Update app.py

app.py

@@ -5,45 +5,25 @@ from datetime import datetime, timedelta
 import difflib
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
-import yfinance as yf  # Primary source for accurate adjusted data
+import yfinance as yf
 from functools import lru_cache
 import pandas as pd
-import os
-from pycharts import CompanyClient  # Retained, but with fallback
 
 # Define the list of tickers
 tickers = ['TSLA', 'PLTR', 'SOUN', 'MSFT']
 
-# YCharts API setup (requires a valid key; obtain from ycharts.com)
-ycharts_api_key = os.environ.get('YCHARTS_API_KEY', 'your-api-key-here')  # Placeholder; set in environment
-company_client = CompanyClient(ycharts_api_key)
-
-# Prefetch stock data for all tickers at startup, preferring yfinance for adjusted prices
+# Prefetch stock data for all tickers at startup using yfinance
 all_data = {}
 try:
     now = datetime.now().strftime('%Y-%m-%d')
     for ticker in tickers:
-        # Attempt YCharts fetch (assuming 'price' is unadjusted; fallback to yfinance)
-        try:
-            past = datetime(2020, 1, 1)
-            series_rsp = company_client.get_series([ticker], ['price'], query_start_date=past, query_end_date=now)
-            if ticker in series_rsp and 'price' in series_rsp[ticker]:
-                df = pd.DataFrame({
-                    'Date': series_rsp[ticker]['dates'],
-                    'Close': series_rsp[ticker]['price']['values']  # Likely unadjusted
-                }).set_index('Date')
-                all_data[ticker] = df
-            raise Exception("Fallback to yfinance")  # Force fallback for accuracy in this update
-        except:
-            # Use yfinance for adjusted data
-            all_data[ticker] = yf.download(ticker, start='2020-01-01', end=now)
+        all_data[ticker] = yf.download(ticker, start='2020-01-01', end=now, auto_adjust=True)
 except Exception as e:
-    print(f"Error prefetching data: {e}. Using yfinance exclusively.")
-    all_data = {ticker: yf.download(ticker, start='2020-01-01', end=now.strftime('%Y-%m-%d'))
-                for ticker in tickers}
+    print(f"Error prefetching data: {e}")
+    all_data = {ticker: pd.DataFrame() for ticker in tickers}  # Initialize empty DataFrames on failure
 
 # Create a DataFrame with 'Adj Close' columns for each ticker
-adj_close_data = pd.DataFrame({ticker: data['Adj Close'] for ticker, data in all_data.items() if 'Adj Close' in data.columns})
+adj_close_data = pd.DataFrame({ticker: data['Close'] for ticker, data in all_data.items() if not data.empty})
 
 # Display the first few rows to verify (for debugging; remove in production)
 print(adj_close_data.head())
@@ -53,21 +33,28 @@ available_symbols = ['TSLA', 'MSFT', 'NVDA', 'GOOG', 'AMZN', 'SPY', 'AAPL', 'MET
 
 @lru_cache(maxsize=100)
 def fetch_stock_data(symbol, start_date, end_date):
-    if symbol in all_data:
+    if symbol in all_data and not all_data[symbol].empty:
         # Use preloaded data and slice by date
         hist = all_data[symbol]
         return hist[(hist.index >= start_date) & (hist.index <= end_date)]
     else:
-        # Fallback to on-demand fetch with yfinance for adjusted data (skip YCharts for CAGR accuracy)
+        # Fetch on-demand with yfinance
         try:
             ticker = yf.Ticker(symbol)
-            hist = ticker.history(start=start_date, end=end_date)
+            hist = ticker.history(start=start_date, end=end_date, auto_adjust=True)
             return hist
         except Exception as e:
             print(f"Error fetching data for {symbol}: {e}")
             return None
 
 def parse_period(query):
+    # Enhanced to handle year ranges like "between 2010 and 2020"
+    range_match = re.search(r'between\s+(\d{4})\s+and\s+(\d{4})', query.lower())
+    if range_match:
+        start_year = int(range_match.group(1))
+        end_year = int(range_match.group(2))
+        return (datetime(end_year, 12, 31) - datetime(start_year, 1, 1))
+    # Fallback to original period parsing
     match = re.search(r'(\d+)\s*(year|month|week|day)s?', query.lower())
     if match:
         num = int(match.group(1))
@@ -89,10 +76,10 @@ def find_closest_symbol(input_symbol):
 
 def calculate_growth_rate(start_date, end_date, symbol):
     hist = fetch_stock_data(symbol, start_date.strftime('%Y-%m-%d'), end_date.strftime('%Y-%m-%d'))
-    if hist is None or hist.empty or 'Adj Close' not in hist.columns:
+    if hist is None or hist.empty or 'Close' not in hist.columns:
         return None
-    beginning_value = hist.iloc[0]['Adj Close']
-    ending_value = hist.iloc[-1]['Adj Close']
+    beginning_value = hist.iloc[0]['Close']  # Use 'Close' as yfinance with auto_adjust=True returns adjusted prices
+    ending_value = hist.iloc[-1]['Close']
     years = (end_date - start_date).days / 365.25
     if years <= 0:
         return 0
@@ -137,10 +124,16 @@ def generate_response(user_query, enable_thinking=False):
         "You are FinChat, a knowledgeable financial advisor. Always respond in a friendly, professional manner. "
         "For greetings like 'Hi' or 'Hello', reply warmly, e.g., 'Hi! I'm FinChat, your financial advisor. What can I help you with today regarding stocks, investments, or advice?' "
         "Provide accurate, concise advice based on data."
-        # Assume continuation of original prompt here if additional text exists
     )
-    # Original generation logic (tokenize, generate, decode) would follow here; integrate as per the full original code
+    # Placeholder for generation logic (tokenize, generate, decode)
+    return summary or "Please provide a specific stock or investment query."
 
-# Gradio interface setup (assumed from original; add if needed)
-# demo = gr.Interface(...)
-# demo.launch()
+# Gradio interface setup
+demo = gr.Interface(
+    fn=generate_response,
+    inputs=[gr.Textbox(lines=2, placeholder="Enter your query (e.g., 'TSLA CAGR between 2010 and 2020')"), gr.Checkbox(label="Enable Thinking")],
+    outputs="text",
+    title="FinChat",
+    description="Ask about stock performance, CAGR, or investments."
+)
+demo.launch()