app.py

import streamlit as st
import requests
from Pandas_Market_Predictor import Pandas_Market_Predictor
import pandas as pd
from sklearn.decomposition import PCA

# Hard-coded API key for demonstration purposes
API_KEY = "QR8F9B7T6R2SWTAT"

def fetch_alpha_vantage_data(api_key):
    url = f'https://www.alphavantage.co/query?function=TIME_SERIES_INTRADAY&symbol=IBM&interval=60min&apikey={api_key}'
    response = requests.get(url)
    alpha_vantage_data = response.json()
    return alpha_vantage_data

def calculate_indicators(data):
    # Convert all columns to numeric
    data = data.apply(pd.to_numeric, errors='coerce')

    # Example: Simple condition for doji and inside
    data['Doji'] = abs(data['close'] - data['open']) <= 0.01 * (data['high'] - data['low'])
    data['Inside'] = (data['high'] < data['high'].shift(1)) & (data['low'] > data['low'].shift(1))

    # Include the necessary columns from the original dataset
    data['open'] = data['open_original']
    data['high'] = data['high_original']
    data['low'] = data['low_original']

    return data

def apply_pca(data, n_components=5):
    pca = PCA(n_components=n_components)
    reduced_data = pca.fit_transform(data)
    return pd.DataFrame(reduced_data, index=data.index, columns=[f'component_{i+1}' for i in range(n_components)])

def main():
    st.title("Stock Trend Predictor")

    # Use the hard-coded API key
    api_key = API_KEY

    # Fetch Alpha Vantage data
    alpha_vantage_data = fetch_alpha_vantage_data(api_key)

    # Extract relevant data from Alpha Vantage response
    alpha_vantage_time_series = alpha_vantage_data.get('Time Series (60min)', {})
    df = pd.DataFrame(alpha_vantage_time_series).T
    df.index = pd.to_datetime(df.index)
    df = df.dropna(axis=0)

    # Rename columns
    df = df.rename(columns={'1. open': 'open_original', '2. high': 'high_original', '3. low': 'low_original', '4. close': 'Close', '5. volume': 'volume'})

    # Apply PCA for dimensionality reduction
    df_reduced = apply_pca(df[['open_original', 'high_original', 'low_original', 'Close', 'volume']])

    # Calculate indicators on reduced data
    df_reduced = calculate_indicators(df_reduced)

    # Create predictor
    my_market_predictor = Pandas_Market_Predictor(df_reduced)

    # Predict Trend
    indicators = ["Doji", "Inside"]
    trend = my_market_predictor.Trend_Detection(indicators, 10)

    # Display results
    st.subheader("Predicted Trend:")
    st.write("Buy Trend :", trend['BUY'])
    st.write("Sell Trend :", trend['SELL'])
    st.write(f"Standard Deviation Percentage: {my_market_predictor.PERCENT_STD}%")

    # Delete the DataFrame to release memory
    del df

if __name__ == "__main__":
    main()