from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder import time import json from tradingagents.agents.utils.agent_utils import get_yfinance_news, get_analyst_sentiment, get_sector_performance, execute_text_tool_calls, needs_followup_call, execute_default_tools, generate_analysis_from_data from tradingagents.dataflows.config import get_config from tradingagents.log_utils import add_log, step_timer, symbol_progress ANALYST_RESPONSE_FORMAT = """ RESPONSE FORMAT (follow this structure exactly): ## EXECUTIVE SUMMARY 2-3 sentences: Key sentiment finding and directional bias (BULLISH / BEARISH / NEUTRAL). ## KEY DATA POINTS - Bullet list of the 5 most significant sentiment signals with specific numbers - Include analyst consensus, price target implied upside, sector positioning ## SIGNAL ASSESSMENT Your overall sentiment reading: BULLISH / BEARISH / NEUTRAL 1-2 sentences explaining why, referencing specific data. ## RISK FACTORS 2-3 specific risks or sentiment divergences. ## CONFIDENCE: HIGH / MEDIUM / LOW 1 sentence justifying your confidence level. | Signal Source | Finding | Sentiment | Weight | |--------------|---------|-----------|--------| | (fill with key sentiment signals) | RULES: - Maximum 3000 characters total - Do NOT repeat raw data verbatim — summarize trends and insights - Complete your ENTIRE analysis in a SINGLE response""" def create_social_media_analyst(llm): def social_media_analyst_node(state): current_date = state["trade_date"] ticker = state["company_of_interest"] company_name = state["company_of_interest"] tools = [ get_yfinance_news, get_analyst_sentiment, get_sector_performance, ] system_message = ( "You are a sentiment and market perception analyst. Your job is to assess the overall market sentiment around a company " "by synthesizing multiple signal sources:\n" "- `get_yfinance_news`: Curated news from Yahoo Finance (multiple publishers) — analyze headlines, publishers, recency, and tone\n" "- `get_analyst_sentiment`: Wall Street consensus — price targets, buy/sell/hold distribution, implied upside/downside\n" "- `get_sector_performance`: Sector context — how the stock is positioned vs moving averages, 52-week range, beta, and index\n\n" "Synthesize these into a unified sentiment assessment. Quantify sentiment where possible (e.g., '70% of analysts rate Buy', " "'trading at 85% of 52-week range', 'implied upside of 15%'). Identify sentiment divergences (e.g., analysts bullish but " "price below moving averages). Do not simply state trends are mixed — provide specific, actionable insights." + ANALYST_RESPONSE_FORMAT, ) prompt = ChatPromptTemplate.from_messages( [ ( "system", "You are a helpful AI assistant, collaborating with other assistants." " Use the provided tools to progress towards answering the question." " If you are unable to fully answer, that's OK; another assistant with different tools" " will help where you left off. Execute what you can to make progress." " If you or any other assistant has the FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL** or deliverable," " prefix your response with FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL** so the team knows to stop." " You have access to the following tools: {tool_names}.\n{system_message}" "For your reference, the current date is {current_date}. The current company we want to analyze is {ticker}", ), MessagesPlaceholder(variable_name="messages"), ] ) prompt = prompt.partial(system_message=system_message) prompt = prompt.partial(tool_names=", ".join([tool.name for tool in tools])) prompt = prompt.partial(current_date=current_date) prompt = prompt.partial(ticker=ticker) chain = prompt | llm.bind_tools(tools) step_timer.start_step("social_media_analyst") add_log("agent", "social_analyst", f"💬 Social Media Analyst calling LLM for {ticker}...") t0 = time.time() result = chain.invoke(state["messages"]) elapsed = time.time() - t0 report = "" if len(result.tool_calls) == 0: report = result.content add_log("llm", "social_analyst", f"LLM responded in {elapsed:.1f}s ({len(report)} chars)") tool_results = execute_text_tool_calls(report, tools) if tool_results: add_log("data", "social_analyst", f"Executed {len(tool_results)} tool calls: {', '.join(t['name'] for t in tool_results)}") else: add_log("agent", "social_analyst", f"🔄 No tool calls found, proactively fetching data for {ticker}...") tool_results = execute_default_tools(tools, ticker, current_date) add_log("data", "social_analyst", f"Proactively fetched {len(tool_results)} data sources") if tool_results and needs_followup_call(report): add_log("agent", "social_analyst", f"🔄 Generating analysis from {len(tool_results)} tool results...") t1 = time.time() followup = generate_analysis_from_data(llm, tool_results, system_message, ticker, current_date) elapsed2 = time.time() - t1 if followup and len(followup) > 100: report = followup add_log("llm", "social_analyst", f"Follow-up analysis generated in {elapsed2:.1f}s ({len(report)} chars)") add_log("agent", "social_analyst", f"✅ Sentiment report ready: {report[:300]}...") step_timer.end_step("social_media_analyst", "completed", report[:200]) symbol_progress.step_done(ticker, "social_media_analyst") step_timer.update_details("social_media_analyst", { "system_prompt": system_message[:2000], "user_prompt": f"Analyze social media sentiment for {ticker} on {current_date}", "response": report[:3000], "tool_calls": tool_results if tool_results else [], }) else: tool_call_info = [{"name": tc["name"], "args": str(tc.get("args", {}))[:200]} for tc in result.tool_calls] step_timer.set_details("social_media_analyst", { "system_prompt": system_message[:2000], "user_prompt": f"Analyze social media sentiment for {ticker} on {current_date}", "response": "(Pending - tool calls in progress)", "tool_calls": tool_call_info, }) add_log("data", "social_analyst", f"LLM requested {len(result.tool_calls)} tool calls in {elapsed:.1f}s: {', '.join(tc['name'] for tc in result.tool_calls)}") return { "messages": [result], "sentiment_report": report, } return social_media_analyst_node