[CHORE] 감정 예측 엔드포인트 정리 및 LABELS 상수 정리

config.py

@@ -1,9 +1,9 @@
 import os
+
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
+
 class Settings(BaseSettings):
-    # mongodb_uri: str = "mongodb://localhost:27017"
-    # db_name: str = "emotion_db"
     collection_name: str = "predictions"
     model_dir: str = os.getenv("MODEL_DIR", "models/0717_kobert_5_emotion_model")
     title: str = "MovieMood - KoBERT Emotion API"

const.py

@@ -1,11 +1,5 @@
 # 7 가지 감정 상태
 # 기쁨, 슬픔, 분노, 놀람, 혐오, 공포, 중립
-LABELS_6 = [
-    "anger", "fear", "joy", "neutral", "sadness", "surprise"
-]
-# LABELS_5 = [
-#     "anger", "fear", "joy", "neutral", "sadness"
-# ]
 
 LABELS_5 = [
     "anger", "disgust", "fear", "joy", "sadness"

main.py

@@ -1,5 +1,6 @@
 import uvicorn
 from fastapi import FastAPI
+
 from config import settings
 from routers.home import router as home_router
 from routers.predict import router as predict_router

routers/home.py

@@ -1,4 +1,5 @@
 from fastapi import APIRouter
+
 from config import settings
 
 router = APIRouter()

routers/predict.py

@@ -1,41 +1,14 @@
-from fastapi import APIRouter, HTTPException
 from datetime import datetime
+
+from fastapi import APIRouter, HTTPException
+
 from schemas import TextItem, Prediction
 from services.prediction import (
-    predict_emotion_full,
-    predict_emotion_split_avg,
     predict_emotion_overall_avg,
 )
 
 router = APIRouter(prefix="/predict", tags=["Prediction"])
 
-@router.post("/full", response_model=Prediction)
-async def predict_full(item: TextItem):
-    try:
-        probs = predict_emotion_full(item.text)
-        record = {
-            "text": item.text,
-            "probabilities": probs,
-            "timestamp": datetime.utcnow()
-        }
-        # await collection.insert_one(record)
-        return record
-    except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
-
-@router.post("/split_avg", response_model=Prediction)
-async def predict_split_avg(item: TextItem):
-    try:
-        probs = predict_emotion_split_avg(item.text)
-        record = {
-            "text": item.text,
-            "probabilities": probs,
-            "timestamp": datetime.utcnow()
-        }
-        # await collection.insert_one(record)
-        return record
-    except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
 
 @router.post("/overall_avg", response_model=Prediction)
 async def predict_overall_avg(item: TextItem):
@@ -46,7 +19,6 @@ async def predict_overall_avg(item: TextItem):
             "probabilities": probs,
             "timestamp": datetime.utcnow()
         }
-        # await collection.insert_one(record)
         return record
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))

routers/review.py

@@ -1,8 +1,7 @@
 # TODO: 데이터 저장하기(mongo or mysql)
-from fastapi import APIRouter, HTTPException
-from pydantic import BaseModel
-
 import requests
+from fastapi import APIRouter
+from pydantic import BaseModel
 
 router = APIRouter(
   prefix = "/review",

schemas.py

@@ -3,6 +3,7 @@
 
 from pydantic import BaseModel
 
+
 class TextItem(BaseModel):
     text: str
 

services/prediction.py

@@ -1,8 +1,9 @@
 import torch
-from transformers import AutoModelForSequenceClassification
 from kobert_tokenizer import KoBERTTokenizer
-from const import LABELS_5
+from transformers import AutoModelForSequenceClassification
+
 from config import settings
+from const import LABELS_5
 
 _tokenizer = KoBERTTokenizer.from_pretrained(settings.model_dir)
 _model     = AutoModelForSequenceClassification.from_pretrained(settings.model_dir)