Update app.py

app.py

@@ -1,5 +1,4 @@
 from fastapi import FastAPI
-from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
 import joblib
 import logging
@@ -11,18 +10,10 @@ logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(
 
 app = FastAPI()
 
-# === CORS (fixes connection issues with frontend) ===
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
 models = None
 
 def to_native(value):
+    """Fix for numpy.int64 / numpy types that FastAPI can't serialize"""
     if isinstance(value, np.integer):
         return int(value)
     elif isinstance(value, np.floating):
@@ -34,29 +25,23 @@ def load_models():
     if models is not None:
         return
     logging.info("Loading models...")
-    try:
-        # New English Model
-        repo_id = "E-motionAssistant/English_LR_Model_New"
-        en_vectorizer = joblib.load(hf_hub_download(repo_id, "tfidf_vectorizer.joblib"))
-        en_classifier = joblib.load(hf_hub_download(repo_id, "logreg_model.joblib"))
-        en_label_encoder = joblib.load(hf_hub_download(repo_id, "label_encoder.joblib"))
-
-        # Sinhala
-        si_vectorizer = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "tfidf_vectorizer.joblib"))
-        si_classifier = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "logreg_model.joblib"))
-        si_label_encoder = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "label_encoder.joblib"))
-
-        # Tamil
-        tamil_pipe = pipeline("text-classification", 
-                            model="E-motionAssistant/Tamil_Emotion_Recognition_Model", 
-                            device=-1)
-
-        models = (en_vectorizer, en_classifier, en_label_encoder,
-                  si_vectorizer, si_classifier, si_label_encoder, tamil_pipe)
-        logging.info("✅ All models loaded successfully.")
-    except Exception as e:
-        logging.error(f"Model loading failed: {type(e).__name__} - {e}")
-        raise
+    
+    # ==================== CHANGED ONLY THIS PART ====================
+    en_vectorizer = joblib.load(hf_hub_download("E-motionAssistant/English_LR_Model_New", "tfidf_vectorizer.joblib"))
+    en_classifier = joblib.load(hf_hub_download("E-motionAssistant/English_LR_Model_New", "logreg_model.joblib"))
+    en_label_encoder = joblib.load(hf_hub_download("E-motionAssistant/English_LR_Model_New", "label_encoder.joblib"))
+    # ================================================================
+
+    si_vectorizer = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "tfidf_vectorizer.joblib"))
+    si_classifier = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "logreg_model.joblib"))
+    si_label_encoder = joblib.load(hf_hub_download("E-motionAssistant/Sinhala_Text_Emotion_Model_LR", "label_encoder.joblib"))
+
+    tamil_pipe = pipeline("text-classification", model="E-motionAssistant/Tamil_Emotion_Recognition_Model", device=-1)
+
+    models = (en_vectorizer, en_classifier, en_label_encoder,
+              si_vectorizer, si_classifier, si_label_encoder, tamil_pipe)
+    
+    logging.info("✅ All models loaded.")
 
 @app.on_event("startup")
 def startup_event():
@@ -72,36 +57,35 @@ def root():
 
 @app.post("/predict")
 def predict(req: PredictRequest):
-    if not req.text or not req.text.strip():
+    if not req.text.strip():
         return {"error": "Text cannot be empty"}
 
     en_vec, en_clf, en_le, si_vec, si_clf, si_le, tamil_pipe = models
 
     try:
-        lang = req.language.lower()
-        if lang == "english":
+        if req.language.lower() == "english":          # Made case-insensitive (safer)
             X = en_vec.transform([req.text])
             pred = en_clf.predict(X)[0]
             emotion = en_le.inverse_transform([pred])[0]
-            emotion = to_native(emotion)
+            emotion = to_native(emotion)               # ← This fixes the 500 error
             return {"emotion": emotion, "language": "English"}
 
-        elif lang == "sinhala":
+        elif req.language.lower() == "sinhala":
             X = si_vec.transform([req.text])
             pred = si_clf.predict(X)[0]
             emotion = si_le.inverse_transform([pred])[0]
             emotion = to_native(emotion)
             return {"emotion": emotion, "language": "Sinhala"}
 
-        elif lang == "tamil":
+        elif req.language.lower() == "tamil":
             res = tamil_pipe(req.text)[0]
-            return {
-                "emotion": res["label"],
-                "confidence": round(float(res["score"]), 3),
-                "language": "Tamil"
-            }
+            return {"emotion": res["label"], 
+                    "confidence": round(res["score"], 3), 
+                    "language": "Tamil"}
+
         else:
             return {"error": f"Unsupported language: {req.language}"}
+
     except Exception as e:
-        logging.error(f"Prediction error: {type(e).__name__} - {e}")
-        return {"error": "Prediction failed. Please try again later."}
+        logging.error(f"Prediction error: {e}")
+        return {"error": "Prediction failed. Please try again."}