TradingAgents/tradingagents/agents/managers/research_manager.py

97 lines
4.4 KiB
Python

import time
import json
from tradingagents.agents.utils.agent_utils import (
build_debate_brief,
build_instrument_context,
extract_feedback_snapshot,
get_language_instruction,
get_snapshot_template,
get_snapshot_writing_instruction,
localize_label,
localize_rating_term,
localize_role_name,
truncate_for_prompt,
)
def create_research_manager(llm, memory):
def research_manager_node(state) -> dict:
instrument_context = build_instrument_context(state["company_of_interest"])
market_research_report = truncate_for_prompt(state["market_report"])
sentiment_report = truncate_for_prompt(state["sentiment_report"])
news_report = truncate_for_prompt(state["news_report"])
fundamentals_report = truncate_for_prompt(state["fundamentals_report"])
investment_debate_state = state["investment_debate_state"]
bull_snapshot = investment_debate_state.get("bull_snapshot", "")
bear_snapshot = investment_debate_state.get("bear_snapshot", "")
debate_brief = investment_debate_state.get("debate_brief", "")
curr_situation = f"{market_research_report}\n\n{sentiment_report}\n\n{news_report}\n\n{fundamentals_report}"
past_memories = memory.get_memories(curr_situation, n_matches=2)
past_memory_str = ""
for i, rec in enumerate(past_memories, 1):
past_memory_str += rec["recommendation"] + "\n\n"
prompt = f"""As the portfolio manager and debate facilitator, your role is to critically evaluate this round of debate and make a definitive decision: align with the {localize_role_name("Bear Analyst")}, the {localize_role_name("Bull Analyst")}, or choose {localize_rating_term("Hold")} only if it is strongly justified based on the arguments presented.
Summarize the key points from both sides concisely, focusing on the most compelling evidence or reasoning. Your recommendation—{localize_rating_term("Buy")}, {localize_rating_term("Sell")}, or {localize_rating_term("Hold")}—must be clear and actionable. Avoid defaulting to {localize_rating_term("Hold")} simply because both sides have valid points; commit to a stance grounded in the debate's strongest arguments.
Additionally, develop a detailed investment plan for the trader. This should include:
Your Recommendation: A decisive stance supported by the most convincing arguments.
Rationale: An explanation of why these arguments lead to your conclusion.
Strategic Actions: Concrete steps for implementing the recommendation.
Take into account your past mistakes on similar situations. Use these insights to refine your decision-making and ensure you are learning and improving. Present your analysis conversationally, as if speaking naturally, without special formatting.
After your analysis, append a feedback block in this exact format:
{get_snapshot_template()}
{get_snapshot_writing_instruction()}
Here are your past reflections on mistakes:
\"{past_memory_str}\"
{instrument_context}
Here is the latest debate context:
{localize_label("Rolling debate brief:", "滚动辩论摘要:")}
{debate_brief}
{localize_label("Bull latest snapshot:", f"{localize_role_name('Bull Analyst')} 最新快照:")}
{bull_snapshot}
{localize_label("Bear latest snapshot:", f"{localize_role_name('Bear Analyst')} 最新快照:")}
{bear_snapshot}{get_language_instruction()}
"""
response = llm.invoke(prompt)
judge_snapshot = extract_feedback_snapshot(response.content)
updated_brief = build_debate_brief(
{
"Bull Analyst": bull_snapshot,
"Bear Analyst": bear_snapshot,
"Research Manager": judge_snapshot,
},
latest_speaker="Research Manager",
)
new_investment_debate_state = {
"judge_decision": response.content,
"history": investment_debate_state.get("history", ""),
"bear_history": investment_debate_state.get("bear_history", ""),
"bull_history": investment_debate_state.get("bull_history", ""),
"current_response": response.content,
"bull_snapshot": bull_snapshot,
"bear_snapshot": bear_snapshot,
"debate_brief": updated_brief,
"latest_speaker": "Research Manager",
"count": investment_debate_state["count"],
}
return {
"investment_debate_state": new_investment_debate_state,
"investment_plan": response.content,
}
return research_manager_node