | auth_headers: dict[str, str], |
| 639 | ) -> None: |
| 640 | """Results are grouped by repo — each group has repoId, repoName, matches list.""" |
| 641 | repo_a = await _make_repo(db_session, name="repo-alpha", visibility="public") |
| 642 | repo_b = await _make_repo(db_session, name="repo-beta", visibility="public") |
| 643 | |
| 644 | await _make_commit( |
| 645 | db_session, repo_a, commit_id="a001abc123", message="bossa nova rhythm" |
| 646 | ) |
| 647 | await _make_commit( |
| 648 | db_session, repo_a, commit_id="a002abc123", message="bossa nova variation" |
| 649 | ) |
| 650 | await _make_commit( |
| 651 | db_session, repo_b, commit_id="b001abc123", message="bossa nova groove" |
| 652 | ) |
| 653 | |
| 654 | response = await client.get( |
| 655 | "/api/v1/musehub/search?q=bossa+nova", |
| 656 | headers=auth_headers, |
| 657 | ) |
| 658 | assert response.status_code == 200 |
| 659 | data = response.json() |
| 660 | groups = data["groups"] |
| 661 | |
| 662 | group_repo_ids = {g["repoId"] for g in groups} |
| 663 | assert repo_a in group_repo_ids |
| 664 | assert repo_b in group_repo_ids |
| 665 | |
| 666 | for group in groups: |
| 667 | assert "repoId" in group |
| 668 | assert "repoName" in group |
| 669 | assert "repoOwner" in group |
| 670 | assert "repoSlug" in group # PR #282: slug required for UI link construction |
| 671 | assert "repoVisibility" in group |
| 672 | assert "matches" in group |
| 673 | assert "totalMatches" in group |
| 674 | assert isinstance(group["matches"], list) |
| 675 | assert isinstance(group["repoSlug"], str) |
| 676 | assert group["repoSlug"] != "" |
| 677 | |
| 678 | group_a = next(g for g in groups if g["repoId"] == repo_a) |
| 679 | assert group_a["totalMatches"] == 2 |
| 680 | assert len(group_a["matches"]) == 2 |
| 681 | |
| 682 | |
| 683 | @pytest.mark.anyio |
| 684 | async def test_global_search_empty_query_handled( |
| 685 | client: AsyncClient, |
| 686 | db_session: AsyncSession, |
| 687 | auth_headers: dict[str, str], |
| 688 | ) -> None: |
| 689 | """A query that matches nothing returns empty groups and valid pagination metadata.""" |
| 690 | await _make_repo(db_session, name="silent-repo", visibility="public") |
| 691 | |
| 692 | response = await client.get( |
| 693 | "/api/v1/musehub/search?q=zyxqwvutsr_no_match", |
| 694 | headers=auth_headers, |
| 695 | ) |
| 696 | assert response.status_code == 200 |
| 697 | data = response.json() |
| 698 | assert data["groups"] == [] |
| 699 | assert data["page"] == 1 |
| 700 | assert "totalReposSearched" in data |
| 701 | |
| 702 | |
| 703 | @pytest.mark.anyio |
| 704 | async def test_global_search_keyword_mode( |
| 705 | client: AsyncClient, |
| 706 | db_session: AsyncSession, |
| 707 | auth_headers: dict[str, str], |
| 708 | ) -> None: |
| 709 | """Keyword mode matches any term in the query (OR logic, case-insensitive).""" |
| 710 | repo_id = await _make_repo(db_session, name="jazz-lab", visibility="public") |
| 711 | await _make_commit( |
| 712 | db_session, repo_id, commit_id="kw001abcde", message="Blues Shuffle in E" |
| 713 | ) |
| 714 | await _make_commit( |
| 715 | db_session, repo_id, commit_id="kw002abcde", message="Jazz Waltz Trio" |
| 716 | ) |
| 717 | |
| 718 | response = await client.get( |
| 719 | "/api/v1/musehub/search?q=blues&mode=keyword", |
| 720 | headers=auth_headers, |
| 721 | ) |
| 722 | assert response.status_code == 200 |
| 723 | data = response.json() |
| 724 | group = next((g for g in data["groups"] if g["repoId"] == repo_id), None) |
| 725 | assert group is not None |
| 726 | messages = [m["message"] for m in group["matches"]] |
| 727 | assert any("Blues" in msg for msg in messages) |
| 728 | |
| 729 | |
| 730 | @pytest.mark.anyio |
| 731 | async def test_global_search_pattern_mode( |
| 732 | client: AsyncClient, |
| 733 | db_session: AsyncSession, |
| 734 | auth_headers: dict[str, str], |
| 735 | ) -> None: |
| 736 | """Pattern mode applies a raw SQL LIKE pattern to commit messages.""" |
| 737 | repo_id = await _make_repo(db_session, name="pattern-lab", visibility="public") |
| 738 | await _make_commit( |
| 739 | db_session, repo_id, commit_id="pt001abcde", message="minor pentatonic run" |
| 740 | ) |
| 741 | await _make_commit( |
| 742 | db_session, repo_id, commit_id="pt002abcde", message="major scale exercise" |
| 743 | ) |
| 744 | |
| 745 | response = await client.get( |
| 746 | "/api/v1/musehub/search?q=%25minor%25&mode=pattern", |
| 747 | headers=auth_headers, |
| 748 | ) |
| 749 | assert response.status_code == 200 |
| 750 | data = response.json() |
| 751 | group = next((g for g in data["groups"] if g["repoId"] == repo_id), None) |
| 752 | assert group is not None |
| 753 | assert group["totalMatches"] == 1 |
| 754 | assert "minor" in group["matches"][0]["message"] |
| 755 | |
| 756 | |
| 757 | @pytest.mark.anyio |
| 758 | async def test_global_search_pagination( |
| 759 | client: AsyncClient, |
| 760 | db_session: AsyncSession, |
| 761 | auth_headers: dict[str, str], |
| 762 | ) -> None: |
| 763 | """page and page_size parameters control repo-group pagination.""" |
| 764 | ids = [] |
| 765 | for i in range(3): |
| 766 | rid = await _make_repo( |
| 767 | db_session, name=f"paged-repo-{i}", visibility="public", owner=f"owner-{i}" |
| 768 | ) |
| 769 | ids.append(rid) |
| 770 | await _make_commit( |
| 771 | db_session, rid, commit_id=f"pg{i:03d}abcde", message="paginate funk groove" |
| 772 | ) |
| 773 | |
| 774 | response = await client.get( |
| 775 | "/api/v1/musehub/search?q=paginate&page=1&page_size=2", |
| 776 | headers=auth_headers, |
| 777 | ) |
| 778 | assert response.status_code == 200 |
| 779 | data = response.json() |
| 780 | assert len(data["groups"]) <= 2 |
| 781 | assert data["page"] == 1 |
| 782 | assert data["pageSize"] == 2 |
| 783 | |
| 784 | response2 = await client.get( |
| 785 | "/api/v1/musehub/search?q=paginate&page=2&page_size=2", |
| 786 | headers=auth_headers, |
| 787 | ) |
| 788 | assert response2.status_code == 200 |
| 789 | data2 = response2.json() |
| 790 | assert data2["page"] == 2 |
| 791 | |
| 792 | |
| 793 | @pytest.mark.anyio |
| 794 | async def test_global_search_match_contains_required_fields( |
| 795 | client: AsyncClient, |
| 796 | db_session: AsyncSession, |
| 797 | auth_headers: dict[str, str], |
| 798 | ) -> None: |
| 799 | """Each match entry contains commitId, message, author, branch, timestamp, repoId.""" |
| 800 | repo_id = await _make_repo(db_session, name="fields-check", visibility="public") |
| 801 | await _make_commit( |
| 802 | db_session, |
| 803 | repo_id, |
| 804 | commit_id="fc001abcde", |
| 805 | message="swing feel experiment", |
| 806 | author="charlie", |
| 807 | branch="main", |
| 808 | ) |
| 809 | |
| 810 | response = await client.get( |
| 811 | "/api/v1/musehub/search?q=swing", |
| 812 | headers=auth_headers, |
| 813 | ) |
| 814 | assert response.status_code == 200 |
| 815 | data = response.json() |
| 816 | group = next((g for g in data["groups"] if g["repoId"] == repo_id), None) |
| 817 | assert group is not None |
| 818 | match = group["matches"][0] |
| 819 | assert match["commitId"] == "fc001abcde" |
| 820 | assert match["message"] == "swing feel experiment" |
| 821 | assert match["author"] == "charlie" |
| 822 | assert match["branch"] == "main" |
| 823 | assert "timestamp" in match |
| 824 | assert match["repoId"] == repo_id |
| 825 | |
| 826 | |
| 827 | # --------------------------------------------------------------------------- |
| 828 | # Global search — audio preview batching |
| 829 | # --------------------------------------------------------------------------- |
| 830 | |
| 831 | |
| 832 | @pytest.mark.anyio |
| 833 | async def test_global_search_audio_preview_populated_for_multiple_repos( |
| 834 | client: AsyncClient, |
| 835 | db_session: AsyncSession, |
| 836 | auth_headers: dict[str, str], |
| 837 | ) -> None: |
| 838 | """Audio preview object IDs are resolved via a single batched query for all repos. |
| 839 | |
| 840 | Verifies that when N repos all have audio files, each GlobalSearchRepoGroup |
| 841 | contains the correct audioObjectId — confirming the batched path works |
| 842 | end-to-end and produces the same result as the old N+1 per-repo loop. |
| 843 | |
| 844 | Regression test for the N+1 bug fixed. |
| 845 | """ |
| 846 | repo_a = await _make_repo(db_session, name="audio-repo-alpha", visibility="public") |
| 847 | repo_b = await _make_repo(db_session, name="audio-repo-beta", visibility="public") |
| 848 | |
| 849 | await _make_commit( |
| 850 | db_session, repo_a, commit_id="ap001abcde", message="funky groove jam" |
| 851 | ) |
| 852 | await _make_commit( |
| 853 | db_session, repo_b, commit_id="ap002abcde", message="funky bass session" |
| 854 | ) |
| 855 | |
| 856 | obj_a = MusehubObject( |
| 857 | object_id="sha256:audio-preview-alpha", |
| 858 | repo_id=repo_a, |
| 859 | path="preview.mp3", |
| 860 | size_bytes=1024, |
| 861 | disk_path="/tmp/preview-alpha.mp3", |
| 862 | ) |
| 863 | obj_b = MusehubObject( |
| 864 | object_id="sha256:audio-preview-beta", |
| 865 | repo_id=repo_b, |
| 866 | path="preview.ogg", |
| 867 | size_bytes=2048, |
| 868 | disk_path="/tmp/preview-beta.ogg", |
| 869 | ) |
| 870 | db_session.add(obj_a) |
| 871 | db_session.add(obj_b) |
| 872 | await db_session.commit() |
| 873 | |
| 874 | response = await client.get( |
| 875 | "/api/v1/musehub/search?q=funky", |
| 876 | headers=auth_headers, |
| 877 | ) |
| 878 | assert response.status_code == 200 |
| 879 | data = response.json() |
| 880 | |
| 881 | groups_by_id = {g["repoId"]: g for g in data["groups"]} |
| 882 | assert repo_a in groups_by_id |
| 883 | assert repo_b in groups_by_id |
| 884 | |
| 885 | assert groups_by_id[repo_a]["matches"][0]["audioObjectId"] == "sha256:audio-preview-alpha" |
| 886 | assert groups_by_id[repo_b]["matches"][0]["audioObjectId"] == "sha256:audio-preview-beta" |
| 887 | |
| 888 | |
| 889 | @pytest.mark.anyio |
| 890 | async def test_global_search_audio_preview_absent_when_no_audio_objects( |
| 891 | client: AsyncClient, |
| 892 | db_session: AsyncSession, |
| 893 | auth_headers: dict[str, str], |
| 894 | ) -> None: |
| 895 | """Repos without audio objects return null audioObjectId in search results.""" |
| 896 | repo_id = await _make_repo(db_session, name="no-audio-repo", visibility="public") |
| 897 | await _make_commit( |
| 898 | db_session, repo_id, commit_id="na001abcde", message="silent ambient piece" |
| 899 | ) |
| 900 | |
| 901 | response = await client.get( |
| 902 | "/api/v1/musehub/search?q=silent", |
| 903 | headers=auth_headers, |
| 904 | ) |
| 905 | assert response.status_code == 200 |
| 906 | data = response.json() |
| 907 | group = next((g for g in data["groups"] if g["repoId"] == repo_id), None) |
| 908 | assert group is not None |
| 909 | assert group["matches"][0]["audioObjectId"] is None |
| 910 | |
| 911 | |
| 912 | # --------------------------------------------------------------------------- |
| 913 | # In-repo search — UI page |
| 914 | # --------------------------------------------------------------------------- |
| 915 | |
| 916 | |
| 917 | @pytest.mark.anyio |
| 918 | async def test_search_page_renders( |
| 919 | client: AsyncClient, |
| 920 | db_session: AsyncSession, |
| 921 | ) -> None: |
| 922 | """GET /musehub/ui/{repo_id}/search returns 200 HTML with mode tabs.""" |
| 923 | repo_id = await _make_search_repo(db_session) |
| 924 | response = await client.get("/musehub/ui/testuser/search-test-repo/search") |
| 925 | assert response.status_code == 200 |
| 926 | assert "text/html" in response.headers["content-type"] |
| 927 | body = response.text |
| 928 | assert "Muse Hub" in body |
| 929 | assert "Search Commits" in body |
| 930 | assert "Keyword" in body |
| 931 | assert "Natural Language" in body |
| 932 | assert "Pattern" in body |
| 933 | assert "Musical Properties" in body |
| 934 | assert "inp-since" in body |
| 935 | assert "inp-until" in body |
| 936 | |
| 937 | |
| 938 | @pytest.mark.anyio |
| 939 | async def test_search_page_no_auth_required( |
| 940 | client: AsyncClient, |
| 941 | db_session: AsyncSession, |
| 942 | ) -> None: |
| 943 | """Search UI page is accessible without a JWT (HTML shell, JS handles auth).""" |
| 944 | repo_id = await _make_search_repo(db_session) |
| 945 | response = await client.get("/musehub/ui/testuser/search-test-repo/search") |
| 946 | assert response.status_code == 200 |
| 947 | |
| 948 | |
| 949 | # --------------------------------------------------------------------------- |
| 950 | # In-repo search — authentication |
| 951 | # --------------------------------------------------------------------------- |
| 952 | |
| 953 | |
| 954 | @pytest.mark.anyio |
| 955 | async def test_search_requires_auth( |
| 956 | client: AsyncClient, |
| 957 | db_session: AsyncSession, |
| 958 | ) -> None: |
| 959 | """GET /api/v1/musehub/repos/{repo_id}/search returns 401 without a token.""" |
| 960 | repo_id = await _make_search_repo(db_session) |
| 961 | response = await client.get(f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=jazz") |
| 962 | assert response.status_code == 401 |
| 963 | |
| 964 | |
| 965 | @pytest.mark.anyio |
| 966 | async def test_search_unknown_repo( |
| 967 | client: AsyncClient, |
| 968 | db_session: AsyncSession, |
| 969 | auth_headers: dict[str, str], |
| 970 | ) -> None: |
| 971 | """GET /api/v1/musehub/repos/{unknown}/search returns 404.""" |
| 972 | response = await client.get( |
| 973 | "/api/v1/musehub/repos/does-not-exist/search?mode=keyword&q=test", |
| 974 | headers=auth_headers, |
| 975 | ) |
| 976 | assert response.status_code == 404 |
| 977 | |
| 978 | |
| 979 | @pytest.mark.anyio |
| 980 | async def test_search_invalid_mode( |
| 981 | client: AsyncClient, |
| 982 | db_session: AsyncSession, |
| 983 | auth_headers: dict[str, str], |
| 984 | ) -> None: |
| 985 | """GET search with an unknown mode returns 422.""" |
| 986 | repo_id = await _make_search_repo(db_session) |
| 987 | response = await client.get( |
| 988 | f"/api/v1/musehub/repos/{repo_id}/search?mode=badmode&q=x", |
| 989 | headers=auth_headers, |
| 990 | ) |
| 991 | assert response.status_code == 422 |
| 992 | |
| 993 | |
| 994 | # --------------------------------------------------------------------------- |
| 995 | # In-repo search — keyword mode |
| 996 | # --------------------------------------------------------------------------- |
| 997 | |
| 998 | |
| 999 | @pytest.mark.anyio |
| 1000 | async def test_search_keyword_mode( |
| 1001 | client: AsyncClient, |
| 1002 | db_session: AsyncSession, |
| 1003 | auth_headers: dict[str, str], |
| 1004 | ) -> None: |
| 1005 | """Keyword search returns commits whose messages overlap with the query.""" |
| 1006 | repo_id = await _make_search_repo(db_session) |
| 1007 | await db_session.commit() |
| 1008 | |
| 1009 | await _make_search_commit(db_session, repo_id=repo_id, message="dark jazz bassline in Dm") |
| 1010 | await _make_search_commit(db_session, repo_id=repo_id, message="classical piano intro section") |
| 1011 | await _make_search_commit(db_session, repo_id=repo_id, message="hip hop drum fill pattern") |
| 1012 | await db_session.commit() |
| 1013 | |
| 1014 | response = await client.get( |
| 1015 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=jazz+bassline", |
| 1016 | headers=auth_headers, |
| 1017 | ) |
| 1018 | assert response.status_code == 200 |
| 1019 | data = response.json() |
| 1020 | assert data["mode"] == "keyword" |
| 1021 | assert data["query"] == "jazz bassline" |
| 1022 | assert any("jazz" in m["message"].lower() for m in data["matches"]) |
| 1023 | |
| 1024 | |
| 1025 | @pytest.mark.anyio |
| 1026 | async def test_search_keyword_empty_query( |
| 1027 | client: AsyncClient, |
| 1028 | db_session: AsyncSession, |
| 1029 | auth_headers: dict[str, str], |
| 1030 | ) -> None: |
| 1031 | """Empty keyword query returns empty matches (no tokens → no overlap).""" |
| 1032 | repo_id = await _make_search_repo(db_session) |
| 1033 | await db_session.commit() |
| 1034 | await _make_search_commit(db_session, repo_id=repo_id, message="some commit") |
| 1035 | await db_session.commit() |
| 1036 | |
| 1037 | response = await client.get( |
| 1038 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=", |
| 1039 | headers=auth_headers, |
| 1040 | ) |
| 1041 | assert response.status_code == 200 |
| 1042 | data = response.json() |
| 1043 | assert data["mode"] == "keyword" |
| 1044 | assert data["matches"] == [] |
| 1045 | |
| 1046 | |
| 1047 | @pytest.mark.anyio |
| 1048 | async def test_search_json_response( |
| 1049 | client: AsyncClient, |
| 1050 | db_session: AsyncSession, |
| 1051 | auth_headers: dict[str, str], |
| 1052 | ) -> None: |
| 1053 | """Search response has the expected SearchResponse JSON shape.""" |
| 1054 | repo_id = await _make_search_repo(db_session) |
| 1055 | await db_session.commit() |
| 1056 | await _make_search_commit(db_session, repo_id=repo_id, message="piano chord progression F Bb Eb") |
| 1057 | await db_session.commit() |
| 1058 | |
| 1059 | response = await client.get( |
| 1060 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=piano", |
| 1061 | headers=auth_headers, |
| 1062 | ) |
| 1063 | assert response.status_code == 200 |
| 1064 | data = response.json() |
| 1065 | |
| 1066 | assert "mode" in data |
| 1067 | assert "query" in data |
| 1068 | assert "matches" in data |
| 1069 | assert "totalScanned" in data |
| 1070 | assert "limit" in data |
| 1071 | |
| 1072 | if data["matches"]: |
| 1073 | m = data["matches"][0] |
| 1074 | assert "commitId" in m |
| 1075 | assert "branch" in m |
| 1076 | assert "message" in m |
| 1077 | assert "author" in m |
| 1078 | assert "timestamp" in m |
| 1079 | assert "score" in m |
| 1080 | assert "matchSource" in m |
| 1081 | |
| 1082 | |
| 1083 | # --------------------------------------------------------------------------- |
| 1084 | # In-repo search — musical property mode |
| 1085 | # --------------------------------------------------------------------------- |
| 1086 | |
| 1087 | |
| 1088 | @pytest.mark.anyio |
| 1089 | async def test_search_musical_property( |
| 1090 | client: AsyncClient, |
| 1091 | db_session: AsyncSession, |
| 1092 | auth_headers: dict[str, str], |
| 1093 | ) -> None: |
| 1094 | """Property mode filters commits containing the harmony string.""" |
| 1095 | repo_id = await _make_search_repo(db_session) |
| 1096 | await db_session.commit() |
| 1097 | |
| 1098 | await _make_search_commit(db_session, repo_id=repo_id, message="add harmony=Eb bridge section") |
| 1099 | await _make_search_commit(db_session, repo_id=repo_id, message="drum groove tweak no harmony") |
| 1100 | await db_session.commit() |
| 1101 | |
| 1102 | response = await client.get( |
| 1103 | f"/api/v1/musehub/repos/{repo_id}/search?mode=property&harmony=Eb", |
| 1104 | headers=auth_headers, |
| 1105 | ) |
| 1106 | assert response.status_code == 200 |
| 1107 | data = response.json() |
| 1108 | assert data["mode"] == "property" |
| 1109 | assert len(data["matches"]) >= 1 |
| 1110 | assert all("Eb" in m["message"] for m in data["matches"]) |
| 1111 | |
| 1112 | |
| 1113 | # --------------------------------------------------------------------------- |
| 1114 | # In-repo search — natural language (ask) mode |
| 1115 | # --------------------------------------------------------------------------- |
| 1116 | |
| 1117 | |
| 1118 | @pytest.mark.anyio |
| 1119 | async def test_search_natural_language( |
| 1120 | client: AsyncClient, |
| 1121 | db_session: AsyncSession, |
| 1122 | auth_headers: dict[str, str], |
| 1123 | ) -> None: |
| 1124 | """Ask mode extracts keywords and returns relevant commits.""" |
| 1125 | repo_id = await _make_search_repo(db_session) |
| 1126 | await db_session.commit() |
| 1127 | |
| 1128 | await _make_search_commit(db_session, repo_id=repo_id, message="switched tempo to 140bpm for drop") |
| 1129 | await _make_search_commit(db_session, repo_id=repo_id, message="piano melody in minor key") |
| 1130 | await db_session.commit() |
| 1131 | |
| 1132 | response = await client.get( |
| 1133 | f"/api/v1/musehub/repos/{repo_id}/search?mode=ask&q=what+tempo+changes+did+I+make", |
| 1134 | headers=auth_headers, |
| 1135 | ) |
| 1136 | assert response.status_code == 200 |
| 1137 | data = response.json() |
| 1138 | assert data["mode"] == "ask" |
| 1139 | assert any("tempo" in m["message"].lower() for m in data["matches"]) |
| 1140 | |
| 1141 | |
| 1142 | # --------------------------------------------------------------------------- |
| 1143 | # In-repo search — pattern mode |
| 1144 | # --------------------------------------------------------------------------- |
| 1145 | |
| 1146 | |
| 1147 | @pytest.mark.anyio |
| 1148 | async def test_search_pattern_message( |
| 1149 | client: AsyncClient, |
| 1150 | db_session: AsyncSession, |
| 1151 | auth_headers: dict[str, str], |
| 1152 | ) -> None: |
| 1153 | """Pattern mode matches substring in commit message.""" |
| 1154 | repo_id = await _make_search_repo(db_session) |
| 1155 | await db_session.commit() |
| 1156 | |
| 1157 | await _make_search_commit(db_session, repo_id=repo_id, message="add Cm7 chord voicing in bridge") |
| 1158 | await _make_search_commit(db_session, repo_id=repo_id, message="fix timing on verse drums") |
| 1159 | await db_session.commit() |
| 1160 | |
| 1161 | response = await client.get( |
| 1162 | f"/api/v1/musehub/repos/{repo_id}/search?mode=pattern&q=Cm7", |
| 1163 | headers=auth_headers, |
| 1164 | ) |
| 1165 | assert response.status_code == 200 |
| 1166 | data = response.json() |
| 1167 | assert data["mode"] == "pattern" |
| 1168 | assert len(data["matches"]) == 1 |
| 1169 | assert "Cm7" in data["matches"][0]["message"] |
| 1170 | assert data["matches"][0]["matchSource"] == "message" |
| 1171 | |
| 1172 | |
| 1173 | @pytest.mark.anyio |
| 1174 | async def test_search_pattern_branch( |
| 1175 | client: AsyncClient, |
| 1176 | db_session: AsyncSession, |
| 1177 | auth_headers: dict[str, str], |
| 1178 | ) -> None: |
| 1179 | """Pattern mode matches substring in branch name when message doesn't match.""" |
| 1180 | repo_id = await _make_search_repo(db_session) |
| 1181 | await db_session.commit() |
| 1182 | |
| 1183 | await _make_search_commit( |
| 1184 | db_session, |
| 1185 | repo_id=repo_id, |
| 1186 | message="rough cut", |
| 1187 | branch="feature/hip-hop-session", |
| 1188 | ) |
| 1189 | await db_session.commit() |
| 1190 | |
| 1191 | response = await client.get( |
| 1192 | f"/api/v1/musehub/repos/{repo_id}/search?mode=pattern&q=hip-hop", |
| 1193 | headers=auth_headers, |
| 1194 | ) |
| 1195 | assert response.status_code == 200 |
| 1196 | data = response.json() |
| 1197 | assert data["mode"] == "pattern" |
| 1198 | assert len(data["matches"]) == 1 |
| 1199 | assert data["matches"][0]["matchSource"] == "branch" |
| 1200 | |
| 1201 | |
| 1202 | # --------------------------------------------------------------------------- |
| 1203 | # In-repo search — date range filters |
| 1204 | # --------------------------------------------------------------------------- |
| 1205 | |
| 1206 | |
| 1207 | @pytest.mark.anyio |
| 1208 | async def test_search_date_range_since( |
| 1209 | client: AsyncClient, |
| 1210 | db_session: AsyncSession, |
| 1211 | auth_headers: dict[str, str], |
| 1212 | ) -> None: |
| 1213 | """since filter excludes commits committed before the given datetime.""" |
| 1214 | repo_id = await _make_search_repo(db_session) |
| 1215 | await db_session.commit() |
| 1216 | |
| 1217 | old_ts = datetime(2024, 1, 1, tzinfo=timezone.utc) |
| 1218 | new_ts = datetime(2026, 1, 1, tzinfo=timezone.utc) |
| 1219 | |
| 1220 | await _make_search_commit(db_session, repo_id=repo_id, message="old jazz commit", committed_at=old_ts) |
| 1221 | await _make_search_commit(db_session, repo_id=repo_id, message="new jazz commit", committed_at=new_ts) |
| 1222 | await db_session.commit() |
| 1223 | |
| 1224 | response = await client.get( |
| 1225 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=jazz&since=2025-06-01T00:00:00Z", |
| 1226 | headers=auth_headers, |
| 1227 | ) |
| 1228 | assert response.status_code == 200 |
| 1229 | data = response.json() |
| 1230 | assert all(m["message"] != "old jazz commit" for m in data["matches"]) |
| 1231 | assert any(m["message"] == "new jazz commit" for m in data["matches"]) |
| 1232 | |
| 1233 | |
| 1234 | @pytest.mark.anyio |
| 1235 | async def test_search_date_range_until( |
| 1236 | client: AsyncClient, |
| 1237 | db_session: AsyncSession, |
| 1238 | auth_headers: dict[str, str], |
| 1239 | ) -> None: |
| 1240 | """until filter excludes commits committed after the given datetime.""" |
| 1241 | repo_id = await _make_search_repo(db_session) |
| 1242 | await db_session.commit() |
| 1243 | |
| 1244 | old_ts = datetime(2024, 1, 1, tzinfo=timezone.utc) |
| 1245 | new_ts = datetime(2026, 1, 1, tzinfo=timezone.utc) |
| 1246 | |
| 1247 | await _make_search_commit(db_session, repo_id=repo_id, message="old piano commit", committed_at=old_ts) |
| 1248 | await _make_search_commit(db_session, repo_id=repo_id, message="new piano commit", committed_at=new_ts) |
| 1249 | await db_session.commit() |
| 1250 | |
| 1251 | response = await client.get( |
| 1252 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=piano&until=2025-06-01T00:00:00Z", |
| 1253 | headers=auth_headers, |
| 1254 | ) |
| 1255 | assert response.status_code == 200 |
| 1256 | data = response.json() |
| 1257 | assert any(m["message"] == "old piano commit" for m in data["matches"]) |
| 1258 | assert all(m["message"] != "new piano commit" for m in data["matches"]) |
| 1259 | |
| 1260 | |
| 1261 | # --------------------------------------------------------------------------- |
| 1262 | # In-repo search — limit |
| 1263 | # --------------------------------------------------------------------------- |
| 1264 | |
| 1265 | |
| 1266 | @pytest.mark.anyio |
| 1267 | async def test_search_limit_respected( |
| 1268 | client: AsyncClient, |
| 1269 | db_session: AsyncSession, |
| 1270 | auth_headers: dict[str, str], |
| 1271 | ) -> None: |
| 1272 | """The limit parameter caps the number of results returned.""" |
| 1273 | repo_id = await _make_search_repo(db_session) |
| 1274 | await db_session.commit() |
| 1275 | |
| 1276 | for i in range(10): |
| 1277 | await _make_search_commit(db_session, repo_id=repo_id, message=f"bass groove iteration {i}") |
| 1278 | await db_session.commit() |
| 1279 | |
| 1280 | response = await client.get( |
| 1281 | f"/api/v1/musehub/repos/{repo_id}/search?mode=keyword&q=bass&limit=3", |
| 1282 | headers=auth_headers, |
| 1283 | ) |
| 1284 | assert response.status_code == 200 |
| 1285 | data = response.json() |
| 1286 | assert len(data["matches"]) <= 3 |
| 1287 | assert data["limit"] == 3 |