StockSight / app.py
Toyman's picture
User Deepseek R1 for Sentimental analysis
31f5268 verified
import gradio as gr
from huggingface_hub import InferenceClient
from GoogleNews import GoogleNews
import logging
import warnings
import textwrap
from tabulate import tabulate
import yfinance as yf
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import requests
from fuzzywuzzy import process
import re
# Suppress warnings
warnings.filterwarnings("ignore", category=UserWarning, module="fuzzywuzzy")
# Set up logging
logging.basicConfig(
level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s"
)
class FinancialAnalyzer:
def __init__(self):
# Load the DeepSeek model directly from Hugging Face Hub
self.client = InferenceClient("deepseek-ai/DeepSeek-R1-Distill-Qwen-32B")
self.ta_config = {
'rsi_window': 14,
'macd_fast': 12,
'macd_slow': 26,
'macd_signal': 9,
'bollinger_window': 20,
'sma_windows': [20, 50, 200],
'ema_windows': [12, 26],
'volatility_window': 30
}
logging.info("Initialized Financial Analyzer")
def resolve_ticker_symbol(self, query: str) -> str:
"""Convert company names to valid Yahoo Finance tickers"""
logging.info(f"Resolving ticker symbol for query: {query}")
url = "https://query2.finance.yahoo.com/v1/finance/search"
headers = {"User-Agent": "Mozilla/5.0"}
params = {"q": query, "quotesCount": 5, "country": "India"}
try:
response = requests.get(url, headers=headers, params=params, timeout=10)
response.raise_for_status()
data = response.json()
if not data.get("quotes"):
raise ValueError(f"No ticker found for: {query}")
quotes = data["quotes"]
names = [quote.get("longname") or quote.get("shortname", "") for quote in quotes]
best_match, score = process.extractOne(query, names)
if not best_match or score < 60:
raise ValueError(f"No matching ticker found for: {query}")
index = names.index(best_match)
best_quote = quotes[index]
resolved_ticker = best_quote["symbol"]
exchange_code = best_quote.get("exchange", "").upper()
exchange_suffix_map = {
"NSI": ".NS", # NSE
"BOM": ".BO", # BSE
"BSE": ".BO",
"NSE": ".NS",
}
suffix = exchange_suffix_map.get(exchange_code, ".NS")
if not resolved_ticker.endswith(suffix):
resolved_ticker += suffix
logging.info(f"Resolved ticker symbol: {resolved_ticker}")
return resolved_ticker
except Exception as e:
logging.error(f"Ticker resolution failed: {str(e)}")
raise
def fetch_stock_data(self, ticker):
"""Fetch historical data and technical indicators"""
logging.info(f"Fetching stock data for ticker: {ticker}")
try:
stock = yf.Ticker(ticker)
history = stock.history(period="1y", interval="1d")
if history.empty:
logging.error(f"No data found for {ticker}")
return {"error": f"No data found for {ticker}"}
logging.info(f"Successfully fetched stock data for {ticker}")
return {
'history': history,
'current_price': history['Close'].iloc[-1],
'indicators': self.calculate_technical_indicators(history),
'info': stock.info
}
except Exception as e:
logging.error(f"Error fetching stock data: {str(e)}")
return {"error": str(e)}
def calculate_technical_indicators(self, history):
"""Calculate technical analysis metrics"""
logging.info("Calculating technical indicators")
ta = {}
# RSI
delta = history['Close'].diff()
gain = delta.where(delta > 0, 0)
loss = -delta.where(delta < 0, 0)
avg_gain = gain.rolling(self.ta_config['rsi_window']).mean()
avg_loss = loss.rolling(self.ta_config['rsi_window']).mean()
rs = avg_gain / avg_loss
ta['rsi'] = 100 - (100 / (1 + rs)).iloc[-1]
# MACD
ema_fast = history['Close'].ewm(span=self.ta_config['macd_fast'], adjust=False).mean()
ema_slow = history['Close'].ewm(span=self.ta_config['macd_slow'], adjust=False).mean()
macd = ema_fast - ema_slow
signal = macd.ewm(span=self.ta_config['macd_signal'], adjust=False).mean()
ta['macd'] = macd.iloc[-1]
ta['macd_signal'] = signal.iloc[-1]
# Bollinger Bands
sma = history['Close'].rolling(self.ta_config['bollinger_window']).mean()
std = history['Close'].rolling(self.ta_config['bollinger_window']).std()
ta['bollinger_upper'] = (sma + 2 * std).iloc[-1]
ta['bollinger_lower'] = (sma - 2 * std).iloc[-1]
logging.info("Technical indicators calculated")
return ta
def generate_price_chart(self, history):
"""Generate interactive price chart"""
logging.info("Generating price chart")
try:
fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(10, 6), sharex=True)
# Price plot
history['Close'].plot(ax=ax1, label='Price')
ax1.set_title('Price Trend')
ax1.legend()
# Volume plot
history['Volume'].plot(ax=ax2, kind='bar', color='skyblue')
ax2.set_title('Trading Volume')
plt.tight_layout()
logging.info("Price chart generated")
return fig
except Exception as e:
logging.error(f"Chart generation failed: {str(e)}")
return self.create_error_plot("Chart unavailable")
def create_error_plot(self, message):
"""Create a placeholder plot for error messages"""
fig, ax = plt.subplots(figsize=(10, 2))
ax.text(0.5, 0.5, message,
ha='center', va='center',
fontsize=12, color='red')
ax.axis('off')
return fig
def fetch_articles(self, query):
"""Fetch news articles from Google News"""
logging.info(f"Fetching news articles for query: {query}")
try:
googlenews = GoogleNews(lang="en")
googlenews.search(query)
articles = googlenews.result()
logging.info(f"Fetched {len(articles)} news articles")
return articles[:5] # Limit to 5 articles
except Exception as e:
logging.error(f"Error fetching articles: {str(e)}")
return []
def analyze_article_sentiment(self, article):
"""Analyze article sentiment using DeepSeek model with improved parsing"""
logging.info(f"Analyzing sentiment for article: {article['title']}")
prompt = f"""
Analyze the sentiment and provide a brief analysis of this news article about a financial asset.
Respond EXACTLY in this format:
SENTIMENT: [POSITIVE/NEGATIVE/NEUTRAL]
ANALYSIS: [2-3 sentence analysis]
Title: {article['title']}
Description: {article['desc']}
"""
try:
response = self.client.chat.completions.create(
model="deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
messages=[{"role": "user", "content": prompt}],
temperature=0.1,
max_tokens=150
)
response_text = response.choices[0].message.content.strip()
# Improved parsing using regular expressions
sentiment_match = re.search(r"SENTIMENT:\s*(POSITIVE|NEGATIVE|NEUTRAL)", response_text, re.IGNORECASE)
analysis_match = re.search(r"ANALYSIS:\s*(.+)$", response_text, re.DOTALL)
sentiment = "neutral" # Default value
if sentiment_match:
sentiment = sentiment_match.group(1).lower()
else:
logging.warning(f"Failed to parse sentiment from response: {response_text}")
analysis = "Sentiment analysis unavailable"
if analysis_match:
analysis = analysis_match.group(1).strip()
# Validate sentiment value
if sentiment not in ['positive', 'negative', 'neutral']:
sentiment = 'neutral'
logging.warning(f"Invalid sentiment value: {sentiment}")
logging.info(f"Sentiment analysis complete: {sentiment}")
return {
**article,
"sentiment": sentiment,
"analysis": analysis
}
except Exception as e:
logging.error(f"Sentiment analysis failed: {str(e)}")
return {
**article,
"sentiment": "neutral",
"analysis": "Sentiment analysis failed"
}
def generate_recommendation(self, articles, stock_data):
"""Generate investment recommendation with fallback values"""
logging.info("Generating investment recommendation")
# Initialize sentiment scores with default values
sentiment_scores = {
'positive': 0,
'negative': 0,
'neutral': 0
}
for article in articles:
sentiment = article.get('sentiment', 'neutral')
if sentiment in sentiment_scores:
sentiment_scores[sentiment] += 1
# Technical analysis with fallback values
ta = stock_data.get('indicators', {})
price_change = stock_data['history']['Close'].pct_change().iloc[-1] if not stock_data['history'].empty else 0
# Recommendation logic with safeguards
recommendation = "HOLD"
reasons = []
try:
rsi = ta.get('rsi', 50)
if rsi < 30 and sentiment_scores['positive'] > sentiment_scores['negative']:
recommendation = "BUY"
reasons.append("Oversold condition with positive news sentiment")
elif rsi > 70 and sentiment_scores['negative'] > sentiment_scores['positive']:
recommendation = "SELL"
reasons.append("Overbought condition with negative news sentiment")
elif price_change > 0.05 and sentiment_scores['positive'] > 3:
recommendation = "STRONG BUY"
reasons.append("Strong positive momentum and news sentiment")
elif price_change < -0.05 and sentiment_scores['negative'] > 3:
recommendation = "STRONG SELL"
reasons.append("Significant downward pressure and negative news")
except Exception as e:
logging.error(f"Recommendation logic failed: {str(e)}")
recommendation = "HOLD"
reasons.append("Analysis incomplete due to data issues")
logging.info(f"Recommendation generated: {recommendation}")
return {
"recommendation": recommendation,
"reasons": reasons,
"sentiment_distribution": sentiment_scores,
"technical_indicators": ta
}
def format_analysis_output(analyzer, articles, stock_data, recommendation):
"""Format all analysis components for display with error handling"""
logging.info("Formatting analysis output")
try:
# News table
news_table = []
for article in articles:
news_table.append([
article.get('date', 'N/A'),
textwrap.fill(article.get('title', 'No title'), 40),
textwrap.fill(article.get('analysis', 'No analysis'), 60),
"🟒" if article.get('sentiment') == 'positive' else "πŸ”΄" if article.get('sentiment') == 'negative' else "βšͺ"
])
# Stock info with fallback values
info = stock_data.get('info', {})
stock_info = f"""
<div style="padding: 20px; background: #f8f9fa; border-radius: 10px;">
<h3>{info.get('longName', 'N/A')} ({info.get('symbol', 'N/A')})</h3>
<p>Price: ${stock_data.get('current_price', 0):.2f}</p>
<p>Market Cap: {info.get('marketCap', 'N/A')}</p>
<p>PE Ratio: {info.get('trailingPE', 'N/A')}</p>
</div>
"""
# Recommendation styling
rec_style = {
"BUY": ("#d4edda", "🟒"),
"STRONG BUY": ("#d4edda", "🟒"),
"SELL": ("#f8d7da", "πŸ”΄"),
"STRONG SELL": ("#f8d7da", "πŸ”΄"),
"HOLD": ("#fff3cd", "βšͺ")
}.get(recommendation['recommendation'].split()[0], ("#ffffff", "βšͺ"))
rec_html = f"""
<div style="padding: 20px; background: {rec_style[0]}; border-radius: 10px;">
<h2>{rec_style[1]} Recommendation: {recommendation['recommendation']}</h2>
<ul>
{"".join(f'<li>{reason}</li>' for reason in recommendation.get('reasons', ['No analysis available']))}
</ul>
</div>
"""
# Generate chart
chart = analyzer.generate_price_chart(stock_data['history'])
return {
"news_table": tabulate(news_table, headers=["Date", "Title", "Analysis", "Sentiment"], tablefmt="html"),
"stock_info": stock_info,
"recommendation": rec_html,
"chart": chart
}
except Exception as e:
logging.error(f"Formatting failed: {str(e)}")
return {
"error": f"Output formatting failed: {str(e)}"
}
def analyze_asset(asset_input):
logging.info(f"Analyzing asset: {asset_input}")
analyzer = FinancialAnalyzer()
try:
# Resolve ticker symbol
ticker = analyzer.resolve_ticker_symbol(asset_input)
# Fetch data
stock_data = analyzer.fetch_stock_data(ticker)
if 'error' in stock_data:
raise ValueError(stock_data['error'])
articles = analyzer.fetch_articles(asset_input)
analyzed_articles = [analyzer.analyze_article_sentiment(a) for a in articles]
# Generate recommendation
recommendation = analyzer.generate_recommendation(analyzed_articles, stock_data)
# Format output
results = format_analysis_output(analyzer, analyzed_articles, stock_data, recommendation)
logging.info(f"Analysis complete for asset: {asset_input}")
return results
except Exception as e:
logging.error(f"Analysis failed: {str(e)}")
return {"error": f"Analysis failed: {str(e)}"}
def main():
with gr.Blocks(theme=gr.themes.Default()) as app:
gr.Markdown("# Advanced Stock Analysis Suite")
with gr.Row():
asset_input = gr.Textbox(label="Stock/Company Name", placeholder="Enter stock name or symbol...")
analyze_btn = gr.Button("Analyze", variant="primary")
with gr.Tabs():
with gr.Tab("News Sentiment"):
news_table = gr.HTML(label="News Analysis")
with gr.Tab("Technical Analysis"):
stock_info = gr.HTML()
price_chart = gr.Plot()
with gr.Tab("Recommendation"):
recommendation = gr.HTML()
@analyze_btn.click(inputs=[asset_input], outputs=[news_table, stock_info, price_chart, recommendation])
def update_analysis(asset):
logging.info(f"Update analysis triggered for asset: {asset}")
results = analyze_asset(asset)
if 'error' in results:
logging.error(f"Error in analysis: {results['error']}")
return [f"<div style='color: red'>{results['error']}</div>"]*4
logging.info(f"Analysis results returned for asset: {asset}")
return [
results["news_table"],
results["stock_info"],
results["chart"],
results["recommendation"]
]
app.launch()
if __name__ == "__main__":
main()