import os
import pickle
import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Input, Embedding, LSTM, Bidirectional, GlobalAveragePooling1D, Dense, Dropout
import gradio as gr


MAX_LEN = 200
MODEL_PATH = "sentiment_model.keras"
TOKENIZER_PATH = "tokenizer.pickle"




if os.path.exists(TOKENIZER_PATH):
    with open(TOKENIZER_PATH, "rb") as handle:
        tokenizer = pickle.load(handle)
    print("Tokenizer loaded successfully.")
else:
    raise FileNotFoundError(f"Missing critical asset: {TOKENIZER_PATH}")




def build_model():
    
    model=Sequential()
    model.add(Input(shape=(200,)))
    model.add(Embedding(5000,128))
    model.add(LSTM(128,dropout=0.2))
    model.add(Dense(1,activation='sigmoid'))
    return model

model = build_model()


model.load_weights("sentiment_weights.weights.h5")



def analyze_text(text):
    # Handle empty inputs gracefully
    if not text or not text.strip():
        return {"Invalid Input: Please type some text.": 1.0}
    
    # Preprocessing pipeline
    sequences = tokenizer.texts_to_sequences([text])
    padded = pad_sequences(sequences, maxlen=MAX_LEN)
    
    # Model inference
    prediction = float(model.predict(padded)[0][0])
    
    
    return {
    "Positive 😊": prediction,
    "Negative 😞": 1.0 - prediction
}

#  GRADIO BLOCKS LAYOUT


with gr.Blocks(theme=gr.themes.Soft(), title="Sentiment Analyzer Pro") as demo:
    
    # Header Section
    gr.Markdown(
        """
        # 🧠 Deep Learning Sentiment Classifier
        Analyze the emotional polarity of text using an optimized **Bidirectional LSTM** recurrent neural network.
        """
    )
    
    # Workspace Layout (
    with gr.Row():
       
        with gr.Column(scale=2):
            input_text = gr.Textbox(
                label="Input Review / Text Passage",
                placeholder="Type or paste your text here (e.g., product review, tweet, movie feedback)...",
                lines=6,
                max_lines=10,
                interactive=True
            )
            
            with gr.Row():
                clear_btn = gr.Button("Clear", variant="secondary")
                submit_btn = gr.Button("Analyze Sentiment", variant="primary")
        
       
        with gr.Column(scale=1):
            output_label = gr.Label(
                label="Classification Output",
                num_top_classes=2,
                show_label=True
            )
            
  
    gr.Examples(
        examples=[
            ["The cinematic direction was absolutely breathtaking, and the cast delivered a masterful performance. Highly recommended!"],
            ["An absolute disaster. The pacing was agonizingly slow, the dialogue felt completely forced, and I walked out early."],
            ["The device has decent battery life and a nice screen, but the software bugs make it frustrating to use daily."]
        ],
        inputs=input_text,
        label="Click an example to test instantly:"
    )
    
    
    submit_btn.click(
        fn=analyze_text, 
        inputs=input_text, 
        outputs=output_label
    )

    input_text.submit(
        fn=analyze_text, 
        inputs=input_text, 
        outputs=output_label
    )

    clear_btn.click(
        fn=lambda: ("", None), 
        inputs=None, 
        outputs=[input_text, output_label]
    )


if __name__ == "__main__":
    
    demo.launch()