app.py

#!/usr/bin/env python3
"""
FinGPT-Forecaster Hugging Face Space
Market Forecaster Agent - Predict Stock Movements Direction
"""

import os
import json
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import mplfinance as mpf
from datetime import datetime, timedelta
import finnhub
import yfinance as yf
import streamlit as st
import warnings
warnings.filterwarnings('ignore')

# Set page config
st.set_page_config(
    page_title="FinGPT-Forecaster",
    page_icon="📈",
    layout="wide",
    initial_sidebar_state="expanded"
)

class MarketForecaster:
    def __init__(self):
        """Initialize the Market Forecaster"""
        self.finnhub_client = None
        self.setup_finnhub()
        
    def setup_finnhub(self):
        """Setup Finnhub client with API key from environment or use demo mode"""
        finnhub_api_key = os.getenv('FINNHUB_API_KEY')
        if finnhub_api_key:
            try:
                self.finnhub_client = finnhub.Client(api_key=finnhub_api_key)
                st.success("✅ Connected to Finnhub API")
            except Exception as e:
                st.warning(f"⚠️ Finnhub API connection failed: {e}")
                self.finnhub_client = None
        else:
            st.info("ℹ️ Running in demo mode (no Finnhub API key provided)")
            self.finnhub_client = None
    
    def get_company_profile(self, symbol):
        """Get company profile from Finnhub or return demo data"""
        if self.finnhub_client:
            try:
                profile = self.finnhub_client.company_profile2(symbol=symbol)
                return profile
            except Exception as e:
                st.warning(f"Error getting company profile: {e}")
        
        # Return demo data
        return {
            'name': f'{symbol} Corporation',
            'finnhubIndustry': 'Technology',
            'marketCapitalization': 1000000000,
            'country': 'US',
            'currency': 'USD'
        }
    
    def get_company_news(self, symbol, start_date, end_date):
        """Get company news from Finnhub or return demo data"""
        if self.finnhub_client:
            try:
                start_ts = int(datetime.strptime(start_date, '%Y-%m-%d').timestamp())
                end_ts = int(datetime.strptime(end_date, '%Y-%m-%d').timestamp())
                news = self.finnhub_client.company_news(symbol, _from=start_ts, to=end_ts)
                return news
            except Exception as e:
                st.warning(f"Error getting news: {e}")
        
        # Return demo news
        return [
            {
                "headline": f"{symbol} shows strong quarterly performance",
                "summary": f"Recent earnings report shows {symbol} exceeding expectations with robust growth in key segments."
            },
            {
                "headline": f"Market analysts upgrade {symbol} rating",
                "summary": f"Several analysts have upgraded their rating for {symbol} citing strong fundamentals and growth prospects."
            },
            {
                "headline": f"{symbol} announces new strategic initiatives",
                "summary": f"Company announces new strategic initiatives aimed at expanding market presence and driving innovation."
            }
        ]
    
    def get_stock_data(self, symbol, start_date, end_date):
        """Get stock price data from Yahoo Finance"""
        try:
            ticker = yf.Ticker(symbol)
            data = ticker.history(start=start_date, end=end_date)
            return data
        except Exception as e:
            st.error(f"Error getting stock data for {symbol}: {e}")
            return None
    
    def calculate_rsi(self, prices, window=14):
        """Calculate RSI indicator"""
        delta = prices.diff()
        gain = (delta.where(delta > 0, 0)).rolling(window=window).mean()
        loss = (-delta.where(delta < 0, 0)).rolling(window=window).mean()
        rs = gain / loss
        rsi = 100 - (100 / (1 + rs))
        return rsi
    
    def analyze_stock_movement(self, symbol, days_back=30):
        """Analyze stock movement and generate prediction"""
        # Get current date and calculate date range
        end_date = datetime.now().strftime('%Y-%m-%d')
        start_date = (datetime.now() - timedelta(days=days_back)).strftime('%Y-%m-%d')
        
        # Get data from various sources
        with st.spinner(f"📊 Fetching data for {symbol}..."):
            # Company profile
            profile = self.get_company_profile(symbol)
            
            # Recent news
            news = self.get_company_news(symbol, start_date, end_date)
            
            # Stock price data
            stock_data = self.get_stock_data(symbol, start_date, end_date)
        
        if stock_data is None or stock_data.empty:
            st.error(f"❌ No stock data available for {symbol}")
            return None
        
        # Calculate technical indicators
        stock_data['SMA_20'] = stock_data['Close'].rolling(window=20).mean()
        stock_data['SMA_50'] = stock_data['Close'].rolling(window=50).mean()
        stock_data['RSI'] = self.calculate_rsi(stock_data['Close'])
        
        # Recent price performance
        recent_close = stock_data['Close'].iloc[-1]
        week_ago_close = stock_data['Close'].iloc[-5] if len(stock_data) >= 5 else recent_close
        month_ago_close = stock_data['Close'].iloc[-20] if len(stock_data) >= 20 else recent_close
        
        week_change = ((recent_close - week_ago_close) / week_ago_close) * 100
        month_change = ((recent_close - month_ago_close) / month_ago_close) * 100
        
        # Generate analysis
        analysis = self.generate_analysis(symbol, profile, news, stock_data, recent_close, week_change, month_change)
        
        return analysis, stock_data
    
    def generate_analysis(self, symbol, profile, news, stock_data, current_price, week_change, month_change):
        """Generate comprehensive analysis and prediction"""
        
        # Analyze news sentiment
        positive_factors = []
        negative_factors = []
        
        if news:
            for article in news[:10]:  # Analyze top 10 recent news
                headline = article.get('headline', '').lower()
                summary = article.get('summary', '').lower()
                
                # Simple keyword-based sentiment analysis
                positive_keywords = ['growth', 'profit', 'revenue', 'beat', 'exceed', 'strong', 'positive', 'upgrade', 'buy', 'bullish']
                negative_keywords = ['loss', 'decline', 'miss', 'weak', 'negative', 'downgrade', 'sell', 'bearish', 'concern', 'risk']
                
                pos_score = sum(1 for word in positive_keywords if word in headline or word in summary)
                neg_score = sum(1 for word in negative_keywords if word in headline or word in summary)
                
                if pos_score > neg_score:
                    positive_factors.append(article.get('headline', '')[:100])
                elif neg_score > pos_score:
                    negative_factors.append(article.get('headline', '')[:100])
        
        # Technical analysis
        recent_rsi = stock_data['RSI'].iloc[-1] if not stock_data['RSI'].isna().iloc[-1] else 50
        sma_20 = stock_data['SMA_20'].iloc[-1] if not stock_data['SMA_20'].isna().iloc[-1] else current_price
        sma_50 = stock_data['SMA_50'].iloc[-1] if not stock_data['SMA_50'].isna().iloc[-1] else current_price
        
        # Generate prediction
        prediction_score = 0
        
        # RSI analysis
        if recent_rsi < 30:
            prediction_score += 2  # Oversold, potential bounce
            positive_factors.append("RSI indicates oversold conditions")
        elif recent_rsi > 70:
            prediction_score -= 2  # Overbought, potential pullback
            negative_factors.append("RSI indicates overbought conditions")
        
        # Moving average analysis
        if current_price > sma_20 > sma_50:
            prediction_score += 1
            positive_factors.append("Price above both 20-day and 50-day moving averages")
        elif current_price < sma_20 < sma_50:
            prediction_score -= 1
            negative_factors.append("Price below both 20-day and 50-day moving averages")
        
        # Recent performance
        if week_change > 2:
            prediction_score += 1
            positive_factors.append(f"Strong weekly performance (+{week_change:.1f}%)")
        elif week_change < -2:
            prediction_score -= 1
            negative_factors.append(f"Weak weekly performance ({week_change:.1f}%)")
        
        # News sentiment
        prediction_score += len(positive_factors) * 0.5
        prediction_score -= len(negative_factors) * 0.5
        
        # Generate prediction
        if prediction_score >= 2:
            direction = "UP"
            confidence = min(abs(prediction_score) * 10, 80)
            price_change = f"+{confidence/10:.1f}%"
        elif prediction_score <= -2:
            direction = "DOWN"
            confidence = min(abs(prediction_score) * 10, 80)
            price_change = f"-{confidence/10:.1f}%"
        else:
            direction = "SIDEWAYS"
            confidence = 50
            price_change = "±1%"
        
        analysis = {
            'symbol': symbol,
            'current_price': current_price,
            'prediction_direction': direction,
            'prediction_change': price_change,
            'confidence': confidence,
            'positive_factors': positive_factors[:4],
            'negative_factors': negative_factors[:4],
            'technical_indicators': {
                'rsi': recent_rsi,
                'sma_20': sma_20,
                'sma_50': sma_50,
                'week_change': week_change,
                'month_change': month_change
            },
            'news_count': len(news) if news else 0,
            'company_name': profile.get('name', 'N/A'),
            'industry': profile.get('finnhubIndustry', 'N/A'),
            'market_cap': profile.get('marketCapitalization', 0)
        }
        
        return analysis

def create_chart(symbol, stock_data):
    """Create candlestick chart with technical indicators"""
    try:
        # Prepare data for mplfinance
        df = stock_data.copy()
        df.index = pd.to_datetime(df.index)
        
        # Create the chart
        fig, axes = mpf.plot(df, type='candle', style='charles',
                           title=f'{symbol} Stock Price Analysis',
                           ylabel='Price ($)',
                           volume=True,
                           mav=(20, 50),
                           figsize=(12, 8),
                           returnfig=True)
        
        return fig
    except Exception as e:
        st.error(f"Error creating chart: {e}")
        return None

def main():
    """Main Streamlit app"""
    
    # Header
    st.title("📈 FinGPT-Forecaster")
    st.markdown("**AI-Powered Stock Market Prediction System**")
    st.markdown("---")
    
    # Sidebar
    st.sidebar.header("🔧 Configuration")
    
    # Stock symbol input
    symbol = st.sidebar.text_input(
        "Stock Symbol", 
        value="AAPL", 
        help="Enter a stock ticker symbol (e.g., AAPL, MSFT, NVDA)"
    ).upper()
    
    # Analysis period
    days_back = st.sidebar.slider(
        "Analysis Period (days)", 
        min_value=30, 
        max_value=365, 
        value=90,
        help="Number of days to look back for analysis"
    )
    
    # API Key input
    st.sidebar.subheader("🔑 API Configuration")
    finnhub_key = st.sidebar.text_input(
        "Finnhub API Key (Optional)", 
        type="password",
        help="Get your free API key from finnhub.io for enhanced data"
    )
    
    if finnhub_key:
        os.environ['FINNHUB_API_KEY'] = finnhub_key
    
    # Analyze button
    if st.sidebar.button("🚀 Analyze Stock", type="primary"):
        if not symbol:
            st.error("Please enter a stock symbol")
        else:
            # Initialize forecaster
            forecaster = MarketForecaster()
            
            # Perform analysis
            result = forecaster.analyze_stock_movement(symbol, days_back)
            
            if result:
                analysis, stock_data = result
                
                # Display results
                st.header(f"📊 Analysis Results for {symbol}")
                
                # Company info
                col1, col2, col3, col4 = st.columns(4)
                with col1:
                    st.metric("Company", analysis['company_name'])
                with col2:
                    st.metric("Industry", analysis['industry'])
                with col3:
                    st.metric("Current Price", f"${analysis['current_price']:.2f}")
                with col4:
                    st.metric("Market Cap", f"${analysis['market_cap']:,.0f}")
                
                # Prediction
                st.subheader("🎯 Prediction")
                col1, col2, col3 = st.columns(3)
                
                with col1:
                    direction_color = "🟢" if analysis['prediction_direction'] == "UP" else "🔴" if analysis['prediction_direction'] == "DOWN" else "🟡"
                    st.metric("Direction", f"{direction_color} {analysis['prediction_direction']}")
                
                with col2:
                    st.metric("Expected Change", analysis['prediction_change'])
                
                with col3:
                    st.metric("Confidence", f"{analysis['confidence']:.1f}%")
                
                # Technical indicators
                st.subheader("📈 Technical Indicators")
                tech = analysis['technical_indicators']
                
                col1, col2, col3, col4, col5 = st.columns(5)
                with col1:
                    st.metric("RSI", f"{tech['rsi']:.1f}")
                with col2:
                    st.metric("20-day SMA", f"${tech['sma_20']:.2f}")
                with col3:
                    st.metric("50-day SMA", f"${tech['sma_50']:.2f}")
                with col4:
                    st.metric("1-week Change", f"{tech['week_change']:+.2f}%")
                with col5:
                    st.metric("1-month Change", f"{tech['month_change']:+.2f}%")
                
                # Factors
                col1, col2 = st.columns(2)
                
                with col1:
                    st.subheader("✅ Positive Factors")
                    if analysis['positive_factors']:
                        for i, factor in enumerate(analysis['positive_factors'], 1):
                            st.write(f"{i}. {factor}")
                    else:
                        st.write("No significant positive factors identified")
                
                with col2:
                    st.subheader("⚠️ Potential Concerns")
                    if analysis['negative_factors']:
                        for i, factor in enumerate(analysis['negative_factors'], 1):
                            st.write(f"{i}. {factor}")
                    else:
                        st.write("No significant concerns identified")
                
                # Chart
                st.subheader("📊 Price Chart")
                fig = create_chart(symbol, stock_data)
                if fig:
                    st.pyplot(fig)
                
                # News summary
                st.subheader("📰 News Analysis")
                st.write(f"Analyzed {analysis['news_count']} recent news articles")
    
    # Footer
    st.markdown("---")
    st.markdown("""
    <div style='text-align: center; color: #666;'>
        <p><strong>Disclaimer:</strong> This analysis is for educational purposes only and should not be considered as financial advice.</p>
        <p>Powered by FinGPT-Forecaster | Built with Streamlit</p>
    </div>
    """, unsafe_allow_html=True)

if __name__ == "__main__":
    main()