""" test_phase3.py — 測試 tagger.py, indexer.py, run.py """ import json import os import sys import tempfile from datetime import datetime, timezone from pathlib import Path # 加入 src 到路徑 sys.path.insert(0, str(Path(__file__).parent.parent)) from src.capturers.base import RawNote from src.organizer.tagger import Tagger, KeywordTagger, TagSuggestion from src.organizer.indexer import NoteIndexer from src.utils.config import ConfigLoader def make_temp_data_dir(vocabulary: dict = None, config: dict = None) -> Path: """建立暫時的 _data 目錄，帶 config/vocabulary""" import yaml tmp = Path(tempfile.mkdtemp()) default_vocab = { "技術": { "AI/ML": {"keywords": ["llm", "gpt", "pytorch"]}, "程式語言": {"keywords": ["python", "javascript"]}, } } default_config = { "vault": {"root": str(tmp / "vault"), "output_dir": "0.raw"}, "sources": {}, } vocab_data = vocabulary or default_vocab cfg_data = config or default_config (tmp).mkdir(parents=True, exist_ok=True) with open(tmp / "vocabulary.yaml", "w") as f: yaml.dump(vocab_data, f, allow_unicode=True) with open(tmp / "config.yaml", "w") as f: yaml.dump(cfg_data, f, allow_unicode=True) return tmp def make_note(content: str, tags_raw=None, source="test") -> RawNote: return RawNote( source=source, source_id="abc12345", content=content, created_at="2026-04-30T10:00:00+00:00", tags_raw=tags_raw or [], ) # ── Test 1: KeywordTagger — 命中關鍵字 ── def test_keyword_tagger_hit(): data_dir = make_temp_data_dir() config_loader = ConfigLoader(data_dir=data_dir) tagger = KeywordTagger(config_loader) note = make_note("今天研究 LLM 和 GPT-4 的 fine-tuning 方法") suggestions = tagger.tag(note) tags = [s.tag for s in suggestions] assert any("AI/ML" in t or "llm" in t.lower() for t in tags), f"Expected AI tag, got {tags}" print(f" ✅ KeywordTagger hit: {tags}") # ── Test 2: KeywordTagger — 無命中 ── def test_keyword_tagger_miss(): data_dir = make_temp_data_dir() config_loader = ConfigLoader(data_dir=data_dir) tagger = KeywordTagger(config_loader) note = make_note("今天天氣很好，去公園散步") suggestions = tagger.tag(note) # 不應有高 confidence 的技術標籤 high_conf = [s for s in suggestions if s.confidence > 0.5 and s.method == "keyword"] assert len(high_conf) == 0, f"Unexpected high-conf tags: {high_conf}" print(f" ✅ KeywordTagger no hit for non-tech content") # ── Test 3: Tagger.top_tags confidence 過濾 ── def test_tagger_top_tags(): data_dir = make_temp_data_dir() config_loader = ConfigLoader(data_dir=data_dir) tagger = Tagger(config_loader) note = make_note("python 和 javascript 開發技巧", tags_raw=["#programming"]) suggestions = tagger.tag(note, use_llm=False) top = tagger.top_tags(suggestions, max_tags=5, min_confidence=0.3) assert len(top) >= 1, f"Expected at least 1 tag, got {top}" print(f" ✅ Tagger top_tags: {top}") # ── Test 4: existing tags 優先（高 confidence）── def test_existing_tags_high_confidence(): data_dir = make_temp_data_dir() config_loader = ConfigLoader(data_dir=data_dir) tagger = KeywordTagger(config_loader) note = make_note("隨便寫的內容", tags_raw=["#myProject", "#2026"]) suggestions = tagger.tag(note) existing = [s for s in suggestions if s.method == "existing"] assert len(existing) == 2, f"Expected 2 existing tags, got {existing}" assert all(s.confidence >= 0.9 for s in existing), "Existing tags should have high confidence" print(f" ✅ Existing tags confidence: {[s.confidence for s in existing]}") # ── Test 5: NoteIndexer — 寫入 vault ── def test_indexer_write(): data_dir = make_temp_data_dir() vault_root = data_dir / "vault" config_loader = ConfigLoader(data_dir=data_dir) indexer = NoteIndexer(config_loader, vault_root=vault_root) note = make_note("# 測試筆記\n\n今天學了 python 的 dataclass 用法") tags = ["技術/程式語言", "學習"] note_id = "test-20260430-abc12345" filepath = indexer.index_note(note, tags, note_id) assert filepath is not None, "Should have written a file" assert filepath.exists(), f"File not found: {filepath}" content = filepath.read_text(encoding="utf-8") assert "note_id: test-20260430-abc12345" in content assert "技術/程式語言" in content assert "測試筆記" in content print(f" ✅ Indexer wrote to {filepath.name}") # ── Test 6: NoteIndexer — 重複寫入不覆寫 ── def test_indexer_no_overwrite(): data_dir = make_temp_data_dir() vault_root = data_dir / "vault" config_loader = ConfigLoader(data_dir=data_dir) indexer = NoteIndexer(config_loader, vault_root=vault_root) note = make_note("# 原始內容") note_id = "test-20260430-abc12345" p1 = indexer.index_note(note, [], note_id) p2 = indexer.index_note(note, [], note_id, overwrite=False) assert p1 is not None assert p2 is None, "Second write without overwrite should return None" print(f" ✅ Indexer no-overwrite works") # ── Test 7: NoteIndexer — content hash 計算 ── def test_indexer_content_hash(): data_dir = make_temp_data_dir() config_loader = ConfigLoader(data_dir=data_dir) indexer = NoteIndexer(config_loader, vault_root=data_dir / "vault") note = make_note("hello world") h1 = indexer.get_note_content_hash(note, ["tag1"]) h2 = indexer.get_note_content_hash(note, ["tag1"]) h3 = indexer.get_note_content_hash(note, ["tag2"]) assert h1 == h2 assert h1 != h3 print(f" ✅ Content hash stable: {h1}") if __name__ == "__main__": tests = [ test_keyword_tagger_hit, test_keyword_tagger_miss, test_tagger_top_tags, test_existing_tags_high_confidence, test_indexer_write, test_indexer_no_overwrite, test_indexer_content_hash, ] passed = 0 failed = 0 for test in tests: try: print(f"\n🧪 {test.__name__}") test() passed += 1 except Exception as e: import traceback print(f" ❌ FAIL: {e}") traceback.print_exc() failed += 1 print(f"\n{'='*40}") print(f"結果：{passed} passed, {failed} failed") sys.exit(0 if failed == 0 else 1)