Update app.py

app.py

@@ -39,7 +39,7 @@ def download_dataset():
 # ✅ Download dataset at startup
 download_dataset()
 
-# ✅ Load dataset with proper error handling
+# ✅ Load dataset with error handling
 try:
     df = pd.read_csv(DATASET_PATH, encoding="utf-8", delimiter=",", on_bad_lines="skip")
 except:
@@ -54,14 +54,17 @@ for col in required_columns:
     if col not in df.columns:
         raise Exception(f"❌ Column '{col}' is missing from the dataset!")
 
+# ✅ Convert "Case Problem" column to lowercase for consistency
+df["Case Problem"] = df["Case Problem"].str.lower()
+
 # ✅ Encode categorical variables
 encoder = LabelEncoder()
-df["Case Problem"] = encoder.fit_transform(df["Case Problem"])
-df["Feedback"] = encoder.fit_transform(df["Feedback"])
+df["Case Problem Encoded"] = encoder.fit_transform(df["Case Problem"])
+df["Feedback Encoded"] = encoder.fit_transform(df["Feedback"])
 
 # ✅ Train Model
-X = df[["Case Problem"]]
-y = df["Feedback"]
+X = df[["Case Problem Encoded"]]
+y = df["Feedback Encoded"]
 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
 model = RandomForestClassifier(n_estimators=100, random_state=42)
 model.fit(X_train, y_train)
@@ -70,6 +73,9 @@ model.fit(X_train, y_train)
 joblib.dump(model, "feedback_model.pkl")
 print("✅ Model trained successfully!")
 
+# ✅ Save encoder classes for future use
+joblib.dump(encoder, "case_problem_encoder.pkl")
+
 # ✅ API Input Model
 class PredictionInput(BaseModel):
     case_problem: str
@@ -80,26 +86,34 @@ def predict_feedback(data: PredictionInput):
     if model is None:
         return {"error": "Model is not trained yet."}
 
-    try:
-        case_problem_encoded = encoder.transform([data.case_problem])
-        prediction = model.predict([[case_problem_encoded[0]]])
-        feedback_predicted = encoder.inverse_transform(prediction)[0]
-        return {"Predicted Feedback": feedback_predicted}
-    except Exception as e:
-        return {"error": str(e)}
+    # ✅ Convert input to lowercase to match training data
+    case_problem_lower = data.case_problem.lower()
+
+    if case_problem_lower not in df["Case Problem"].values:
+        return {"error": "Invalid case problem. Please enter a valid category from the dataset."}
+
+    case_problem_encoded = encoder.transform([case_problem_lower])
+    prediction = model.predict([[case_problem_encoded[0]]])
+    feedback_predicted = encoder.inverse_transform(prediction)[0]
+
+    return {"Predicted Feedback": feedback_predicted}
 
 # ✅ Gradio UI
 def gradio_interface(case_problem):
     if model is None:
         return "Model not trained yet."
 
-    try:
-        case_problem_encoded = encoder.transform([case_problem])
-        prediction = model.predict([[case_problem_encoded[0]]])
-        feedback_predicted = encoder.inverse_transform(prediction)[0]
-        return f"Predicted Feedback: {feedback_predicted}"
-    except:
-        return "Invalid case problem. Please enter a valid category."
+    # ✅ Convert input to lowercase for consistency
+    case_problem_lower = case_problem.lower()
+
+    if case_problem_lower not in df["Case Problem"].values:
+        return "Invalid case problem. Please enter a valid category from the dataset."
+
+    case_problem_encoded = encoder.transform([case_problem_lower])
+    prediction = model.predict([[case_problem_encoded[0]]])
+    feedback_predicted = encoder.inverse_transform(prediction)[0]
+
+    return f"Predicted Feedback: {feedback_predicted}"
 
 # ✅ Start both API & Gradio
 def start_app():