TradingAgents/tradingagents/agents/managers/risk_manager.py

105 lines
4.3 KiB
Python

import time
import json
from langchain_core.messages import SystemMessage, HumanMessage
from tradingagents.log_utils import add_log, step_timer, symbol_progress
def create_risk_manager(llm, memory):
def risk_manager_node(state) -> dict:
company_name = state["company_of_interest"]
history = state["risk_debate_state"]["history"]
risk_debate_state = state["risk_debate_state"]
market_research_report = state["market_report"]
news_report = state["news_report"]
fundamentals_report = state["news_report"]
sentiment_report = state["sentiment_report"]
trader_plan = state["investment_plan"]
curr_situation = f"{market_research_report}\n\n{sentiment_report}\n\n{news_report}\n\n{fundamentals_report}"
past_memory_str = ""
if memory is not None:
past_memories = memory.get_memories(curr_situation, n_matches=2)
for i, rec in enumerate(past_memories, 1):
past_memory_str += rec["recommendation"] + "\n\n"
system_prompt = """You are a Risk Manager at a financial advisory firm making the final investment decision. You MUST stay in character as a financial professional at all times.
CRITICAL RULES:
- NEVER mention that you are an AI, Claude, a language model, or an assistant
- NEVER offer to help with code, software, or implementation tasks
- NEVER say "I don't have access to" or "I can't see the data" — analyze whatever data is provided below
- If data sections are empty, state that data is unavailable and make a decision based on available information
Your task: Evaluate the risk debate between Aggressive, Neutral, and Conservative analysts.
Your response must include:
1. FINAL DECISION: BUY, SELL, or HOLD
2. HOLD_DAYS: Number of trading days to hold the position before exiting (for BUY/HOLD only, write N/A for SELL)
3. RISK ASSESSMENT: Summary of key risks identified
4. RATIONALE: Why this decision balances risk and reward appropriately
RESPONSE FORMAT:
- Maximum 1500 characters. Lead with your decision, then key rationale.
- Complete your ENTIRE response in a SINGLE message.
Respond only with your analysis and decision. No disclaimers or meta-commentary."""
user_prompt = f"""Make the final risk-adjusted investment decision:
COMPANY: {company_name}
ORIGINAL TRADER PLAN:
{trader_plan}
RISK ANALYSTS DEBATE:
{history}
PAST LEARNINGS:
{past_memory_str if past_memory_str else "None"}
Based on the risk analysis above, what is your final investment decision?"""
messages = [
SystemMessage(content=system_prompt),
HumanMessage(content=user_prompt)
]
step_timer.start_step("risk_manager")
add_log("agent", "risk_manager", f"🛡️ Risk Manager making final decision for {company_name}...")
t0 = time.time()
response = llm.invoke(messages)
elapsed = time.time() - t0
add_log("llm", "risk_manager", f"LLM responded in {elapsed:.1f}s ({len(response.content)} chars)")
add_log("agent", "risk_manager", f"✅ Final decision: {response.content[:300]}...")
step_timer.end_step("risk_manager", "completed", response.content[:200])
symbol_progress.step_done(company_name, "risk_manager")
step_timer.set_details("risk_manager", {
"system_prompt": system_prompt,
"user_prompt": user_prompt[:3000],
"response": response.content[:3000],
"tool_calls": [],
})
new_risk_debate_state = {
"judge_decision": response.content,
"history": risk_debate_state["history"],
"risky_history": risk_debate_state["risky_history"],
"safe_history": risk_debate_state["safe_history"],
"neutral_history": risk_debate_state["neutral_history"],
"latest_speaker": "Judge",
"current_risky_response": risk_debate_state["current_risky_response"],
"current_safe_response": risk_debate_state["current_safe_response"],
"current_neutral_response": risk_debate_state["current_neutral_response"],
"count": risk_debate_state["count"],
}
return {
"risk_debate_state": new_risk_debate_state,
"final_trade_decision": response.content,
}
return risk_manager_node