add proxy tools use case

Author: qdriven

Diff for: agents/maigctester/.evn.example

@@ -0,0 +1,167 @@
+
+from langchain_core.messages import HumanMessage
+
+from agents.state import AgentState, show_agent_reasoning
+
+import json
+
+##### Fundamental Agent #####
+def fundamentals_agent(state: AgentState):
+    """Analyzes fundamental data and generates trading signals."""
+    show_reasoning = state["metadata"]["show_reasoning"]
+    data = state["data"]
+    metrics = data["financial_metrics"][0]
+    financial_line_item = data["financial_line_items"][0]
+    market_cap = data["market_cap"]
+
+    # Initialize signals list for different fundamental aspects
+    signals = []
+    reasoning = {}
+    
+    # 1. Profitability Analysis
+    profitability_score = 0
+    if metrics["return_on_equity"] > 0.15:  # Strong ROE above 15%
+        profitability_score += 1
+    if metrics["net_margin"] > 0.20:  # Healthy profit margins
+        profitability_score += 1
+    if metrics["operating_margin"] > 0.15:  # Strong operating efficiency
+        profitability_score += 1
+        
+    signals.append('bullish' if profitability_score >= 2 else 'bearish' if profitability_score == 0 else 'neutral')
+    reasoning["Profitability"] = {
+        "signal": signals[0],
+        "details": f"ROE: {metrics['return_on_equity']:.2%}, Net Margin: {metrics['net_margin']:.2%}, Op Margin: {metrics['operating_margin']:.2%}"
+    }
+    
+    # 2. Growth Analysis
+    growth_score = 0
+    if metrics["revenue_growth"] > 0.10:  # 10% revenue growth
+        growth_score += 1
+    if metrics["earnings_growth"] > 0.10:  # 10% earnings growth
+        growth_score += 1
+    if metrics["book_value_growth"] > 0.10:  # 10% book value growth
+        growth_score += 1
+        
+    signals.append('bullish' if growth_score >= 2 else 'bearish' if growth_score == 0 else 'neutral')
+    reasoning["Growth"] = {
+        "signal": signals[1],
+        "details": f"Revenue Growth: {metrics['revenue_growth']:.2%}, Earnings Growth: {metrics['earnings_growth']:.2%}"
+    }
+    
+    # 3. Financial Health
+    health_score = 0
+    if metrics["current_ratio"] > 1.5:  # Strong liquidity
+        health_score += 1
+    if metrics["debt_to_equity"] < 0.5:  # Conservative debt levels
+        health_score += 1
+    if metrics["free_cash_flow_per_share"] > metrics["earnings_per_share"] * 0.8:  # Strong FCF conversion
+        health_score += 1
+        
+    signals.append('bullish' if health_score >= 2 else 'bearish' if health_score == 0 else 'neutral')
+    reasoning["Financial_Health"] = {
+        "signal": signals[2],
+        "details": f"Current Ratio: {metrics['current_ratio']:.2f}, D/E: {metrics['debt_to_equity']:.2f}"
+    }
+    
+    # 4. Price to X ratios
+    pe_ratio = metrics["price_to_earnings_ratio"]
+    pb_ratio = metrics["price_to_book_ratio"]
+    ps_ratio = metrics["price_to_sales_ratio"]
+    
+    price_ratio_score = 0
+    if pe_ratio < 25:  # Reasonable P/E ratio
+        price_ratio_score += 1
+    if pb_ratio < 3:  # Reasonable P/B ratio
+        price_ratio_score += 1
+    if ps_ratio < 5:  # Reasonable P/S ratio
+        price_ratio_score += 1
+        
+    signals.append('bullish' if price_ratio_score >= 2 else 'bearish' if price_ratio_score == 0 else 'neutral')
+    reasoning["Price_Ratios"] = {
+        "signal": signals[3],
+        "details": f"P/E: {pe_ratio:.2f}, P/B: {pb_ratio:.2f}, P/S: {ps_ratio:.2f}"
+    }
+
+    # 5. Calculate intrinsic value and compare to market cap
+    free_cash_flow = financial_line_item.get('free_cash_flow')
+    intrinsic_value = calculate_intrinsic_value(
+        free_cash_flow=free_cash_flow,
+        growth_rate=metrics["earnings_growth"],
+        discount_rate=0.10,
+        terminal_growth_rate=0.03,
+        num_years=5,
+    )
+    if market_cap < intrinsic_value:
+        signals.append('bullish')
+    else:
+        signals.append('bearish')
+
+    reasoning["Intrinsic_Value"] = {
+        "signal": signals[4],
+        "details": f"Intrinsic Value: ${intrinsic_value:,.2f}, Market Cap: ${market_cap:,.2f}"
+    }
+    
+    # Determine overall signal
+    bullish_signals = signals.count('bullish')
+    bearish_signals = signals.count('bearish')
+    
+    if bullish_signals > bearish_signals:
+        overall_signal = 'bullish'
+    elif bearish_signals > bullish_signals:
+        overall_signal = 'bearish'
+    else:
+        overall_signal = 'neutral'
+    
+    # Calculate confidence level
+    total_signals = len(signals)
+    confidence = max(bullish_signals, bearish_signals) / total_signals
+    
+    message_content = {
+        "signal": overall_signal,
+        "confidence": f"{round(confidence * 100)}%",
+        "reasoning": reasoning
+    }
+    
+    # Create the fundamental analysis message
+    message = HumanMessage(
+        content=json.dumps(message_content),
+        name="fundamentals_agent",
+    )
+    
+    # Print the reasoning if the flag is set
+    if show_reasoning:
+        show_agent_reasoning(message_content, "Fundamental Analysis Agent")
+    
+    return {
+        "messages": [message],
+        "data": data,
+    }
+
+def calculate_intrinsic_value(
+    free_cash_flow: float,
+    growth_rate: float = 0.05,
+    discount_rate: float = 0.10,
+    terminal_growth_rate: float = 0.02,
+    num_years: int = 5,
+) -> float:
+    """
+    Computes the discounted cash flow (DCF) for a given company based on the current free cash flow.
+    Use this function to calculate the intrinsic value of a stock.
+    """
+    # Estimate the future cash flows based on the growth rate
+    cash_flows = [free_cash_flow * (1 + growth_rate) ** i for i in range(num_years)]
+
+    # Calculate the present value of projected cash flows
+    present_values = []
+    for i in range(num_years):
+        present_value = cash_flows[i] / (1 + discount_rate) ** (i + 1)
+        present_values.append(present_value)
+
+    # Calculate the terminal value
+    terminal_value = cash_flows[-1] * (1 + terminal_growth_rate) / (discount_rate - terminal_growth_rate)
+    terminal_present_value = terminal_value / (1 + discount_rate) ** num_years
+
+    # Sum up the present values and terminal value
+    dcf_value = sum(present_values) + terminal_present_value
+
+    return dcf_value

Diff for: agents/maigctester/more.md

@@ -0,0 +1,74 @@
+
+from langchain_openai.chat_models import ChatOpenAI
+
+from agents.state import AgentState
+from tools.api import search_line_items, get_financial_metrics, get_insider_trades, get_market_cap, get_prices
+
+from datetime import datetime
+
+llm = ChatOpenAI(model="gpt-4o")
+
+def market_data_agent(state: AgentState):
+    """Responsible for gathering and preprocessing market data"""
+    messages = state["messages"]
+    data = state["data"]
+
+    # Set default dates
+    end_date = data["end_date"] or datetime.now().strftime('%Y-%m-%d')
+    if not data["start_date"]:
+        # Calculate 3 months before end_date
+        end_date_obj = datetime.strptime(end_date, '%Y-%m-%d')
+        start_date = end_date_obj.replace(month=end_date_obj.month - 3) if end_date_obj.month > 3 else \
+            end_date_obj.replace(year=end_date_obj.year - 1, month=end_date_obj.month + 9)
+        start_date = start_date.strftime('%Y-%m-%d')
+    else:
+        start_date = data["start_date"]
+
+    # Get the historical price data
+    prices = get_prices(
+        ticker=data["ticker"], 
+        start_date=start_date, 
+        end_date=end_date,
+    )
+
+    # Get the financial metrics
+    financial_metrics = get_financial_metrics(
+        ticker=data["ticker"], 
+        report_period=end_date, 
+        period='ttm', 
+        limit=1,
+    )
+
+    # Get the insider trades
+    insider_trades = get_insider_trades(
+        ticker=data["ticker"], 
+        end_date=end_date,
+        limit=5,
+    )
+
+    # Get the market cap
+    market_cap = get_market_cap(
+        ticker=data["ticker"],
+    )
+
+    # Get the line_items
+    financial_line_items = search_line_items(
+        ticker=data["ticker"], 
+        line_items=["free_cash_flow"],
+        period='ttm',
+        limit=1,
+    )
+
+    return {
+        "messages": messages,
+        "data": {
+            **data, 
+            "prices": prices, 
+            "start_date": start_date, 
+            "end_date": end_date,
+            "financial_metrics": financial_metrics,
+            "insider_trades": insider_trades,
+            "market_cap": market_cap,
+            "financial_line_items": financial_line_items,
+        }
+    }

Diff for: agents/maigctester/src/maigctester/fi/README.md

@@ -0,0 +1,117 @@
+
+from langchain_core.messages import HumanMessage
+from langchain_core.prompts import ChatPromptTemplate
+from langchain_openai.chat_models import ChatOpenAI
+
+from agents.state import AgentState, show_agent_reasoning
+
+
+##### Portfolio Management Agent #####
+def portfolio_management_agent(state: AgentState):
+    """Makes final trading decisions and generates orders"""
+    show_reasoning = state["metadata"]["show_reasoning"]
+    portfolio = state["data"]["portfolio"]
+
+    # Get the technical analyst, fundamentals agent, and risk management agent messages
+    technical_message = next(msg for msg in state["messages"] if msg.name == "technical_analyst_agent")
+    fundamentals_message = next(msg for msg in state["messages"] if msg.name == "fundamentals_agent")
+    sentiment_message = next(msg for msg in state["messages"] if msg.name == "sentiment_agent")
+    risk_message = next(msg for msg in state["messages"] if msg.name == "risk_management_agent")
+
+    # Create the prompt template
+    template = ChatPromptTemplate.from_messages(
+        [
+            (
+                "system",
+                """You are a portfolio manager making final trading decisions.
+                Your job is to make a trading decision based on the team's analysis while strictly adhering
+                to risk management constraints.
+
+                RISK MANAGEMENT CONSTRAINTS:
+                - You MUST NOT exceed the max_position_size specified by the risk manager
+                - You MUST follow the trading_action (buy/sell/hold) recommended by risk management
+                - These are hard constraints that cannot be overridden by other signals
+
+                When weighing the different signals for direction and timing:
+                1. Fundamental Analysis (50% weight)
+                   - Primary driver of trading decisions
+                   - Should determine overall direction
+                
+                2. Technical Analysis (35% weight)
+                   - Secondary confirmation
+                   - Helps with entry/exit timing
+                
+                3. Sentiment Analysis (15% weight)
+                   - Final consideration
+                   - Can influence sizing within risk limits
+                
+                The decision process should be:
+                1. First check risk management constraints
+                2. Then evaluate fundamental outlook
+                3. Use technical analysis for timing
+                4. Consider sentiment for final adjustment
+                
+                Provide the following in your output:
+                - "action": "buy" | "sell" | "hold",
+                - "quantity": <positive integer>
+                - "confidence": <float between 0 and 1>
+                - "agent_signals": <list of agent signals including agent name, signal (bullish | bearish | neutral), and their confidence>
+                - "reasoning": <concise explanation of the decision including how you weighted the signals>
+
+                Trading Rules:
+                - Never exceed risk management position limits
+                - Only buy if you have available cash
+                - Only sell if you have shares to sell
+                - Quantity must be ≤ current position for sells
+                - Quantity must be ≤ max_position_size from risk management"""
+            ),
+            (
+                "human",
+                """Based on the team's analysis below, make your trading decision.
+
+                Technical Analysis Trading Signal: {technical_message}
+                Fundamental Analysis Trading Signal: {fundamentals_message}
+                Sentiment Analysis Trading Signal: {sentiment_message}
+                Risk Management Trading Signal: {risk_message}
+
+                Here is the current portfolio:
+                Portfolio:
+                Cash: {portfolio_cash}
+                Current Position: {portfolio_stock} shares
+
+                Only include the action, quantity, reasoning, confidence, and agent_signals in your output as JSON.  Do not include any JSON markdown.
+
+                Remember, the action must be either buy, sell, or hold.
+                You can only buy if you have available cash.
+                You can only sell if you have shares in the portfolio to sell.
+                """
+            ),
+        ]
+    )
+
+    # Generate the prompt
+    prompt = template.invoke(
+        {
+            "technical_message": technical_message.content, 
+            "fundamentals_message": fundamentals_message.content,
+            "sentiment_message": sentiment_message.content,
+            "risk_message": risk_message.content,
+            "portfolio_cash": f"{portfolio['cash']:.2f}",
+            "portfolio_stock": portfolio["stock"]
+        }
+    )
+    # Invoke the LLM
+    llm = ChatOpenAI(model="gpt-4o")
+    result = llm.invoke(prompt)
+
+    # Create the portfolio management message
+    message = HumanMessage(
+        content=result.content,
+        name="portfolio_management",
+    )
+
+    # Print the decision if the flag is set
+    if show_reasoning:
+        show_agent_reasoning(message.content, "Portfolio Management Agent")
+
+    return {"messages": state["messages"] + [message]}