import gradio as gr
import tensorflow as tf
from transformers import AutoTokenizer

# Load the model
from huggingface_hub import hf_hub_download

model_path = hf_hub_download(repo_id="anand6572r/my-keras-model11", filename="trained_model.keras")

model = tf.keras.models.load_model(model_path)



# Load a Hugging Face tokenizer (use a compatible model's tokenizer)
tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")

# Define preprocessing function
def preprocess_input(query):
    """
    Tokenizes and formats the input text for the model using Hugging Face's tokenizer.
    """
    inputs = tokenizer(
        query,
        padding="max_length",
        max_length=100,
        truncation=True,
        return_tensors="tf",
    )
    return inputs["input_ids"]

# Define postprocessing function
def postprocess_output(prediction):
    """
    Converts model predictions into a user-friendly response.
    """
    if prediction[0] > 0.5:  # Adjust threshold if necessary
        return "Yes, this relates to a Wikipedia article."
    else:
        return "No, this does not relate to a Wikipedia article."

# Define prediction function
def predict(query):
    """
    Predicts whether the query relates to a Wikipedia article.
    """
    # Preprocess the input query
    input_data = preprocess_input(query)
    
    # Get prediction from the model
    prediction = model.predict(input_data)
    
    # Postprocess the prediction
    response = postprocess_output(prediction[0])
    return response

# Gradio interface
interface = gr.Interface(
    fn=predict,
    inputs=gr.Textbox(lines=2, placeholder="Enter your query here..."),
    outputs=gr.Textbox(label="Response"),
    title="Wikipedia Article Query Predictor",
    description="This model predicts whether a query relates to a Wikipedia article.",
)

# Launch the Gradio app
if __name__ == "__main__":
    interface.launch()