""" test_transcript_learner.py scripts/transcript-learner.py 단위 테스트 (TDD) 테스트 항목: 1. 지적 패턴(criticism) 감지 2. 칭찬 패턴(praise) 감지 3. 선호 패턴(preference) 감지 4. 중복 방지 (같은 텍스트 2번 → count 증가) 5. 빈 입력 graceful 처리 6. learnings.json 없는 경우 (신규 생성) 7. learnings.json 있는 경우 (누적) 8. stdin 읽기 테스트 (monkeypatch) 9. 여러 키워드가 한 줄에 있는 경우 10. stdout 출력 형식 검증 """ import importlib.util import json import sys from pathlib import Path import pytest # scripts 디렉토리를 import path에 추가 _SCRIPTS_DIR = Path(__file__).parent.parent sys.path.insert(0, str(_SCRIPTS_DIR)) # transcript-learner.py는 하이픈이 있으므로 importlib으로 임포트 _MODULE_PATH = _SCRIPTS_DIR / "transcript-learner.py" spec = importlib.util.spec_from_file_location("transcript_learner", _MODULE_PATH) assert spec is not None transcript_learner = importlib.util.module_from_spec(spec) assert spec.loader is not None spec.loader.exec_module(transcript_learner) # --------------------------------------------------------------------------- # Helper # --------------------------------------------------------------------------- def make_learnings(tmp_path: Path, data: dict) -> Path: whisper_dir = tmp_path / "whisper" whisper_dir.mkdir(parents=True, exist_ok=True) f = whisper_dir / "learnings.json" f.write_text(json.dumps(data, ensure_ascii=False), encoding="utf-8") return tmp_path def get_learnings_path(tmp_path: Path) -> Path: return tmp_path / "whisper" / "learnings.json" # --------------------------------------------------------------------------- # 1. 지적 패턴 감지 테스트 # --------------------------------------------------------------------------- class TestCriticismDetection: """지적 패턴 키워드 감지""" def test_detect_igeo_wae(self): line = "이거 왜 이렇게 느려?" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) assert any(p["keyword"] == "이거 왜" for p in result) def test_detect_haetjana(self): line = "분명히 했잖아 그거를" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_ani_geuge(self): line = "아니 그게 맞냐고" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_dasi(self): line = "다시 해줘 이거" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_akka_malhaetjana(self): line = "아까 말했잖아 그 부분을" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_imi_haetjana(self): line = "이미 했잖아 이거" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_tto(self): line = "또? 왜 또 그래" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) def test_detect_anirago(self): line = "아니라고 했잖아" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "criticism" for p in result) # --------------------------------------------------------------------------- # 2. 칭찬 패턴 감지 테스트 # --------------------------------------------------------------------------- class TestPraiseDetection: """칭찬 패턴 키워드 감지""" def test_detect_joha(self): line = "오 좋아 이렇게 해줘" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_o_gwaenchanheunde(self): line = "오 괜찮은데 이거" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_igeo_majeo(self): line = "이거 맞어 이렇게 하면 돼" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_jalhaesseo(self): line = "잘했어 이번엔" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_johne(self): line = "결과 좋네 마음에 들어" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_gwaenchanta(self): line = "괜찮다 이 방식" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) def test_detect_geuraee_igeoya(self): line = "그래 이거야 딱 원하던 거" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "praise" for p in result) # --------------------------------------------------------------------------- # 3. 선호 패턴 감지 테스트 # --------------------------------------------------------------------------- class TestPreferenceDetection: """선호 패턴 키워드 감지""" def test_detect_hangsang(self): line = "항상 이렇게 해줘" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "preference" for p in result) def test_detect_haji_ma(self): line = "그렇게 하지 마" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "preference" for p in result) def test_detect_euro_hae(self): line = "영어로 해줘" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "preference" for p in result) def test_detect_nan_ga_joha(self): line = "난 이 방식이 좋아" result = transcript_learner.extract_patterns_from_line(line) assert any(p["type"] == "preference" for p in result) # --------------------------------------------------------------------------- # 4. 매칭 없는 줄 테스트 # --------------------------------------------------------------------------- class TestNoMatch: """패턴 없는 줄""" def test_no_match_plain_text(self): line = "오늘 날씨가 맑습니다" result = transcript_learner.extract_patterns_from_line(line) assert result == [] def test_no_match_empty_line(self): line = "" result = transcript_learner.extract_patterns_from_line(line) assert result == [] def test_no_match_whitespace_only(self): line = " " result = transcript_learner.extract_patterns_from_line(line) assert result == [] # --------------------------------------------------------------------------- # 5. 여러 키워드가 한 줄에 있는 경우 # --------------------------------------------------------------------------- class TestMultipleKeywordsOneLine: """한 줄에 여러 키워드""" def test_criticism_and_praise_same_line(self): line = "이거 왜 이러지 그래도 좋아" result = transcript_learner.extract_patterns_from_line(line) types = [p["type"] for p in result] assert "criticism" in types assert "praise" in types def test_multiple_criticism_keywords(self): line = "이거 왜 다시 했잖아" result = transcript_learner.extract_patterns_from_line(line) criticism_results = [p for p in result if p["type"] == "criticism"] assert len(criticism_results) >= 2 def test_context_is_full_line(self): line = "이거 왜 이렇게 했어?" result = transcript_learner.extract_patterns_from_line(line) assert all(p["context"] == line for p in result) # --------------------------------------------------------------------------- # 6. learnings.json 없는 경우 (신규 생성) # --------------------------------------------------------------------------- class TestNewLearningsFile: """learnings.json 신규 생성""" def test_load_learnings_missing_returns_empty(self, tmp_path: Path): result = transcript_learner.load_learnings(base_dir=tmp_path) assert result["version"] == 1 assert result["patterns"] == [] def test_process_transcript_creates_file(self, tmp_path: Path): text = "이거 왜 이렇게 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) path = get_learnings_path(tmp_path) assert path.exists() def test_process_transcript_new_pattern_saved(self, tmp_path: Path): text = "이거 왜 이렇게 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) assert len(data["patterns"]) >= 1 assert any(p["keyword"] == "이거 왜" for p in data["patterns"]) def test_process_transcript_version_1(self, tmp_path: Path): text = "좋아 이 방식이 마음에 들어" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) assert data["version"] == 1 def test_process_transcript_has_last_updated(self, tmp_path: Path): text = "잘했어" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) assert "last_updated" in data assert data["last_updated"] != "" def test_pattern_has_required_fields(self, tmp_path: Path): text = "이거 왜 이렇게 됐어?" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) pattern = next(p for p in data["patterns"] if p["keyword"] == "이거 왜") assert "type" in pattern assert "keyword" in pattern assert "context" in pattern assert "count" in pattern assert "first_seen" in pattern assert "last_seen" in pattern assert pattern["count"] == 1 # --------------------------------------------------------------------------- # 7. learnings.json 있는 경우 (누적) # --------------------------------------------------------------------------- class TestExistingLearningsFile: """기존 learnings.json에 누적""" def test_accumulates_new_pattern(self, tmp_path: Path): existing = { "version": 1, "last_updated": "2026-03-10T09:00:00", "patterns": [ { "type": "criticism", "keyword": "이거 왜", "context": "이거 왜 이렇게 느려?", "count": 2, "first_seen": "2026-03-10", "last_seen": "2026-03-10", } ], } make_learnings(tmp_path, existing) text = "잘했어 이번엔" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) # 기존 패턴 + 새 패턴 assert len(data["patterns"]) >= 2 def test_load_learnings_existing(self, tmp_path: Path): existing = { "version": 1, "last_updated": "2026-03-10T09:00:00", "patterns": [ { "type": "praise", "keyword": "좋아", "context": "좋아 이거", "count": 1, "first_seen": "2026-03-10", "last_seen": "2026-03-10", } ], } make_learnings(tmp_path, existing) result = transcript_learner.load_learnings(base_dir=tmp_path) assert len(result["patterns"]) == 1 assert result["patterns"][0]["keyword"] == "좋아" def test_load_learnings_corrupted_returns_empty(self, tmp_path: Path): """손상된 learnings.json → 빈 데이터로 계속""" whisper_dir = tmp_path / "whisper" whisper_dir.mkdir(parents=True, exist_ok=True) (whisper_dir / "learnings.json").write_text("{ not valid json!!!", encoding="utf-8") result = transcript_learner.load_learnings(base_dir=tmp_path) assert result["patterns"] == [] # --------------------------------------------------------------------------- # 8. 중복 방지 테스트 # --------------------------------------------------------------------------- class TestDuplicatePrevention: """중복 학습 방지""" def test_same_context_increases_count(self, tmp_path: Path): text = "이거 왜 이렇게 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) pattern = next(p for p in data["patterns"] if p["keyword"] == "이거 왜") assert pattern["count"] == 2 def test_same_context_no_duplicate_entry(self, tmp_path: Path): text = "이거 왜 이렇게 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) matching = [p for p in data["patterns"] if p["keyword"] == "이거 왜" and p["context"] == text.strip()] assert len(matching) == 1 def test_different_context_adds_new_entry(self, tmp_path: Path): text1 = "이거 왜 이렇게 느려?" text2 = "이거 왜 안 되는 거야?" transcript_learner.process_transcript(text1, base_dir=tmp_path) transcript_learner.process_transcript(text2, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) igeo_wae_patterns = [p for p in data["patterns"] if p["keyword"] == "이거 왜"] assert len(igeo_wae_patterns) == 2 def test_last_seen_updated_on_duplicate(self, tmp_path: Path): existing = { "version": 1, "last_updated": "2026-03-10T09:00:00", "patterns": [ { "type": "criticism", "keyword": "이거 왜", "context": "이거 왜 이렇게 느려?", "count": 1, "first_seen": "2026-03-10", "last_seen": "2026-03-10", } ], } make_learnings(tmp_path, existing) text = "이거 왜 이렇게 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) pattern = next(p for p in data["patterns"] if p["keyword"] == "이거 왜") assert pattern["last_seen"] >= "2026-03-10" assert pattern["count"] == 2 # --------------------------------------------------------------------------- # 9. 빈 입력 graceful 처리 # --------------------------------------------------------------------------- class TestEmptyInput: """빈 입력 처리""" def test_empty_string_no_exception(self, tmp_path: Path): try: transcript_learner.process_transcript("", base_dir=tmp_path) except Exception as e: pytest.fail(f"빈 입력에서 예외 발생: {e}") def test_empty_string_returns_zero_new(self, tmp_path: Path): result = transcript_learner.process_transcript("", base_dir=tmp_path) assert result["new_patterns"] == 0 assert result["updated_patterns"] == 0 def test_whitespace_only_no_exception(self, tmp_path: Path): try: transcript_learner.process_transcript(" \n\n ", base_dir=tmp_path) except Exception as e: pytest.fail(f"공백 입력에서 예외 발생: {e}") def test_no_matching_text_returns_zero_new(self, tmp_path: Path): result = transcript_learner.process_transcript("오늘 날씨가 좋다", base_dir=tmp_path) assert result["new_patterns"] == 0 assert result["updated_patterns"] == 0 # --------------------------------------------------------------------------- # 10. stdout 출력 형식 검증 # --------------------------------------------------------------------------- class TestOutputFormat: """process_transcript() 반환값 형식""" def test_returns_dict_with_required_keys(self, tmp_path: Path): result = transcript_learner.process_transcript("이거 왜 느려?", base_dir=tmp_path) assert "new_patterns" in result assert "updated_patterns" in result assert "total_patterns" in result def test_new_patterns_count(self, tmp_path: Path): text = "이거 왜 느려?\n좋아 이거" result = transcript_learner.process_transcript(text, base_dir=tmp_path) assert result["new_patterns"] >= 2 def test_updated_patterns_count(self, tmp_path: Path): text = "이거 왜 느려?" transcript_learner.process_transcript(text, base_dir=tmp_path) result = transcript_learner.process_transcript(text, base_dir=tmp_path) assert result["updated_patterns"] >= 1 def test_total_patterns_matches_file(self, tmp_path: Path): text = "이거 왜 느려?\n잘했어 이번엔" result = transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) assert result["total_patterns"] == len(data["patterns"]) def test_total_accumulates_across_calls(self, tmp_path: Path): transcript_learner.process_transcript("이거 왜 느려?", base_dir=tmp_path) result = transcript_learner.process_transcript("잘했어 이번엔", base_dir=tmp_path) assert result["total_patterns"] >= 2 # --------------------------------------------------------------------------- # 11. stdin 읽기 테스트 (monkeypatch) # --------------------------------------------------------------------------- class TestStdinReading: """stdin 입력 처리""" def test_read_from_stdin(self, monkeypatch, tmp_path: Path): import io fake_stdin = io.StringIO("이거 왜 이렇게 느려?\n좋아 이거") monkeypatch.setattr(sys, "stdin", fake_stdin) text = transcript_learner.read_input(input_file=None) assert "이거 왜" in text assert "좋아" in text def test_read_from_file(self, tmp_path: Path): f = tmp_path / "session.txt" f.write_text("잘했어 이번에\n항상 이렇게 해줘", encoding="utf-8") text = transcript_learner.read_input(input_file=str(f)) assert "잘했어" in text assert "항상" in text def test_read_from_missing_file_raises(self, tmp_path: Path): with pytest.raises((FileNotFoundError, OSError)): transcript_learner.read_input(input_file=str(tmp_path / "nonexistent.txt")) def test_stdin_text_processed_correctly(self, monkeypatch, tmp_path: Path): import io fake_stdin = io.StringIO("이거 왜 이렇게 느려?") monkeypatch.setattr(sys, "stdin", fake_stdin) text = transcript_learner.read_input(input_file=None) result = transcript_learner.process_transcript(text, base_dir=tmp_path) assert result["new_patterns"] >= 1 # --------------------------------------------------------------------------- # 12. 멀티라인 텍스트 처리 # --------------------------------------------------------------------------- class TestMultilineProcessing: """여러 줄 텍스트 처리""" def test_multiline_detects_all_patterns(self, tmp_path: Path): text = """오늘 회의 내용입니다 이거 왜 이렇게 됐어? 잘했어 이번엔 항상 이렇게 해줘 그냥 평범한 문장 또? 왜 또 이래 """ result = transcript_learner.process_transcript(text, base_dir=tmp_path) assert result["new_patterns"] >= 4 def test_multiline_saves_each_pattern(self, tmp_path: Path): text = "이거 왜 느려?\n좋아 이거야\n항상 이렇게 해" transcript_learner.process_transcript(text, base_dir=tmp_path) data = json.loads(get_learnings_path(tmp_path).read_text(encoding="utf-8")) types = {p["type"] for p in data["patterns"]} assert "criticism" in types assert "praise" in types assert "preference" in types