from flask import Flask, render_template, request, jsonify
import joblib
import sys
import traceback
import pandas as pd
import re

app = Flask(__name__)

# -----------------------------
# Load kamus CSV (delimiter ;)
# -----------------------------
kamus_df = pd.read_csv('kamus_normalize.csv', sep=';')
kamus_df.columns = kamus_df.columns.str.strip()  # hilangkan spasi di nama kolom
kamus = dict(zip(kamus_df['BEFORE'], kamus_df['AFTER']))

# -----------------------------
# Preprocessing function
# -----------------------------
def preprocess(text):
    text = text.lower()
    for k, v in kamus.items():
        text = text.replace(k, v)
    text = re.sub(r'[^a-z\s]', '', text)
    text = re.sub(r'\s+', ' ', text).strip()
    return text

# -----------------------------
# Load model, vectorizer, LabelEncoder
# -----------------------------
try:
    print("Python version:", sys.version)
    
    print("Loading TF-IDF vectorizer...")
    # Jika vectorizer pernah pakai custom analyzer, definisikan identity dulu:
    # def identity(x): return x
    vectorizer = joblib.load('tfidf_vectorizer.pkl')
    print("Vectorizer loaded:", type(vectorizer))

    print("Loading Naive Bayes model...")
    model = joblib.load('nb_classifier.pkl')
    print("Model loaded:", type(model))

    print("Loading LabelEncoder...")
    le = joblib.load('label_encoder.pkl')
    print("LabelEncoder loaded:", type(le))

except FileNotFoundError as e:
    print(f"Error: {e}")
    raise
except Exception as e:
    print(f"Error loading models: {str(e)}")
    traceback.print_exc()
    raise

# -----------------------------
# Routes
# -----------------------------
@app.route('/')
def home():
    return render_template('index.html')

@app.route('/analyze', methods=['POST'])
def analyze():
    if request.method == 'POST':
        try:
            text = request.form['text']
            print(f"Received text: {text}")

            # Preprocess input
            text_clean = preprocess(text)
            print(f"Preprocessed text: {text_clean}")

            # Transform and predict
            X = vectorizer.transform([text_clean])
            pred_encoded = model.predict(X)[0]
            pred_label = le.inverse_transform([pred_encoded])[0]  # output string
            print(f"Predicted label: {pred_label}")

            # Get prediction probabilities
            probs = model.predict_proba(X)[0]

            # Ambil index probabilitas positif/negatif
            labels = le.inverse_transform(range(len(probs)))  # misal ['negatif', 'positif']
            neg_idx = list(labels).index('negatif')
            pos_idx = list(labels).index('positif')

            # Map ke key JS yang sesuai
            prob_percentages = {
            'Positive': f"{probs[pos_idx]*100:.2f}%",
            'Negative': f"{probs[neg_idx]*100:.2f}%"}

            # Pastikan sentimen dikirim bahasa Inggris untuk JS
            pred_label_lower = pred_label.lower().strip()
            if pred_label_lower == 'positif':
                sentiment_str = 'Positive'
            else:
                sentiment_str = 'Negative'
                
            # Map probabilities to correct labels (hanya negatif/positif)
            # label_order = le.inverse_transform(range(len(probs)))  # ['negatif', 'positif']
            # prob_percentages = {label.capitalize(): f"{prob*100:.2f}%" for label, prob in zip(label_order, probs)}
            
            return jsonify({
                'sentiment': pred_label.capitalize(),
                'probabilities': prob_percentages
            })

        except Exception as e:
            print(f"Error during analysis: {str(e)}")
            traceback.print_exc()
            return jsonify({'error': str(e)}), 500

# -----------------------------
# Run Flask
# -----------------------------
if __name__ == '__main__':
    app.run(debug=True)