TradingAgents/tradingagents/agents/researchers/bear_researcher.py

103 lines
3.9 KiB
Python

from langchain_core.messages import AIMessage, SystemMessage, HumanMessage
import time
import json
from tradingagents.log_utils import add_log, step_timer, symbol_progress
def create_bear_researcher(llm, memory):
def bear_node(state) -> dict:
investment_debate_state = state["investment_debate_state"]
history = investment_debate_state.get("history", "")
bear_history = investment_debate_state.get("bear_history", "")
current_response = investment_debate_state.get("current_response", "")
market_research_report = state["market_report"]
sentiment_report = state["sentiment_report"]
news_report = state["news_report"]
fundamentals_report = state["fundamentals_report"]
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 Bear Analyst at a financial research firm. You MUST stay in character as a financial analyst 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 focus your analysis on the data that IS available
Your role: Present a case AGAINST investing in this stock by highlighting risks, challenges, and negative indicators.
Focus on: downside risks, competitive weaknesses, negative market signals, valuation concerns, macro headwinds.
RESPONSE FORMAT:
- Maximum 2000 characters. Focus on the 3-5 strongest bearish points.
- Complete your ENTIRE argument in a SINGLE response.
Respond only with your bearish financial analysis. No disclaimers or meta-commentary."""
user_prompt = f"""Analyze this stock from a bearish perspective:
MARKET DATA:
{market_research_report}
SENTIMENT:
{sentiment_report}
NEWS:
{news_report}
FUNDAMENTALS:
{fundamentals_report}
DEBATE HISTORY:
{history}
BULL'S LAST ARGUMENT:
{current_response}
PAST LEARNINGS:
{past_memory_str if past_memory_str else "None"}
Provide your bearish case highlighting risks and concerns."""
messages = [
SystemMessage(content=system_prompt),
HumanMessage(content=user_prompt)
]
step_timer.start_step("bear_researcher")
add_log("agent", "bear_researcher", f"🐻 Bear Analyst calling LLM...")
t0 = time.time()
response = llm.invoke(messages)
elapsed = time.time() - t0
add_log("llm", "bear_researcher", f"LLM responded in {elapsed:.1f}s ({len(response.content)} chars)")
add_log("agent", "bear_researcher", f"✅ Bear argument ready: {response.content[:300]}...")
step_timer.end_step("bear_researcher", "completed", response.content[:200])
symbol_progress.step_done(state["company_of_interest"], "bear_researcher")
step_timer.set_details("bear_researcher", {
"system_prompt": system_prompt,
"user_prompt": user_prompt[:3000],
"response": response.content[:3000],
"tool_calls": [],
})
argument = f"Bear Analyst: {response.content}"
new_investment_debate_state = {
"history": history + "\n" + argument,
"bear_history": bear_history + "\n" + argument,
"bull_history": investment_debate_state.get("bull_history", ""),
"current_response": argument,
"count": investment_debate_state["count"] + 1,
}
return {"investment_debate_state": new_investment_debate_state}
return bear_node