import gradio as gr
import pandas as pd
import numpy as np
import yfinance as yf
from sklearn.preprocessing import MinMaxScaler
from tensorflow import keras

# Load your trained model
model = keras.models.load_model('your_model.h5')  # Ensure this path is correct

# Function to predict stock prices
def predict_stock_price(stock_ticker, start_date, end_date):
    # Fetch data
    data = yf.download(stock_ticker, start=start_date, end=end_date)
    
    # Check if data is returned
    if data.empty:
        return "No data available for the selected dates."

    # Preprocess data
    scaler = MinMaxScaler()
    scaled_data = scaler.fit_transform(data['Close'].values.reshape(-1, 1))

    # Prepare input for the model
    input_data = scaled_data[-60:]  # Use the last 60 days of data
    input_data = input_data.reshape((1, input_data.shape[0], 1))

    # Predict stock prices
    prediction = model.predict(input_data)
    predicted_price = scaler.inverse_transform(prediction)  # Rescale back to original price
    
    return f"Predicted stock price for tomorrow: ${predicted_price[0][0]:.2f}"

# Create the Gradio interface
stock_ticker_input = gr.Dropdown(
    choices=["AAPL", "GOOGL", "MSFT", "AMZN", "TSLA"],  # Add more tickers as needed
    label="Select Stock Ticker"
)

start_date_input = gr.Date(label="Start Date")
end_date_input = gr.Date(label="End Date")

iface = gr.Interface(
    fn=predict_stock_price,
    inputs=[
        stock_ticker_input,
        start_date_input,
        end_date_input
    ],
    outputs="text",
    title="Stock Price Prediction App",
    description="Enter the stock ticker and date range to predict the stock price for tomorrow."
)

# Launch the Gradio app
iface.launch()