Update app.py

app.py

@@ -6,7 +6,6 @@ from tensorflow.keras.layers import Dense
 from tensorflow.keras.preprocessing.sequence import pad_sequences
 import numpy as np
 import pickle
-from stable_baselines3 import PPO
 
 app = FastAPI(title="Transaction Classifier API", description="API for classifying banking transactions.")
 print("FastAPI app initialized...")
@@ -72,15 +71,8 @@ def load_resources():
     print(f"Num categories: {len(le_category.classes_)}, Num subcategories: {len(le_subcategory.classes_)}")
     return model, tokenizer, le_category, le_subcategory
 
-def load_ppo_model():
-    print("Loading PPO model...")
-    ppo = PPO.load('ppo_finetuned_model')
-    print("PPO model loaded.")
-    return ppo
-
 class TransactionRequest(BaseModel):
     description: str
-    use_rl: Optional[bool] = False
 
 class PredictionResponse(BaseModel):
     category: str
@@ -96,30 +88,20 @@ async def root():
 async def predict(request: TransactionRequest):
     try:
         model, tokenizer, le_category, le_subcategory = load_resources()
-        num_subcategories = len(le_subcategory.classes_)
         
         seq = tokenizer.texts_to_sequences([request.description])
         pad = pad_sequences(seq, maxlen=max_len, padding='post')
         pred = model.predict(pad, verbose=0)
         
-        if request.use_rl:
-            ppo = load_ppo_model()
-            obs = pad[0]
-            action, _ = ppo.predict(obs)  # Match Colab: no deterministic=True
-            print(f"RL Action: {action}, Observation: {obs}")
-            cat_idx = action // num_subcategories
-            subcat_idx = action % num_subcategories
-            print(f"RL cat_idx: {cat_idx}, subcat_idx: {subcat_idx}")
-        else:
-            cat_probs = pred[0][0]
-            subcat_probs = pred[1][0]
-            cat_idx = np.argmax(cat_probs)
-            subcat_idx = np.argmax(subcat_probs)
+        cat_probs = pred[0][0]
+        subcat_probs = pred[1][0]
+        cat_idx = np.argmax(cat_probs)
+        subcat_idx = np.argmax(subcat_probs)
         
         cat_pred = le_category.inverse_transform([cat_idx])[0]
         subcat_pred = le_subcategory.inverse_transform([subcat_idx])[0]
-        cat_conf = float(pred[0][0][cat_idx] * 100)
-        subcat_conf = float(pred[1][0][subcat_idx] * 100)
+        cat_conf = float(cat_probs[cat_idx] * 100)
+        subcat_conf = float(subcat_probs[subcat_idx] * 100)
         
         print(f"Predicted: category={cat_pred}, subcategory={subcat_pred}")
         return {