mirror of
https://github.com/simonlin1212/TradingAgents-astock.git
synced 2026-10-30 09:29:56 +00:00
simon
c567f8aee9
- 适配A股数据源(mootdx+腾讯财经+akshare)，全免费无积分墙 - 新增3个A股特化分析师(政策/游资/解禁)，共7位AI分析师 - A股交易规则约束(T+1/涨跌停/最小手数/ST标识) - 内置Streamlit Web UI + PDF报告导出 - 支持MiniMax/DeepSeek/智谱/通义/OpenAI/Anthropic等多家LLM - 沪深300基准替代SPY
68 lines
4.0 KiB
Python
68 lines
4.0 KiB
Python
|
|
|
|
def create_bear_researcher(llm):
|
|
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"]
|
|
policy_report = state.get("policy_report", "")
|
|
hot_money_report = state.get("hot_money_report", "")
|
|
lockup_report = state.get("lockup_report", "")
|
|
data_quality_summary = state.get("data_quality_summary", "")
|
|
|
|
prompt = f"""You are a Bear Analyst making the case against investing in this A-share (China mainland) stock. Your goal is to present a well-reasoned argument emphasizing risks, challenges, and negative indicators unique to the Chinese market. Leverage the provided research and data to highlight potential downsides and counter bullish arguments effectively.
|
|
|
|
A-Share Bear Framework — prioritize these China-specific risk factors:
|
|
- Policy Headwinds: Sudden regulatory crackdowns (e.g. industry rectification, antitrust), CSRC window guidance (窗口指导), sector-wide trading restrictions, or political risk signals
|
|
- Lockup & Insider Selling: Upcoming lockup expiry dates with large overhang, controlling shareholders in pre-disclosure reduction windows, equity pledge liquidation risk
|
|
- Hot Money Withdrawal (游资撤退): Volume divergence after limit-ups (放量滞涨), declining limit-up board count (连板断裂), sector rotation moving away from this theme
|
|
- Valuation Bubble: PE far above 30x A-stock growth anchor with EPS unable to digest within 3 years, PEG > 2 indicating overpriced growth, retail-driven speculative premium
|
|
- T+1 Trap: After a sharp rally, buyers today cannot exit until tomorrow — if sentiment reverses overnight or a gap-down opens, losses are locked in
|
|
- Northbound Retreat: Net outflow from Stock Connect signals foreign institutions reducing exposure
|
|
|
|
General bear points:
|
|
- Risks and Challenges: Market saturation, financial instability, or macroeconomic threats
|
|
- Competitive Weaknesses: Weaker market positioning, declining innovation, or competitor threats
|
|
- Negative Indicators: Evidence from financial data, market trends, or adverse news
|
|
- Bull Counterpoints: Expose over-optimistic assumptions with specific data
|
|
- Engagement: Present your argument conversationally, directly engaging with the bull analyst's points
|
|
|
|
Resources available:
|
|
Market research report: {market_research_report}
|
|
Social media sentiment report: {sentiment_report}
|
|
Latest news report: {news_report}
|
|
Company fundamentals report: {fundamentals_report}
|
|
Policy analysis report: {policy_report}
|
|
Hot money / capital flow report: {hot_money_report}
|
|
Lockup expiry / insider reduction report: {lockup_report}
|
|
Data quality assessment: {data_quality_summary}
|
|
Conversation history of the debate: {history}
|
|
Last bull argument: {current_response}
|
|
|
|
⚠️ If the data quality assessment flags any report as low-confidence (grade C/D/F), reduce your reliance on that report and note the data limitation in your argument.
|
|
|
|
Deliver a compelling bear argument grounded in A-share market realities. Refute the bull's claims and demonstrate the risks of investing in this stock within the Chinese regulatory and market structure.
|
|
"""
|
|
|
|
response = llm.invoke(prompt)
|
|
|
|
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
|