Chore[#79]: 임베딩 모델 교체

.env.local

@@ -7,11 +7,11 @@ DB_NAME=leadme
 
 # MONGO
 MONGO_DB_NAME=highfive
-MONGO_DB_HOST=1232132131
-MONGO_DB_PORT=123213
+MONGO_DB_HOST=localhost
+MONGO_DB_PORT=27017
 
 # Word2Vec Model path
-W2V_MODEL_PATH=./app/models/GoogleNews-vectors-negative300.bin
+W2V_MODEL_PATH=./app/models/node2vec.model
 
 # RabbitMQ
 RABBITMQ_URL=localhost

.github/workflows/ci.yml

@@ -27,7 +27,6 @@ jobs:
           echo "DB_PASSWORD=testpass" >> .env.local
           echo "DB_NAME=testdb" >> .env.local
           echo "W2V_MODEL_PATH=./app/assets/word2vec.model" >> .env.local
-          echo "MONGO_URL=mongodb://localhost:27017" >> .env.local
           echo "RABBITMQ_URL=localhost" >> .env.local
           echo "RABBITMQ_PORT=5672" >> .env.local
           echo "DEV_REDIS_HOST=localhost" >> .env.local

.gitignore

@@ -15,4 +15,5 @@ __pycache__/
 *.log
 
 # w2v 모델
-GoogleNews-vectors-negative300.bin.gz
+GoogleNews-vectors-negative300.bin.gz
+node2vec.model

app/models/db_w2v_mapper.py

@@ -6,7 +6,7 @@
     "가족": "Family",
     "코미디": "Comedy",
     "애니메이션": "Animation",
-    "SF": "SF",
+    "SF": "Science Fiction",
     "다큐멘터리": "Documentary",
     "판타지": "Fantasy",
     "드라마": "Drama",
@@ -24,5 +24,14 @@
 }
 
 
-def translate_genre(korean_genre: str) -> str:
-    return _genre_mapping.get(korean_genre, korean_genre)
+def translate_genre(genre: str) -> str | None:
+    genre = genre.strip()
+
+    if genre in _genre_mapping:
+        return _genre_mapping[genre]
+
+    if genre in _genre_mapping.values():
+        return genre
+
+    return None
+

app/models/word2vec_model.py

@@ -1,4 +1,4 @@
-from gensim.models import KeyedVectors
+from gensim.models import Word2Vec
 
 
 class Word2VecModel:
@@ -8,15 +8,15 @@ class Word2VecModel:
     def load_model(cls, model_path: str):
         if cls._model is None:
             print("모델 로딩 중...")
-            cls._model = KeyedVectors.load_word2vec_format(model_path, binary=True)
+            cls._model = Word2Vec.load(model_path)
             print("모델 로드 완료!")
         return cls._model
 
     @classmethod
     def get_vector(cls, word: str):
         if cls._model is None:
             raise RuntimeError("모델이 아직 로드되지 않았습니다.")
-        return cls._model.get_vector(word)
+        return cls._model.wv.get_vector(word)
 
     @classmethod
     def similarity(cls, word1: str, word2: str) -> float: