"""Tests fuer den rekursiven Audio-Scanner.""" from __future__ import annotations from pathlib import Path import pytest from sjaudiosearch.core.metadata import AudioMetadataError, read_metadata from sjaudiosearch.core.scanner import scan from conftest import set_mtime def test_scan_finds_multiple_audio_files(repo, make_wav, tmp_path, processing): music = tmp_path / "music" make_wav("music/a.mp3.wav", 1.0) make_wav("music/sub/b.wav", 1.0) make_wav("music/sub/deep/c.wav", 1.0) (music / "notes.txt").write_text("kein Audio", encoding="utf-8") result = scan(music, repo, processing) assert result.files_found == 3 assert result.added == 3 assert result.changed == 0 assert result.errors == [] records = repo.list_audio() assert len(records) == 3 assert {record.filename for record in records} == {"a.mp3.wav", "b.wav", "c.wav"} for record in records: assert record.status == "pending" assert Path(record.path).is_absolute() assert record.duration == pytest.approx(1.0, abs=0.05) assert record.file_size > 0 def test_scan_detects_duplicate_content(repo, make_wav, tmp_path, processing): root = tmp_path / "music" original = make_wav("music/original.wav", 2.0) copy = root / "kopie.wav" copy.parent.mkdir(parents=True, exist_ok=True) copy.write_bytes(original.read_bytes()) result = scan(root, repo, processing) assert result.added == 2 assert result.duplicates == 1 records = {record.filename: record for record in repo.list_audio()} first, second = records["kopie.wav"], records["original.wav"] # Der Scan ist alphabetisch sortiert, daher ist "kopie.wav" zuerst dran und # "original.wav" wird als Duplikat markiert. Genau einer der beiden muss # auf den anderen zeigen. assert bool(first.duplicate_of) ^ bool(second.duplicate_of) duplicate, keeper = (first, second) if first.duplicate_of else (second, first) assert duplicate.duplicate_of == keeper.id assert duplicate.content_hash == keeper.content_hash # Duplikate werden nicht erneut transkribiert. queue = [record.filename for record in repo.transcription_queue()] assert queue == [keeper.filename] def test_scan_detects_changed_file(repo, make_wav, tmp_path, processing): root = tmp_path / "music" root.mkdir() first = scan(root, repo, processing) assert first.added == 0 path = make_wav("music/folge1.wav", 1.0) assert scan(root, repo, processing).added == 1 record = repo.get_audio_by_path(str(path.resolve())) assert record is not None and record.status == "pending" # Inhalt aendern (andere Laenge) und Aenderungszeit setzen make_wav("music/folge1.wav", 3.0) set_mtime(path, path.stat().st_mtime + 30) result = scan(root, repo, processing) assert result.added == 0 assert result.changed == 1 assert result.unchanged == 0 updated = repo.get_audio(record.id) assert updated is not None assert updated.status == "pending" assert updated.duration == pytest.approx(3.0, abs=0.05) def test_scan_keeps_unchanged_files(repo, make_wav, tmp_path, processing): root = tmp_path / "music" make_wav("music/a.wav", 1.0) make_wav("music/b.wav", 1.0) first = scan(root, repo, processing) assert first.added == 2 second = scan(root, repo, processing) assert second.added == 0 assert second.changed == 0 assert second.unchanged == 2 def test_scan_logs_corrupt_files(repo, tmp_path, processing): root = tmp_path / "music" root.mkdir() broken = root / "kaputt.mp3" broken.write_bytes(b"das ist kein audio") result = scan(root, repo, processing) assert result.files_found == 1 assert result.added == 1 assert len(result.errors) == 1 assert "kaputt.mp3" in result.errors[0].path record = repo.get_audio_by_path(str(broken.resolve())) assert record is not None assert record.status == "error" assert record.error_message def test_scan_rejects_missing_path(repo, tmp_path, processing): with pytest.raises(FileNotFoundError): scan(tmp_path / "gibt-es-nicht", repo, processing) def test_metadata_requires_real_audio(tmp_path): fake = tmp_path / "fake.flac" fake.write_bytes(b"nope") with pytest.raises(AudioMetadataError): read_metadata(fake)