MonteWalk

Instructions

Implementation Reference

• tools/feature_engineering.py:57-172 (handler)

  Core handler function implementing get_technical_summary. Downloads 2y stock data via yfinance, computes RSI(14), MACD, SMA50/200. Applies scoring: RSI<30(+1)/>70(-1), MACD>signal(+1), price>SMA50(+1), price>SMA200(+1). Determines signal (STRONG BUY if score>=2 etc.) and returns formatted summary with reasons.
  def get_technical_summary(symbol: str) -> str: """ Performs a technical analysis summary (RSI, MACD, Moving Averages). Returns a 'Buy', 'Sell', or 'Neutral' signal based on aggregated indicators. """ try: # Need enough data for 200 SMA df = yf.download(symbol, period="2y", progress=False) if df.empty: return f"No data found for {symbol}" # Handle MultiIndex if present # yfinance structure: columns are (Price, Ticker) if isinstance(df.columns, pd.MultiIndex): # If we have a MultiIndex, we want the 'Close' level, and then the specific ticker # But usually df['Close'] gives us the Close column(s). pass close = df['Close'] if isinstance(close, pd.DataFrame): # If multiple tickers or just weird structure, take the first column close = close.iloc[:, 0] # 1. RSI rsi = ta.rsi(close, length=14) if rsi is None or rsi.empty: return f"Not enough data to calculate RSI for {symbol}" current_rsi = rsi.iloc[-1] # 2. MACD macd = ta.macd(close) if macd is None or macd.empty: return f"Not enough data to calculate MACD for {symbol}" # MACD_12_26_9, MACDh_12_26_9 (hist), MACDs_12_26_9 (signal) # Column names depend on pandas_ta version/settings, but usually standard macd_line = macd.iloc[-1, 0] # First column is usually MACD line macd_signal = macd.iloc[-1, 2] # Third column is usually Signal line # 3. Moving Averages sma_50 = ta.sma(close, length=50) sma_200 = ta.sma(close, length=200) if sma_50 is None or sma_50.empty: return "Not enough data for SMA 50" if sma_200 is None or sma_200.empty: # Fallback if 200 SMA not available (e.g. IPO recently) sma_200_val = 0 has_200 = False else: sma_200_val = sma_200.iloc[-1] has_200 = True sma_50_val = sma_50.iloc[-1] current_price = close.iloc[-1] # Scoring Logic score = 0 reasons = [] # RSI Logic if current_rsi < 30: score += 1 reasons.append(f"RSI is Oversold ({current_rsi:.2f})") elif current_rsi > 70: score -= 1 reasons.append(f"RSI is Overbought ({current_rsi:.2f})") else: reasons.append(f"RSI is Neutral ({current_rsi:.2f})") # MACD Logic if macd_line > macd_signal: score += 1 reasons.append("MACD Bullish Crossover") else: score -= 1 reasons.append("MACD Bearish Crossover") # Trend Logic if current_price > sma_50_val: score += 1 reasons.append("Price above 50 SMA (Bullish Trend)") else: score -= 1 reasons.append("Price below 50 SMA (Bearish Trend)") if has_200: if current_price > sma_200_val: score += 1 reasons.append("Price above 200 SMA (Long-term Bullish)") else: score -= 1 reasons.append("Price below 200 SMA (Long-term Bearish)") else: reasons.append("200 SMA not available (Insufficient history)") # Final Verdict if score >= 2: signal = "STRONG BUY" elif score == 1: signal = "BUY" elif score == 0: signal = "NEUTRAL" elif score == -1: signal = "SELL" else: signal = "STRONG SELL" return (f"Technical Summary for {symbol}:\n" f"Signal: {signal} (Score: {score})\n" f"Price: ${current_price:.2f}\n" f"--------------------------------\n" f"Analysis:\n" + "\n".join([f"- {r}" for r in reasons])) except Exception as e: return f"Error performing technical analysis: {str(e)}"
• server.py:390-393 (registration)

  MCP tool registration call that registers get_technical_summary (imported from tools.feature_engineering) using the register_tools helper, which applies @mcp.tool() decorator to each function in the list.
  register_tools( [compute_indicators, rolling_stats, get_technical_summary], "Feature Engineering" )
• server.py:16-16 (registration)

  Import statement bringing get_technical_summary into server.py scope for registration as MCP tool.
  from tools.feature_engineering import compute_indicators, rolling_stats, get_technical_summary