Update app.py

app.py

@@ -7,32 +7,57 @@ import yfinance as yf
 from datetime import datetime, timedelta
 import random
 import gradio as gr
-import io
-import base64
 
-# Keep all the existing functions (fetch_data, calculate_beta, calculate_r_squared, 
-# determine_optimal_clusters, align_data, risk_level) as they are
-
-# Modify the main function to work with Gradio
-def analyze_stocks(stocks_filepath, index_symbol, years=5, progress=gr.Progress()):
+def fetch_data(symbol, start_date, end_date):
+    return yf.download(symbol, start=start_date, end=end_date)
+
+def calculate_beta(stock_returns, market_returns):
+    if len(stock_returns) < 2 or len(market_returns) < 2:
+        return np.nan, np.nan, np.nan
+    covariance_matrix = np.cov(stock_returns, market_returns)
+    beta = covariance_matrix[0, 1] / covariance_matrix[1, 1]
+    return beta, covariance_matrix[0, 1], covariance_matrix[1, 1]
+
+def calculate_r_squared(stock_returns, market_returns):
+    if len(stock_returns) < 2 or len(market_returns) < 2:
+        return np.nan
+    correlation_matrix = np.corrcoef(stock_returns, market_returns)
+    correlation_xy = correlation_matrix[0, 1]
+    r_squared = correlation_xy ** 2
+    return r_squared
+
+def align_data(stock_data, index_data):
+    aligned_data = stock_data.join(index_data, how='inner', lsuffix='_stock', rsuffix='_index')
+    return aligned_data
+
+def risk_level(beta):
+    if beta < 0.5:
+        return "Very Low Risk"
+    elif beta < 1:
+        return "Low Risk"
+    elif beta < 1.5:
+        return "Moderate Risk"
+    elif beta < 2:
+        return "High Risk"
+    else:
+        return "Very High Risk"
+
+def analyze_stocks(stocks_filepath, index_symbol, years=5):
     end_date = datetime.now()
     start_date = end_date - timedelta(days=365 * years)
 
     stock_symbols_df = pd.read_csv(stocks_filepath)
     stock_symbols = stock_symbols_df['Symbol'].tolist()
 
-    progress(0, desc="Fetching market index data")
     index_data = fetch_data(index_symbol, start_date, end_date)
     index_data = index_data['Close'].to_frame(name='Close_index')
 
     betas = {}
     r_squared_values = {}
     latest_close_values = {}
-    symbols = {}
     valid_stocks_count = 0
 
-    for i, symbol in enumerate(stock_symbols):
-        progress((i + 1) / len(stock_symbols), desc=f"Processing {symbol}")
+    for symbol in stock_symbols:
         stock_data = fetch_data(symbol, start_date, end_date)
         stock_data = stock_data['Close'].to_frame(name='Close_stock')
 
@@ -48,7 +73,6 @@ def analyze_stocks(stocks_filepath, index_symbol, years=5, progress=gr.Progress(
                     betas[symbol] = round(beta, 3)
                     r_squared_values[symbol] = round(calculate_r_squared(aligned_data['Close_stock'].dropna(), aligned_data['Close_index'].dropna()), 3)
                     latest_close_values[symbol] = round(stock_data['Close_stock'].iloc[-1], 3)
-                    symbols[symbol] = symbol
                     valid_stocks_count += 1
 
     results_df = pd.DataFrame({
@@ -65,15 +89,13 @@ def analyze_stocks(stocks_filepath, index_symbol, years=5, progress=gr.Progress(
     scaler = StandardScaler()
     features = scaler.fit_transform(features)
 
-    progress(0.9, desc="Determining optimal clusters")
-    optimal_clusters = determine_optimal_clusters(features)
+    optimal_clusters = 5  # You can implement a method to determine this dynamically
 
     gmm = GaussianMixture(n_components=optimal_clusters, random_state=42)
     cluster_labels = gmm.fit_predict(features)
     results_df['Cluster'] = cluster_labels
     results_df['Risk Level'] = results_df['Beta'].apply(risk_level)
 
-    progress(0.95, desc="Generating plot")
     custom_colors = [f'rgb({random.random()}, {random.random()}, {random.random()})' for _ in range(optimal_clusters)]
 
     traces = []
@@ -84,9 +106,9 @@ def analyze_stocks(stocks_filepath, index_symbol, years=5, progress=gr.Progress(
                            mode='markers', marker=dict(size=cluster_df['Latest Close'],
                                                        sizeref=2. * max(cluster_df['Latest Close']) / (60. ** 2),
                                                        sizemode='area'),
-                           hovertext=cluster_df['Symbol'] + '<br>' + cluster_df['Name'] + '<br>Beta: ' + cluster_df[
-                               'Beta'].astype(str) + '<br>R-Squared: ' + cluster_df['R-Squared'].astype(
-                               str) + '<br>Latest Close Price: ' + cluster_df['Latest Close'].astype(str) +
+                           hovertext=cluster_df['Symbol'] + '<br>' + cluster_df['Name'] + '<br>Beta: ' + cluster_df['Beta'].astype(str) +
+                                     '<br>R-Squared: ' + cluster_df['R-Squared'].astype(str) +
+                                     '<br>Latest Close Price: ' + cluster_df['Latest Close'].astype(str) +
                                      '<br>Risk Level: ' + cluster_df['Risk Level'] +
                                      '<br>Cluster Prob: ' + cluster_probs[cluster_df.index, cluster].round(3).astype(str),
                            showlegend=False,
@@ -109,10 +131,8 @@ def analyze_stocks(stocks_filepath, index_symbol, years=5, progress=gr.Progress(
                        text=f"Valid Stocks: {valid_stocks_count} | Date: {latest_date}",
                        showarrow=False, font=dict(size=14, color="black"))
 
-    progress(1.0, desc="Analysis complete")
     return fig
 
-# Gradio interface
 def gradio_interface(years):
     stocks_filepath = "SP500_stock_list_Jan-1-2024.csv"
     index_symbol = "^GSPC"