IB Analytics MCP Server

""" Sentiment analysis module for stock market analysis Provides multi-source sentiment analysis combining news, options market data, and technical indicators to gauge overall market sentiment for stocks and ETFs. Architecture ------------ The module uses a strategy pattern with pluggable sentiment analyzers: 1. **BaseSentimentAnalyzer**: Abstract base class defining the interface 2. **NewsSentimentAnalyzer**: Sentiment from news articles and headlines 3. **OptionsSentimentAnalyzer**: Sentiment from Put/Call ratios, IV, and Max Pain 4. **TechnicalSentimentAnalyzer**: Sentiment from RSI, MACD, trends, and levels 5. **CompositeSentimentAnalyzer**: Weighted aggregation of multiple sources Usage ----- Basic Usage (via MCP tools): >>> from ib_sec_mcp.mcp.tools.sentiment import analyze_market_sentiment >>> # Get composite sentiment from all sources >>> result = await analyze_market_sentiment(symbol="AAPL", sources="composite") >>> # Get technical sentiment only >>> result = await analyze_market_sentiment(symbol="SPY", sources="technical") Programmatic Usage: >>> from ib_sec_mcp.analyzers.sentiment import ( ... NewsSentimentAnalyzer, ... OptionsSentimentAnalyzer, ... TechnicalSentimentAnalyzer, ... CompositeSentimentAnalyzer, ... ) >>> from decimal import Decimal >>> >>> # Create individual analyzers >>> news_analyzer = NewsSentimentAnalyzer() >>> options_analyzer = OptionsSentimentAnalyzer() >>> technical_analyzer = TechnicalSentimentAnalyzer() >>> >>> # Create composite with custom weights >>> composite = CompositeSentimentAnalyzer( ... sources={ ... "news": news_analyzer, ... "options": options_analyzer, ... "technical": technical_analyzer, ... }, ... weights={ ... "news": Decimal("0.4"), # 40% weight ... "options": Decimal("0.3"), # 30% weight ... "technical": Decimal("0.3"), # 30% weight ... }, ... ) >>> >>> # Analyze sentiment >>> result = await composite.analyze_sentiment(symbol="AAPL") >>> print(f"Score: {result.score}, Confidence: {result.confidence}") Sentiment Score Interpretation ------------------------------ Scores range from -1.0 (very bearish) to +1.0 (very bullish): - 0.5 to 1.0: Very Bullish (strong buy signals) - 0.2 to 0.5: Bullish (buy bias with good conviction) - -0.2 to 0.2: Neutral (mixed signals, wait for clarity) - -0.5 to -0.2: Bearish (sell bias) - -1.0 to -0.5: Very Bearish (strong sell signals) Confidence scores range from 0.0 to 1.0: - 0.7-1.0: High confidence (multiple sources agree) - 0.4-0.7: Medium confidence (some agreement) - 0.0-0.4: Low confidence (sources disagree) Components ---------- News Sentiment (40% default weight): - Headlines and article sentiment analysis - Publisher credibility weighting - Time decay for article recency Options Market Sentiment (30% default weight): - Put/Call ratio: <0.7 bullish, >1.3 bearish - IV Rank: High IV favors selling premium - Max Pain: Price gravitation toward max pain strike Technical Sentiment (30% default weight): - RSI: <30 oversold (bullish), >70 overbought (bearish) - MACD: Bullish/bearish crossovers - Trend: Multi-timeframe trend alignment - Support/Resistance: Price proximity to key levels Integration ----------- Sentiment analysis is integrated into: - MCP tools: `analyze_market_sentiment` function - /analyze-symbol command: Automatic sentiment inclusion - Claude Desktop: Natural language queries Examples -------- See examples/sentiment_analysis_example.py for comprehensive usage examples. See Also -------- - README.md section "Market Sentiment Analysis" for complete documentation - .claude/commands/analyze-symbol.md for integration with symbol analysis """ from ib_sec_mcp.analyzers.sentiment.base import BaseSentimentAnalyzer, SentimentScore from ib_sec_mcp.analyzers.sentiment.composite import CompositeSentimentAnalyzer from ib_sec_mcp.analyzers.sentiment.news import NewsSentimentAnalyzer from ib_sec_mcp.analyzers.sentiment.options import OptionsSentimentAnalyzer from ib_sec_mcp.analyzers.sentiment.technical import TechnicalSentimentAnalyzer __all__ = [ "BaseSentimentAnalyzer", "SentimentScore", "NewsSentimentAnalyzer", "OptionsSentimentAnalyzer", "TechnicalSentimentAnalyzer", "CompositeSentimentAnalyzer", ]