From 8536ccacdd1cb05b2e8d2d4aa195d72c5aeb2e0d Mon Sep 17 00:00:00 2001 From: Zhigong Liu Date: Fri, 17 Apr 2026 18:57:31 -0400 Subject: [PATCH 01/12] chore: ignore CLAUDE.md (local AI assistant context file) Co-Authored-By: Claude Sonnet 4.6 --- .gitignore | 3 +++ 1 file changed, 3 insertions(+) diff --git a/.gitignore b/.gitignore index 9a2904a9c..a8b1d6307 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,6 @@ +# Claude Code local files +CLAUDE.md + # Byte-compiled / optimized / DLL files __pycache__/ *.py[codz] From 6abc768c1d0d2d65263140ba6d30f086b4ea98f6 Mon Sep 17 00:00:00 2001 From: Zhigong Liu Date: Sun, 19 Apr 2026 22:13:53 -0400 Subject: [PATCH 02/12] feat: replace per-agent BM25 memory with persistent append-only decision log Co-Authored-By: Claude Sonnet 4.6 --- pyproject.toml | 1 - tests/test_memory_log.py | 648 ++++++++++++++++++ tradingagents/agents/__init__.py | 2 - .../agents/managers/portfolio_manager.py | 15 +- .../agents/managers/research_manager.py | 18 +- .../agents/researchers/bear_researcher.py | 12 +- .../agents/researchers/bull_researcher.py | 12 +- tradingagents/agents/trader/trader.py | 18 +- tradingagents/agents/utils/agent_states.py | 1 + tradingagents/agents/utils/memory.py | 364 ++++++---- tradingagents/default_config.py | 1 + tradingagents/graph/propagation.py | 3 +- tradingagents/graph/reflection.py | 137 +--- tradingagents/graph/setup.py | 28 +- tradingagents/graph/trading_graph.py | 126 +++- 15 files changed, 1046 insertions(+), 340 deletions(-) create mode 100644 tests/test_memory_log.py diff --git a/pyproject.toml b/pyproject.toml index 98385e32e..a1dfcd75e 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -20,7 +20,6 @@ dependencies = [ "parsel>=1.10.0", "pytz>=2025.2", "questionary>=2.1.0", - "rank-bm25>=0.2.2", "redis>=6.2.0", "requests>=2.32.4", "rich>=14.0.0", diff --git a/tests/test_memory_log.py b/tests/test_memory_log.py new file mode 100644 index 000000000..c299e7fa5 --- /dev/null +++ b/tests/test_memory_log.py @@ -0,0 +1,648 @@ +"""Tests for TradingMemoryLog — storage, deferred reflection, PM injection, legacy removal.""" + +import pytest +import pandas as pd +from unittest.mock import MagicMock, patch + +from tradingagents.agents.utils.memory import TradingMemoryLog +from tradingagents.graph.reflection import Reflector +from tradingagents.graph.trading_graph import TradingAgentsGraph +from tradingagents.graph.propagation import Propagator +from tradingagents.agents.managers.portfolio_manager import create_portfolio_manager + +_SEP = TradingMemoryLog._SEPARATOR + +DECISION_BUY = "Rating: Buy\nEnter at $189-192, 6% portfolio cap." +DECISION_OVERWEIGHT = ( + "Rating: Overweight\n" + "Executive Summary: Moderate position, await confirmation.\n" + "Investment Thesis: Strong fundamentals but near-term headwinds." +) +DECISION_SELL = "Rating: Sell\nExit position immediately." +DECISION_NO_RATING = ( + "Executive Summary: Complex situation with multiple competing factors.\n" + "Investment Thesis: No clear directional signal at this time." +) + + +# --------------------------------------------------------------------------- +# Shared helpers +# --------------------------------------------------------------------------- + +def make_log(tmp_path, filename="trading_memory.md"): + config = {"memory_log_path": str(tmp_path / filename)} + return TradingMemoryLog(config) + + +def _seed_completed(tmp_path, ticker, date, decision_text, reflection_text, filename="trading_memory.md"): + """Write a completed entry directly to file, bypassing the API.""" + entry = ( + f"[{date} | {ticker} | Buy | +1.0% | +0.5% | 5d]\n\n" + f"DECISION:\n{decision_text}\n\n" + f"REFLECTION:\n{reflection_text}" + + _SEP + ) + with open(tmp_path / filename, "a", encoding="utf-8") as f: + f.write(entry) + + +def _resolve_entry(log, ticker, date, decision, reflection="Good call."): + """Store a decision then immediately resolve it via the API.""" + log.store_decision(ticker, date, decision) + log.update_with_outcome(ticker, date, 0.05, 0.02, 5, reflection) + + +def _price_df(prices): + """Minimal DataFrame matching yfinance .history() output shape.""" + return pd.DataFrame({"Close": prices}) + + +def _make_pm_state(past_context=""): + """Minimal AgentState dict for portfolio_manager_node.""" + return { + "company_of_interest": "NVDA", + "past_context": past_context, + "risk_debate_state": { + "history": "Risk debate history.", + "aggressive_history": "", + "conservative_history": "", + "neutral_history": "", + "judge_decision": "", + "current_aggressive_response": "", + "current_conservative_response": "", + "current_neutral_response": "", + "count": 1, + }, + "market_report": "Market report.", + "sentiment_report": "Sentiment report.", + "news_report": "News report.", + "fundamentals_report": "Fundamentals report.", + "investment_plan": "Research plan.", + "trader_investment_plan": "Trader plan.", + } + + +# --------------------------------------------------------------------------- +# Core: storage and read path +# --------------------------------------------------------------------------- + +class TestTradingMemoryLogCore: + + def test_store_creates_file(self, tmp_path): + log = make_log(tmp_path) + assert not (tmp_path / "trading_memory.md").exists() + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert (tmp_path / "trading_memory.md").exists() + + def test_store_appends_not_overwrites(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.store_decision("AAPL", "2026-01-11", DECISION_OVERWEIGHT) + entries = log.load_entries() + assert len(entries) == 2 + assert entries[0]["ticker"] == "NVDA" + assert entries[1]["ticker"] == "AAPL" + + def test_store_decision_idempotent(self, tmp_path): + """Calling store_decision twice with same (ticker, date) stores only one entry.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert len(log.load_entries()) == 1 + + def test_batch_update_resolves_multiple_entries(self, tmp_path): + """batch_update_with_outcomes resolves multiple pending entries in one write.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-05", DECISION_BUY) + log.store_decision("NVDA", "2026-01-12", DECISION_SELL) + + updates = [ + {"ticker": "NVDA", "trade_date": "2026-01-05", + "raw_return": 0.05, "alpha_return": 0.02, "holding_days": 5, + "reflection": "First correct."}, + {"ticker": "NVDA", "trade_date": "2026-01-12", + "raw_return": -0.03, "alpha_return": -0.01, "holding_days": 5, + "reflection": "Second correct."}, + ] + log.batch_update_with_outcomes(updates) + + entries = log.load_entries() + assert len(entries) == 2 + assert all(not e["pending"] for e in entries) + assert entries[0]["reflection"] == "First correct." + assert entries[1]["reflection"] == "Second correct." + + def test_pending_tag_format(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + text = (tmp_path / "trading_memory.md").read_text(encoding="utf-8") + assert "[2026-01-10 | NVDA | Buy | pending]" in text + + # Rating parsing + + def test_rating_parsed_buy(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert log.load_entries()[0]["rating"] == "Buy" + + def test_rating_parsed_overweight(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("AAPL", "2026-01-11", DECISION_OVERWEIGHT) + assert log.load_entries()[0]["rating"] == "Overweight" + + def test_rating_fallback_hold(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("MSFT", "2026-01-12", DECISION_NO_RATING) + assert log.load_entries()[0]["rating"] == "Hold" + + def test_rating_priority_over_prose(self, tmp_path): + """'Rating: X' label wins even when an opposing rating word appears earlier in prose.""" + decision = ( + "The sell thesis is weak. The hold case is marginal.\n\n" + "Rating: Buy\n\n" + "Executive Summary: Strong fundamentals support the position." + ) + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", decision) + assert log.load_entries()[0]["rating"] == "Buy" + + # Delimiter robustness + + def test_decision_with_markdown_separator(self, tmp_path): + """LLM decision containing '---' must not corrupt the entry.""" + decision = "Rating: Buy\n\n---\n\nRisk: elevated volatility." + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", decision) + entries = log.load_entries() + assert len(entries) == 1 + assert "Risk: elevated volatility" in entries[0]["decision"] + + # load_entries + + def test_load_entries_empty_file(self, tmp_path): + log = make_log(tmp_path) + assert log.load_entries() == [] + + def test_load_entries_single(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + entries = log.load_entries() + assert len(entries) == 1 + e = entries[0] + assert e["date"] == "2026-01-10" + assert e["ticker"] == "NVDA" + assert e["rating"] == "Buy" + assert e["pending"] is True + assert e["raw"] is None + + def test_load_entries_multiple(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.store_decision("AAPL", "2026-01-11", DECISION_OVERWEIGHT) + log.store_decision("MSFT", "2026-01-12", DECISION_NO_RATING) + entries = log.load_entries() + assert len(entries) == 3 + assert [e["ticker"] for e in entries] == ["NVDA", "AAPL", "MSFT"] + + def test_decision_content_preserved(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert log.load_entries()[0]["decision"] == DECISION_BUY.strip() + + # get_pending_entries + + def test_get_pending_returns_pending_only(self, tmp_path): + log = make_log(tmp_path) + _seed_completed(tmp_path, "NVDA", "2026-01-05", "Buy NVDA.", "Correct.") + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + pending = log.get_pending_entries() + assert len(pending) == 1 + assert pending[0]["ticker"] == "NVDA" + assert pending[0]["date"] == "2026-01-10" + + # get_past_context + + def test_get_past_context_empty(self, tmp_path): + log = make_log(tmp_path) + assert log.get_past_context("NVDA") == "" + + def test_get_past_context_pending_excluded(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert log.get_past_context("NVDA") == "" + + def test_get_past_context_same_ticker(self, tmp_path): + log = make_log(tmp_path) + _seed_completed(tmp_path, "NVDA", "2026-01-05", "Buy NVDA — AI capex thesis intact.", "Directionally correct.") + ctx = log.get_past_context("NVDA") + assert "Past analyses of NVDA" in ctx + assert "Buy NVDA" in ctx + + def test_get_past_context_cross_ticker(self, tmp_path): + log = make_log(tmp_path) + _seed_completed(tmp_path, "AAPL", "2026-01-05", "Buy AAPL — Services growth.", "Correct.") + ctx = log.get_past_context("NVDA") + assert "Recent cross-ticker lessons" in ctx + assert "Past analyses of NVDA" not in ctx + + def test_n_same_limit_respected(self, tmp_path): + """Only the n_same most recent same-ticker entries are included.""" + log = make_log(tmp_path) + for i in range(6): + _seed_completed(tmp_path, "NVDA", f"2026-01-{i+1:02d}", f"Buy entry {i}.", "Correct.") + ctx = log.get_past_context("NVDA", n_same=5) + assert "Buy entry 0" not in ctx + assert "Buy entry 5" in ctx + + def test_n_cross_limit_respected(self, tmp_path): + """Only the n_cross most recent cross-ticker entries are included.""" + log = make_log(tmp_path) + for i, ticker in enumerate(["AAPL", "MSFT", "GOOG", "META"]): + _seed_completed(tmp_path, ticker, f"2026-01-{i+1:02d}", f"Buy {ticker}.", "Correct.") + ctx = log.get_past_context("NVDA", n_cross=3) + assert "AAPL" not in ctx + assert "META" in ctx + + # No-op when config is None + + def test_no_log_path_is_noop(self): + log = TradingMemoryLog(config=None) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + assert log.load_entries() == [] + assert log.get_past_context("NVDA") == "" + + # Rating parsing: markdown bold and numbered list formats + + def test_rating_parsed_from_bold_markdown(self, tmp_path): + """**Rating**: Buy — markdown bold wrapper must not prevent parsing.""" + decision = "**Rating**: Buy\nEnter at $190." + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", decision) + assert log.load_entries()[0]["rating"] == "Buy" + + def test_rating_parsed_from_numbered_list(self, tmp_path): + """1. Rating: Buy — numbered list prefix must not prevent parsing.""" + decision = "1. Rating: Buy\nEnter at $190." + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", decision) + assert log.load_entries()[0]["rating"] == "Buy" + + +# --------------------------------------------------------------------------- +# Deferred reflection: update_with_outcome, Reflector, _fetch_returns +# --------------------------------------------------------------------------- + +class TestDeferredReflection: + + # update_with_outcome + + def test_update_replaces_pending_tag(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.update_with_outcome("NVDA", "2026-01-10", 0.042, 0.021, 5, "Momentum confirmed.") + text = (tmp_path / "trading_memory.md").read_text(encoding="utf-8") + assert "[2026-01-10 | NVDA | Buy | pending]" not in text + assert "+4.2%" in text + assert "+2.1%" in text + assert "5d" in text + + def test_update_appends_reflection(self, tmp_path): + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.update_with_outcome("NVDA", "2026-01-10", 0.042, 0.021, 5, "Momentum confirmed.") + entries = log.load_entries() + assert len(entries) == 1 + e = entries[0] + assert e["pending"] is False + assert e["reflection"] == "Momentum confirmed." + assert e["decision"] == DECISION_BUY.strip() + + def test_update_preserves_other_entries(self, tmp_path): + """Only the matching entry is modified; all other entries remain unchanged.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.store_decision("AAPL", "2026-01-11", "Rating: Hold\nHold AAPL.") + log.store_decision("MSFT", "2026-01-12", DECISION_SELL) + log.update_with_outcome("AAPL", "2026-01-11", 0.01, -0.01, 5, "Neutral result.") + entries = log.load_entries() + assert len(entries) == 3 + nvda, aapl, msft = entries + assert nvda["ticker"] == "NVDA" and nvda["pending"] is True + assert aapl["ticker"] == "AAPL" and aapl["pending"] is False + assert aapl["reflection"] == "Neutral result." + assert msft["ticker"] == "MSFT" and msft["pending"] is True + + def test_update_atomic_write(self, tmp_path): + """A pre-existing .tmp file is overwritten; the log is correctly updated.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + stale_tmp = tmp_path / "trading_memory.tmp" + stale_tmp.write_text("GARBAGE CONTENT — should be overwritten", encoding="utf-8") + log.update_with_outcome("NVDA", "2026-01-10", 0.042, 0.021, 5, "Correct.") + assert not stale_tmp.exists() + entries = log.load_entries() + assert len(entries) == 1 + assert entries[0]["reflection"] == "Correct." + assert entries[0]["pending"] is False + + def test_update_noop_when_no_log_path(self): + log = TradingMemoryLog(config=None) + log.update_with_outcome("NVDA", "2026-01-10", 0.05, 0.02, 5, "Reflection") + + def test_formatting_roundtrip_after_update(self, tmp_path): + """All fields intact and blank line between tag and DECISION preserved after update.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-10", DECISION_BUY) + log.update_with_outcome("NVDA", "2026-01-10", 0.042, 0.021, 5, "Momentum confirmed.") + entries = log.load_entries() + assert len(entries) == 1 + e = entries[0] + assert e["pending"] is False + assert e["decision"] == DECISION_BUY.strip() + assert e["reflection"] == "Momentum confirmed." + assert e["raw"] == "+4.2%" + assert e["alpha"] == "+2.1%" + assert e["holding"] == "5d" + raw_text = (tmp_path / "trading_memory.md").read_text(encoding="utf-8") + assert "[2026-01-10 | NVDA | Buy | +4.2% | +2.1% | 5d]\n\nDECISION:" in raw_text + + # Reflector.reflect_on_final_decision + + def test_reflect_on_final_decision_returns_llm_output(self): + mock_llm = MagicMock() + mock_llm.invoke.return_value.content = "Directionally correct. Thesis confirmed." + reflector = Reflector(mock_llm) + result = reflector.reflect_on_final_decision( + final_decision=DECISION_BUY, raw_return=0.042, alpha_return=0.021 + ) + assert result == "Directionally correct. Thesis confirmed." + mock_llm.invoke.assert_called_once() + + def test_reflect_on_final_decision_includes_returns_in_prompt(self): + """Return figures are present in the human message sent to the LLM.""" + mock_llm = MagicMock() + mock_llm.invoke.return_value.content = "Incorrect call." + reflector = Reflector(mock_llm) + reflector.reflect_on_final_decision( + final_decision=DECISION_SELL, raw_return=-0.08, alpha_return=-0.05 + ) + messages = mock_llm.invoke.call_args[0][0] + human_content = next(content for role, content in messages if role == "human") + assert "-8.0%" in human_content + assert "-5.0%" in human_content + assert "Exit position immediately." in human_content + + # TradingAgentsGraph._fetch_returns + + def test_fetch_returns_valid_ticker(self): + stock_prices = [100.0, 102.0, 104.0, 103.0, 105.0, 106.0] + spy_prices = [400.0, 402.0, 404.0, 403.0, 405.0, 406.0] + mock_graph = MagicMock(spec=TradingAgentsGraph) + with patch("yfinance.Ticker") as mock_ticker_cls: + def _make_ticker(sym): + m = MagicMock() + m.history.return_value = _price_df(spy_prices if sym == "SPY" else stock_prices) + return m + mock_ticker_cls.side_effect = _make_ticker + raw, alpha, days = TradingAgentsGraph._fetch_returns(mock_graph, "NVDA", "2026-01-05") + assert raw is not None and alpha is not None and days is not None + assert isinstance(raw, float) and isinstance(alpha, float) and isinstance(days, int) + assert days == 5 + + def test_fetch_returns_too_recent(self): + """Only 1 data point available → returns (None, None, None), no crash.""" + mock_graph = MagicMock(spec=TradingAgentsGraph) + with patch("yfinance.Ticker") as mock_ticker_cls: + m = MagicMock() + m.history.return_value = _price_df([100.0]) + mock_ticker_cls.return_value = m + raw, alpha, days = TradingAgentsGraph._fetch_returns(mock_graph, "NVDA", "2026-04-19") + assert raw is None and alpha is None and days is None + + def test_fetch_returns_delisted(self): + """Empty DataFrame → returns (None, None, None), no crash.""" + mock_graph = MagicMock(spec=TradingAgentsGraph) + with patch("yfinance.Ticker") as mock_ticker_cls: + m = MagicMock() + m.history.return_value = pd.DataFrame({"Close": []}) + mock_ticker_cls.return_value = m + raw, alpha, days = TradingAgentsGraph._fetch_returns(mock_graph, "XXXXXFAKE", "2026-01-10") + assert raw is None and alpha is None and days is None + + def test_fetch_returns_spy_shorter_than_stock(self): + """SPY having fewer rows than the stock must not raise IndexError.""" + stock_prices = [100.0, 102.0, 104.0, 103.0, 105.0, 106.0] + spy_prices = [400.0, 402.0, 403.0] + mock_graph = MagicMock(spec=TradingAgentsGraph) + with patch("yfinance.Ticker") as mock_ticker_cls: + def _make_ticker(sym): + m = MagicMock() + m.history.return_value = _price_df(spy_prices if sym == "SPY" else stock_prices) + return m + mock_ticker_cls.side_effect = _make_ticker + raw, alpha, days = TradingAgentsGraph._fetch_returns(mock_graph, "NVDA", "2026-01-05") + assert raw is not None and alpha is not None and days is not None + assert days == 2 + + # TradingAgentsGraph._resolve_pending_entries + + def test_resolve_skips_other_tickers(self, tmp_path): + """Pending AAPL entry is not resolved when the run is for NVDA.""" + log = make_log(tmp_path) + log.store_decision("AAPL", "2026-01-10", DECISION_BUY) + mock_graph = MagicMock(spec=TradingAgentsGraph) + mock_graph.memory_log = log + mock_graph._fetch_returns = MagicMock(return_value=(0.05, 0.02, 5)) + TradingAgentsGraph._resolve_pending_entries(mock_graph, "NVDA") + mock_graph._fetch_returns.assert_not_called() + assert len(log.get_pending_entries()) == 1 + + def test_resolve_marks_entry_completed(self, tmp_path): + """After resolve, get_pending_entries() is empty and the entry has a REFLECTION.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-05", DECISION_BUY) + mock_reflector = MagicMock() + mock_reflector.reflect_on_final_decision.return_value = "Momentum confirmed." + mock_graph = MagicMock(spec=TradingAgentsGraph) + mock_graph.memory_log = log + mock_graph.reflector = mock_reflector + mock_graph._fetch_returns = MagicMock(return_value=(0.05, 0.02, 5)) + TradingAgentsGraph._resolve_pending_entries(mock_graph, "NVDA") + assert log.get_pending_entries() == [] + entries = log.load_entries() + assert len(entries) == 1 + assert entries[0]["pending"] is False + assert entries[0]["reflection"] == "Momentum confirmed." + assert "+5.0%" in entries[0]["raw"] + assert "+2.0%" in entries[0]["alpha"] + + +# --------------------------------------------------------------------------- +# Portfolio Manager injection: past_context in state and prompt +# --------------------------------------------------------------------------- + +class TestPortfolioManagerInjection: + + # past_context in initial state + + def test_past_context_in_initial_state(self): + propagator = Propagator() + state = propagator.create_initial_state("NVDA", "2026-01-10", past_context="some context") + assert "past_context" in state + assert state["past_context"] == "some context" + + def test_past_context_defaults_to_empty(self): + propagator = Propagator() + state = propagator.create_initial_state("NVDA", "2026-01-10") + assert state["past_context"] == "" + + # PM prompt + + def test_pm_prompt_includes_past_context(self): + captured = {} + mock_llm = MagicMock() + mock_llm.invoke.side_effect = lambda prompt: ( + captured.__setitem__("prompt", prompt) or MagicMock(content="Rating: Hold\nHold.") + ) + pm_node = create_portfolio_manager(mock_llm) + state = _make_pm_state(past_context="[2026-01-05 | NVDA | Buy | +5.0% | +2.0% | 5d]\nGreat call.") + pm_node(state) + assert "Past decisions on this stock" in captured["prompt"] + assert "Great call." in captured["prompt"] + + def test_pm_no_past_context_no_section(self): + """PM prompt omits the lessons section entirely when past_context is empty.""" + captured = {} + mock_llm = MagicMock() + mock_llm.invoke.side_effect = lambda prompt: ( + captured.__setitem__("prompt", prompt) or MagicMock(content="Rating: Hold\nHold.") + ) + pm_node = create_portfolio_manager(mock_llm) + state = _make_pm_state(past_context="") + pm_node(state) + assert "Past decisions on this stock" not in captured["prompt"] + assert "lessons learned" not in captured["prompt"] + + # get_past_context ordering and limits + + def test_same_ticker_prioritised(self, tmp_path): + """Same-ticker entries in same-ticker section; cross-ticker entries in cross-ticker section.""" + log = make_log(tmp_path) + _resolve_entry(log, "NVDA", "2026-01-05", DECISION_BUY, "Momentum confirmed.") + _resolve_entry(log, "AAPL", "2026-01-06", DECISION_SELL, "Overvalued.") + result = log.get_past_context("NVDA") + assert "Past analyses of NVDA" in result + assert "Recent cross-ticker lessons" in result + same_block, cross_block = result.split("Recent cross-ticker lessons") + assert "NVDA" in same_block + assert "AAPL" in cross_block + + def test_cross_ticker_reflection_only(self, tmp_path): + """Cross-ticker entries show only the REFLECTION text, not the full DECISION.""" + log = make_log(tmp_path) + _resolve_entry(log, "AAPL", "2026-01-06", DECISION_SELL, "Overvalued correction.") + result = log.get_past_context("NVDA") + assert "Overvalued correction." in result + assert "Exit position immediately." not in result + + def test_n_same_limit_respected(self, tmp_path): + """More than 5 same-ticker completed entries → only 5 injected.""" + log = make_log(tmp_path) + for i in range(7): + _resolve_entry(log, "NVDA", f"2026-01-{i+1:02d}", DECISION_BUY, f"Lesson {i}.") + result = log.get_past_context("NVDA", n_same=5) + lessons_present = sum(1 for i in range(7) if f"Lesson {i}." in result) + assert lessons_present == 5 + + def test_n_cross_limit_respected(self, tmp_path): + """More than 3 cross-ticker completed entries → only 3 injected.""" + log = make_log(tmp_path) + tickers = ["AAPL", "MSFT", "TSLA", "AMZN", "GOOG"] + for i, ticker in enumerate(tickers): + _resolve_entry(log, ticker, f"2026-01-{i+1:02d}", DECISION_BUY, f"{ticker} lesson.") + result = log.get_past_context("NVDA", n_cross=3) + cross_count = sum(result.count(f"{t} lesson.") for t in tickers) + assert cross_count == 3 + + # Full A→B→C integration cycle + + def test_full_cycle_store_resolve_inject(self, tmp_path): + """store pending → resolve with outcome → past_context non-empty for PM.""" + log = make_log(tmp_path) + log.store_decision("NVDA", "2026-01-05", DECISION_BUY) + assert len(log.get_pending_entries()) == 1 + assert log.get_past_context("NVDA") == "" + log.update_with_outcome("NVDA", "2026-01-05", 0.05, 0.02, 5, "Correct call.") + assert log.get_pending_entries() == [] + past_ctx = log.get_past_context("NVDA") + assert past_ctx != "" + assert "NVDA" in past_ctx + assert "Correct call." in past_ctx + assert "DECISION:" in past_ctx + assert "REFLECTION:" in past_ctx + + +# --------------------------------------------------------------------------- +# Legacy removal: BM25 / FinancialSituationMemory fully gone +# --------------------------------------------------------------------------- + +class TestLegacyRemoval: + + def test_financial_situation_memory_removed(self): + """FinancialSituationMemory must not be importable from the memory module.""" + import tradingagents.agents.utils.memory as m + assert not hasattr(m, "FinancialSituationMemory") + + def test_bm25_not_imported(self): + """rank_bm25 must not be present in the memory module namespace.""" + import tradingagents.agents.utils.memory as m + assert not hasattr(m, "BM25Okapi") + + def test_reflect_and_remember_removed(self): + """TradingAgentsGraph must not expose reflect_and_remember.""" + assert not hasattr(TradingAgentsGraph, "reflect_and_remember") + + def test_portfolio_manager_no_memory_param(self): + """create_portfolio_manager accepts only llm; passing memory= raises TypeError.""" + mock_llm = MagicMock() + create_portfolio_manager(mock_llm) + with pytest.raises(TypeError): + create_portfolio_manager(mock_llm, memory=MagicMock()) + + def test_full_pipeline_no_regression(self, tmp_path): + """propagate() completes without AttributeError after legacy cleanup.""" + fake_state = { + "final_trade_decision": "Rating: Buy\nBuy NVDA.", + "company_of_interest": "NVDA", + "trade_date": "2026-01-10", + "market_report": "", + "sentiment_report": "", + "news_report": "", + "fundamentals_report": "", + "investment_debate_state": { + "bull_history": "", "bear_history": "", "history": "", + "current_response": "", "judge_decision": "", + }, + "investment_plan": "", + "trader_investment_plan": "", + "risk_debate_state": { + "aggressive_history": "", "conservative_history": "", + "neutral_history": "", "history": "", "judge_decision": "", + "current_aggressive_response": "", "current_conservative_response": "", + "current_neutral_response": "", "count": 1, "latest_speaker": "", + }, + } + mock_graph = MagicMock() + mock_graph.memory_log = TradingMemoryLog({"memory_log_path": str(tmp_path / "mem.md")}) + mock_graph.log_states_dict = {} + mock_graph.debug = False + mock_graph.config = {"results_dir": str(tmp_path)} + mock_graph.graph.invoke.return_value = fake_state + mock_graph.propagator.create_initial_state.return_value = fake_state + mock_graph.propagator.get_graph_args.return_value = {} + mock_graph.signal_processor.process_signal.return_value = "Buy" + TradingAgentsGraph.propagate(mock_graph, "NVDA", "2026-01-10") + entries = mock_graph.memory_log.load_entries() + assert len(entries) == 1 + assert entries[0]["ticker"] == "NVDA" + assert entries[0]["pending"] is True diff --git a/tradingagents/agents/__init__.py b/tradingagents/agents/__init__.py index 1f03642c6..2fb4e1bac 100644 --- a/tradingagents/agents/__init__.py +++ b/tradingagents/agents/__init__.py @@ -1,6 +1,5 @@ from .utils.agent_utils import create_msg_delete from .utils.agent_states import AgentState, InvestDebateState, RiskDebateState -from .utils.memory import FinancialSituationMemory from .analysts.fundamentals_analyst import create_fundamentals_analyst from .analysts.market_analyst import create_market_analyst @@ -20,7 +19,6 @@ from .managers.portfolio_manager import create_portfolio_manager from .trader.trader import create_trader __all__ = [ - "FinancialSituationMemory", "AgentState", "create_msg_delete", "InvestDebateState", diff --git a/tradingagents/agents/managers/portfolio_manager.py b/tradingagents/agents/managers/portfolio_manager.py index 6c69ae9fd..215236156 100644 --- a/tradingagents/agents/managers/portfolio_manager.py +++ b/tradingagents/agents/managers/portfolio_manager.py @@ -1,7 +1,7 @@ from tradingagents.agents.utils.agent_utils import build_instrument_context, get_language_instruction -def create_portfolio_manager(llm, memory): +def create_portfolio_manager(llm): def portfolio_manager_node(state) -> dict: instrument_context = build_instrument_context(state["company_of_interest"]) @@ -15,12 +15,11 @@ def create_portfolio_manager(llm, memory): research_plan = state["investment_plan"] trader_plan = state["trader_investment_plan"] - 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" + past_context = state.get("past_context", "") + lessons_line = ( + f"- Past decisions on this stock and lessons learned:\n{past_context}\n" + if past_context else "" + ) prompt = f"""As the Portfolio Manager, synthesize the risk analysts' debate and deliver the final trading decision. @@ -38,7 +37,7 @@ def create_portfolio_manager(llm, memory): **Context:** - Research Manager's investment plan: **{research_plan}** - Trader's transaction proposal: **{trader_plan}** -- Lessons from past decisions: **{past_memory_str}** +{lessons_line} **Required Output Structure:** 1. **Rating**: State one of Buy / Overweight / Hold / Underweight / Sell. diff --git a/tradingagents/agents/managers/research_manager.py b/tradingagents/agents/managers/research_manager.py index 5b4b4fdc5..3902a60c4 100644 --- a/tradingagents/agents/managers/research_manager.py +++ b/tradingagents/agents/managers/research_manager.py @@ -2,24 +2,13 @@ from tradingagents.agents.utils.agent_utils import build_instrument_context -def create_research_manager(llm, memory): +def create_research_manager(llm): def research_manager_node(state) -> dict: instrument_context = build_instrument_context(state["company_of_interest"]) history = state["investment_debate_state"].get("history", "") - market_research_report = state["market_report"] - sentiment_report = state["sentiment_report"] - news_report = state["news_report"] - fundamentals_report = state["fundamentals_report"] investment_debate_state = state["investment_debate_state"] - 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 bear analyst, the bull analyst, or choose 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—Buy, Sell, or Hold—must be clear and actionable. Avoid defaulting to Hold simply because both sides have valid points; commit to a stance grounded in the debate's strongest arguments. @@ -29,10 +18,7 @@ Additionally, develop a detailed investment plan for the trader. This should inc 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. - -Here are your past reflections on mistakes: -\"{past_memory_str}\" +Present your analysis conversationally, as if speaking naturally, without special formatting. {instrument_context} diff --git a/tradingagents/agents/researchers/bear_researcher.py b/tradingagents/agents/researchers/bear_researcher.py index a44212dc4..9cde9d39c 100644 --- a/tradingagents/agents/researchers/bear_researcher.py +++ b/tradingagents/agents/researchers/bear_researcher.py @@ -1,6 +1,6 @@ -def create_bear_researcher(llm, memory): +def create_bear_researcher(llm): def bear_node(state) -> dict: investment_debate_state = state["investment_debate_state"] history = investment_debate_state.get("history", "") @@ -12,13 +12,6 @@ def create_bear_researcher(llm, memory): 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_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"""You are a Bear Analyst making the case against investing in the stock. Your goal is to present a well-reasoned argument emphasizing risks, challenges, and negative indicators. Leverage the provided research and data to highlight potential downsides and counter bullish arguments effectively. Key points to focus on: @@ -37,8 +30,7 @@ Latest world affairs news: {news_report} Company fundamentals report: {fundamentals_report} Conversation history of the debate: {history} Last bull argument: {current_response} -Reflections from similar situations and lessons learned: {past_memory_str} -Use this information to deliver a compelling bear argument, refute the bull's claims, and engage in a dynamic debate that demonstrates the risks and weaknesses of investing in the stock. You must also address reflections and learn from lessons and mistakes you made in the past. +Use this information to deliver a compelling bear argument, refute the bull's claims, and engage in a dynamic debate that demonstrates the risks and weaknesses of investing in the stock. """ response = llm.invoke(prompt) diff --git a/tradingagents/agents/researchers/bull_researcher.py b/tradingagents/agents/researchers/bull_researcher.py index d23d4d76e..d16bc2371 100644 --- a/tradingagents/agents/researchers/bull_researcher.py +++ b/tradingagents/agents/researchers/bull_researcher.py @@ -1,6 +1,6 @@ -def create_bull_researcher(llm, memory): +def create_bull_researcher(llm): def bull_node(state) -> dict: investment_debate_state = state["investment_debate_state"] history = investment_debate_state.get("history", "") @@ -12,13 +12,6 @@ def create_bull_researcher(llm, memory): 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_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"""You are a Bull Analyst advocating for investing in the stock. Your task is to build a strong, evidence-based case emphasizing growth potential, competitive advantages, and positive market indicators. Leverage the provided research and data to address concerns and counter bearish arguments effectively. Key points to focus on: @@ -35,8 +28,7 @@ Latest world affairs news: {news_report} Company fundamentals report: {fundamentals_report} Conversation history of the debate: {history} Last bear argument: {current_response} -Reflections from similar situations and lessons learned: {past_memory_str} -Use this information to deliver a compelling bull argument, refute the bear's concerns, and engage in a dynamic debate that demonstrates the strengths of the bull position. You must also address reflections and learn from lessons and mistakes you made in the past. +Use this information to deliver a compelling bull argument, refute the bear's concerns, and engage in a dynamic debate that demonstrates the strengths of the bull position. """ response = llm.invoke(prompt) diff --git a/tradingagents/agents/trader/trader.py b/tradingagents/agents/trader/trader.py index 07e9f262c..0ecae8888 100644 --- a/tradingagents/agents/trader/trader.py +++ b/tradingagents/agents/trader/trader.py @@ -3,25 +3,11 @@ import functools from tradingagents.agents.utils.agent_utils import build_instrument_context -def create_trader(llm, memory): +def create_trader(llm): def trader_node(state, name): company_name = state["company_of_interest"] instrument_context = build_instrument_context(company_name) investment_plan = state["investment_plan"] - 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_memories = memory.get_memories(curr_situation, n_matches=2) - - past_memory_str = "" - if past_memories: - for i, rec in enumerate(past_memories, 1): - past_memory_str += rec["recommendation"] + "\n\n" - else: - past_memory_str = "No past memories found." context = { "role": "user", @@ -31,7 +17,7 @@ def create_trader(llm, memory): messages = [ { "role": "system", - "content": f"""You are a trading agent analyzing market data to make investment decisions. Based on your analysis, provide a specific recommendation to buy, sell, or hold. End with a firm decision and always conclude your response with 'FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**' to confirm your recommendation. Apply lessons from past decisions to strengthen your analysis. Here are reflections from similar situations you traded in and the lessons learned: {past_memory_str}""", + "content": "You are a trading agent analyzing market data to make investment decisions. Based on your analysis, provide a specific recommendation to buy, sell, or hold. End with a firm decision and always conclude your response with 'FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**' to confirm your recommendation.", }, context, ] diff --git a/tradingagents/agents/utils/agent_states.py b/tradingagents/agents/utils/agent_states.py index 6423b9363..6151a3863 100644 --- a/tradingagents/agents/utils/agent_states.py +++ b/tradingagents/agents/utils/agent_states.py @@ -70,3 +70,4 @@ class AgentState(MessagesState): RiskDebateState, "Current state of the debate on evaluating risk" ] final_trade_decision: Annotated[str, "Final decision made by the Risk Analysts"] + past_context: Annotated[str, "Memory log context injected at run start (same-ticker decisions + cross-ticker lessons)"] diff --git a/tradingagents/agents/utils/memory.py b/tradingagents/agents/utils/memory.py index 2aefa7a38..fd14449e3 100644 --- a/tradingagents/agents/utils/memory.py +++ b/tradingagents/agents/utils/memory.py @@ -1,144 +1,272 @@ -"""Financial situation memory using BM25 for lexical similarity matching. +"""Append-only markdown decision log for TradingAgents.""" -Uses BM25 (Best Matching 25) algorithm for retrieval - no API calls, -no token limits, works offline with any LLM provider. -""" - -from rank_bm25 import BM25Okapi -from typing import List, Tuple +from typing import List, Optional +from pathlib import Path import re -class FinancialSituationMemory: - """Memory system for storing and retrieving financial situations using BM25.""" +class TradingMemoryLog: + """Append-only markdown log of trading decisions and reflections.""" - def __init__(self, name: str, config: dict = None): - """Initialize the memory system. + RATINGS = {"buy", "overweight", "hold", "underweight", "sell"} + # HTML comment: cannot appear in LLM prose output, safe as a hard delimiter + _SEPARATOR = "\n\n\n\n" + # Precompiled patterns — avoids re-compilation on every load_entries() call + _DECISION_RE = re.compile(r"DECISION:\n(.*?)(?=\nREFLECTION:|\Z)", re.DOTALL) + _REFLECTION_RE = re.compile(r"REFLECTION:\n(.*?)$", re.DOTALL) + _RATING_LABEL_RE = re.compile(r"rating.*?[:\-]\s*(\w+)", re.IGNORECASE) - Args: - name: Name identifier for this memory instance - config: Configuration dict (kept for API compatibility, not used for BM25) - """ - self.name = name - self.documents: List[str] = [] - self.recommendations: List[str] = [] - self.bm25 = None + def __init__(self, config: dict = None): + self._log_path = None + path = (config or {}).get("memory_log_path") + if path: + self._log_path = Path(path).expanduser() + self._log_path.parent.mkdir(parents=True, exist_ok=True) - def _tokenize(self, text: str) -> List[str]: - """Tokenize text for BM25 indexing. + # --- Write path (Phase A) --- - Simple whitespace + punctuation tokenization with lowercasing. - """ - # Lowercase and split on non-alphanumeric characters - tokens = re.findall(r'\b\w+\b', text.lower()) - return tokens + def store_decision( + self, + ticker: str, + trade_date: str, + final_trade_decision: str, + ) -> None: + """Append pending entry at end of propagate(). No LLM call.""" + if not self._log_path: + return + # Idempotency guard: fast raw-text scan instead of full parse + if self._log_path.exists(): + raw = self._log_path.read_text(encoding="utf-8") + for line in raw.splitlines(): + if line.startswith(f"[{trade_date} | {ticker} |") and line.endswith("| pending]"): + return + rating = self._parse_rating(final_trade_decision) + tag = f"[{trade_date} | {ticker} | {rating} | pending]" + entry = f"{tag}\n\nDECISION:\n{final_trade_decision}{self._SEPARATOR}" + with open(self._log_path, "a", encoding="utf-8") as f: + f.write(entry) - def _rebuild_index(self): - """Rebuild the BM25 index after adding documents.""" - if self.documents: - tokenized_docs = [self._tokenize(doc) for doc in self.documents] - self.bm25 = BM25Okapi(tokenized_docs) - else: - self.bm25 = None + # --- Read path (Phase A) --- - def add_situations(self, situations_and_advice: List[Tuple[str, str]]): - """Add financial situations and their corresponding advice. - - Args: - situations_and_advice: List of tuples (situation, recommendation) - """ - for situation, recommendation in situations_and_advice: - self.documents.append(situation) - self.recommendations.append(recommendation) - - # Rebuild BM25 index with new documents - self._rebuild_index() - - def get_memories(self, current_situation: str, n_matches: int = 1) -> List[dict]: - """Find matching recommendations using BM25 similarity. - - Args: - current_situation: The current financial situation to match against - n_matches: Number of top matches to return - - Returns: - List of dicts with matched_situation, recommendation, and similarity_score - """ - if not self.documents or self.bm25 is None: + def load_entries(self) -> List[dict]: + """Parse all entries from log. Returns list of dicts.""" + if not self._log_path or not self._log_path.exists(): return [] + text = self._log_path.read_text(encoding="utf-8") + raw_entries = [e.strip() for e in text.split(self._SEPARATOR) if e.strip()] + entries = [] + for raw in raw_entries: + parsed = self._parse_entry(raw) + if parsed: + entries.append(parsed) + return entries - # Tokenize query - query_tokens = self._tokenize(current_situation) + def get_pending_entries(self) -> List[dict]: + """Return entries with outcome:pending (for Phase B).""" + return [e for e in self.load_entries() if e.get("pending")] - # Get BM25 scores for all documents - scores = self.bm25.get_scores(query_tokens) + def get_past_context(self, ticker: str, n_same: int = 5, n_cross: int = 3) -> str: + """Return formatted past context string for agent prompt injection.""" + entries = [e for e in self.load_entries() if not e.get("pending")] + if not entries: + return "" - # Get top-n indices sorted by score (descending) - top_indices = sorted(range(len(scores)), key=lambda i: scores[i], reverse=True)[:n_matches] + same, cross = [], [] + for e in reversed(entries): + if len(same) >= n_same and len(cross) >= n_cross: + break + if e["ticker"] == ticker and len(same) < n_same: + same.append(e) + elif e["ticker"] != ticker and len(cross) < n_cross: + cross.append(e) - # Build results - results = [] - max_score = float(scores.max()) if len(scores) > 0 and scores.max() > 0 else 1.0 + if not same and not cross: + return "" - for idx in top_indices: - # Normalize score to 0-1 range for consistency - normalized_score = scores[idx] / max_score if max_score > 0 else 0 - results.append({ - "matched_situation": self.documents[idx], - "recommendation": self.recommendations[idx], - "similarity_score": normalized_score, - }) + parts = [] + if same: + parts.append(f"Past analyses of {ticker} (most recent first):") + parts.extend(self._format_full(e) for e in same) + if cross: + parts.append("Recent cross-ticker lessons:") + parts.extend(self._format_reflection_only(e) for e in cross) + return "\n\n".join(parts) - return results + # --- Update path (Phase B) --- - def clear(self): - """Clear all stored memories.""" - self.documents = [] - self.recommendations = [] - self.bm25 = None + def update_with_outcome( + self, + ticker: str, + trade_date: str, + raw_return: float, + alpha_return: float, + holding_days: int, + reflection: str, + ) -> None: + """Replace pending tag and append REFLECTION section using atomic write. + Finds the first pending entry matching (trade_date, ticker), updates + its tag with return figures, and appends a REFLECTION section. Uses + a temp-file + os.replace() so a crash mid-write never corrupts the log. + """ + if not self._log_path or not self._log_path.exists(): + return -if __name__ == "__main__": - # Example usage - matcher = FinancialSituationMemory("test_memory") + text = self._log_path.read_text(encoding="utf-8") + blocks = text.split(self._SEPARATOR) - # Example data - example_data = [ - ( - "High inflation rate with rising interest rates and declining consumer spending", - "Consider defensive sectors like consumer staples and utilities. Review fixed-income portfolio duration.", - ), - ( - "Tech sector showing high volatility with increasing institutional selling pressure", - "Reduce exposure to high-growth tech stocks. Look for value opportunities in established tech companies with strong cash flows.", - ), - ( - "Strong dollar affecting emerging markets with increasing forex volatility", - "Hedge currency exposure in international positions. Consider reducing allocation to emerging market debt.", - ), - ( - "Market showing signs of sector rotation with rising yields", - "Rebalance portfolio to maintain target allocations. Consider increasing exposure to sectors benefiting from higher rates.", - ), - ] + pending_prefix = f"[{trade_date} | {ticker} |" + raw_pct = f"{raw_return:+.1%}" + alpha_pct = f"{alpha_return:+.1%}" - # Add the example situations and recommendations - matcher.add_situations(example_data) + updated = False + new_blocks = [] + for block in blocks: + stripped = block.strip() + if not stripped: + new_blocks.append(block) + continue - # Example query - current_situation = """ - Market showing increased volatility in tech sector, with institutional investors - reducing positions and rising interest rates affecting growth stock valuations - """ + lines = stripped.splitlines() + tag_line = lines[0].strip() - try: - recommendations = matcher.get_memories(current_situation, n_matches=2) + if ( + not updated + and tag_line.startswith(pending_prefix) + and tag_line.endswith("| pending]") + ): + # Parse rating from the existing pending tag + fields = [f.strip() for f in tag_line[1:-1].split("|")] + rating = fields[2] + new_tag = ( + f"[{trade_date} | {ticker} | {rating}" + f" | {raw_pct} | {alpha_pct} | {holding_days}d]" + ) + rest = "\n".join(lines[1:]) + new_blocks.append( + f"{new_tag}\n\n{rest.lstrip()}\n\nREFLECTION:\n{reflection}" + ) + updated = True + else: + new_blocks.append(block) - for i, rec in enumerate(recommendations, 1): - print(f"\nMatch {i}:") - print(f"Similarity Score: {rec['similarity_score']:.2f}") - print(f"Matched Situation: {rec['matched_situation']}") - print(f"Recommendation: {rec['recommendation']}") + if not updated: + return - except Exception as e: - print(f"Error during recommendation: {str(e)}") + new_text = self._SEPARATOR.join(new_blocks) + tmp_path = self._log_path.with_suffix(".tmp") + tmp_path.write_text(new_text, encoding="utf-8") + tmp_path.replace(self._log_path) + + def batch_update_with_outcomes(self, updates: List[dict]) -> None: + """Apply multiple outcome updates in a single read + atomic write. + + Each element of updates must have keys: ticker, trade_date, + raw_return, alpha_return, holding_days, reflection. + """ + if not self._log_path or not self._log_path.exists() or not updates: + return + + text = self._log_path.read_text(encoding="utf-8") + blocks = text.split(self._SEPARATOR) + + # Build lookup keyed by (trade_date, ticker) for O(1) dispatch + update_map = {(u["trade_date"], u["ticker"]): u for u in updates} + + new_blocks = [] + for block in blocks: + stripped = block.strip() + if not stripped: + new_blocks.append(block) + continue + + lines = stripped.splitlines() + tag_line = lines[0].strip() + + matched = False + for (trade_date, ticker), upd in list(update_map.items()): + pending_prefix = f"[{trade_date} | {ticker} |" + if tag_line.startswith(pending_prefix) and tag_line.endswith("| pending]"): + fields = [f.strip() for f in tag_line[1:-1].split("|")] + rating = fields[2] + raw_pct = f"{upd['raw_return']:+.1%}" + alpha_pct = f"{upd['alpha_return']:+.1%}" + new_tag = ( + f"[{trade_date} | {ticker} | {rating}" + f" | {raw_pct} | {alpha_pct} | {upd['holding_days']}d]" + ) + rest = "\n".join(lines[1:]) + new_blocks.append( + f"{new_tag}\n\n{rest.lstrip()}\n\nREFLECTION:\n{upd['reflection']}" + ) + del update_map[(trade_date, ticker)] + matched = True + break + + if not matched: + new_blocks.append(block) + + new_text = self._SEPARATOR.join(new_blocks) + tmp_path = self._log_path.with_suffix(".tmp") + tmp_path.write_text(new_text, encoding="utf-8") + tmp_path.replace(self._log_path) + + # --- Helpers --- + + def _parse_rating(self, text: str) -> str: + # First pass: explicit "Rating: X" label — search handles markdown bold/numbered lists + for line in text.splitlines(): + m = self._RATING_LABEL_RE.search(line) + if m and m.group(1).lower() in self.RATINGS: + return m.group(1).capitalize() + # Fallback: first rating word found anywhere in the text + for line in text.splitlines(): + for word in line.lower().split(): + clean = word.strip("*:.,") + if clean in self.RATINGS: + return clean.capitalize() + return "Hold" + + def _parse_entry(self, raw: str) -> Optional[dict]: + lines = raw.strip().splitlines() + if not lines: + return None + tag_line = lines[0].strip() + if not (tag_line.startswith("[") and tag_line.endswith("]")): + return None + fields = [f.strip() for f in tag_line[1:-1].split("|")] + if len(fields) < 4: + return None + entry = { + "date": fields[0], + "ticker": fields[1], + "rating": fields[2], + "pending": fields[3] == "pending", + "raw": fields[3] if fields[3] != "pending" else None, + "alpha": fields[4] if len(fields) > 4 else None, + "holding": fields[5] if len(fields) > 5 else None, + } + body = "\n".join(lines[1:]).strip() + decision_match = self._DECISION_RE.search(body) + reflection_match = self._REFLECTION_RE.search(body) + entry["decision"] = decision_match.group(1).strip() if decision_match else "" + entry["reflection"] = reflection_match.group(1).strip() if reflection_match else "" + return entry + + def _format_full(self, e: dict) -> str: + raw = e["raw"] or "n/a" + alpha = e["alpha"] or "n/a" + holding = e["holding"] or "n/a" + tag = f"[{e['date']} | {e['ticker']} | {e['rating']} | {raw} | {alpha} | {holding}]" + parts = [tag, f"DECISION:\n{e['decision']}"] + if e["reflection"]: + parts.append(f"REFLECTION:\n{e['reflection']}") + return "\n\n".join(parts) + + def _format_reflection_only(self, e: dict) -> str: + tag = f"[{e['date']} | {e['ticker']} | {e['rating']} | {e['raw'] or 'n/a'}]" + if e["reflection"]: + return f"{tag}\n{e['reflection']}" + text = e["decision"][:300] + suffix = "..." if len(e["decision"]) > 300 else "" + return f"{tag}\n{text}{suffix}" diff --git a/tradingagents/default_config.py b/tradingagents/default_config.py index a9b75e4be..274c29a51 100644 --- a/tradingagents/default_config.py +++ b/tradingagents/default_config.py @@ -6,6 +6,7 @@ DEFAULT_CONFIG = { "project_dir": os.path.abspath(os.path.join(os.path.dirname(__file__), ".")), "results_dir": os.getenv("TRADINGAGENTS_RESULTS_DIR", os.path.join(_TRADINGAGENTS_HOME, "logs")), "data_cache_dir": os.getenv("TRADINGAGENTS_CACHE_DIR", os.path.join(_TRADINGAGENTS_HOME, "cache")), + "memory_log_path": os.path.join(_TRADINGAGENTS_HOME, "memory", "trading_memory.md"), # LLM settings "llm_provider": "openai", "deep_think_llm": "gpt-5.4", diff --git a/tradingagents/graph/propagation.py b/tradingagents/graph/propagation.py index 0fd10c0c3..2a5efb1ba 100644 --- a/tradingagents/graph/propagation.py +++ b/tradingagents/graph/propagation.py @@ -16,13 +16,14 @@ class Propagator: self.max_recur_limit = max_recur_limit def create_initial_state( - self, company_name: str, trade_date: str + self, company_name: str, trade_date: str, past_context: str = "" ) -> Dict[str, Any]: """Create the initial state for the agent graph.""" return { "messages": [("human", company_name)], "company_of_interest": company_name, "trade_date": str(trade_date), + "past_context": past_context, "investment_debate_state": InvestDebateState( { "bull_history": "", diff --git a/tradingagents/graph/reflection.py b/tradingagents/graph/reflection.py index 2a680038c..813114428 100644 --- a/tradingagents/graph/reflection.py +++ b/tradingagents/graph/reflection.py @@ -1,120 +1,53 @@ # TradingAgents/graph/reflection.py -from typing import Any, Dict +from typing import Any class Reflector: - """Handles reflection on decisions and updating memory.""" + """Handles reflection on trading decisions.""" def __init__(self, quick_thinking_llm: Any): """Initialize the reflector with an LLM.""" self.quick_thinking_llm = quick_thinking_llm - self.reflection_system_prompt = self._get_reflection_prompt() + self.log_reflection_prompt = self._get_log_reflection_prompt() - def _get_reflection_prompt(self) -> str: - """Get the system prompt for reflection.""" - return """ -You are an expert financial analyst tasked with reviewing trading decisions/analysis and providing a comprehensive, step-by-step analysis. -Your goal is to deliver detailed insights into investment decisions and highlight opportunities for improvement, adhering strictly to the following guidelines: + def _get_log_reflection_prompt(self) -> str: + """Concise prompt for reflect_on_final_decision (Phase B log entries). -1. Reasoning: - - For each trading decision, determine whether it was correct or incorrect. A correct decision results in an increase in returns, while an incorrect decision does the opposite. - - Analyze the contributing factors to each success or mistake. Consider: - - Market intelligence. - - Technical indicators. - - Technical signals. - - Price movement analysis. - - Overall market data analysis - - News analysis. - - Social media and sentiment analysis. - - Fundamental data analysis. - - Weight the importance of each factor in the decision-making process. + Produces 2-4 sentences of plain prose — compact enough to be re-injected + into future agent prompts without bloating the context window. + """ + return ( + "You are a trading analyst reviewing your own past decision now that the outcome is known.\n" + "Write exactly 2-4 sentences of plain prose (no bullets, no headers, no markdown).\n\n" + "Cover in order:\n" + "1. Was the directional call correct? (cite the alpha figure)\n" + "2. Which part of the investment thesis held or failed?\n" + "3. One concrete lesson to apply to the next similar analysis.\n\n" + "Be specific and terse. Your output will be stored verbatim in a decision log " + "and re-read by future analysts, so every word must earn its place." + ) -2. Improvement: - - For any incorrect decisions, propose revisions to maximize returns. - - Provide a detailed list of corrective actions or improvements, including specific recommendations (e.g., changing a decision from HOLD to BUY on a particular date). - -3. Summary: - - Summarize the lessons learned from the successes and mistakes. - - Highlight how these lessons can be adapted for future trading scenarios and draw connections between similar situations to apply the knowledge gained. - -4. Query: - - Extract key insights from the summary into a concise sentence of no more than 1000 tokens. - - Ensure the condensed sentence captures the essence of the lessons and reasoning for easy reference. - -Adhere strictly to these instructions, and ensure your output is detailed, accurate, and actionable. You will also be given objective descriptions of the market from a price movements, technical indicator, news, and sentiment perspective to provide more context for your analysis. -""" - - def _extract_current_situation(self, current_state: Dict[str, Any]) -> str: - """Extract the current market situation from the state.""" - curr_market_report = current_state["market_report"] - curr_sentiment_report = current_state["sentiment_report"] - curr_news_report = current_state["news_report"] - curr_fundamentals_report = current_state["fundamentals_report"] - - return f"{curr_market_report}\n\n{curr_sentiment_report}\n\n{curr_news_report}\n\n{curr_fundamentals_report}" - - def _reflect_on_component( - self, component_type: str, report: str, situation: str, returns_losses + def reflect_on_final_decision( + self, + final_decision: str, + raw_return: float, + alpha_return: float, ) -> str: - """Generate reflection for a component.""" + """Single reflection call on the final trade decision with outcome context. + + Used by Phase B deferred reflection. The final_trade_decision already + synthesises all analyst insights, so no separate market context is needed. + """ messages = [ - ("system", self.reflection_system_prompt), + ("system", self.log_reflection_prompt), ( "human", - f"Returns: {returns_losses}\n\nAnalysis/Decision: {report}\n\nObjective Market Reports for Reference: {situation}", + ( + f"Raw return: {raw_return:+.1%}\n" + f"Alpha vs SPY: {alpha_return:+.1%}\n\n" + f"Final Decision:\n{final_decision}" + ), ), ] - - result = self.quick_thinking_llm.invoke(messages).content - return result - - def reflect_bull_researcher(self, current_state, returns_losses, bull_memory): - """Reflect on bull researcher's analysis and update memory.""" - situation = self._extract_current_situation(current_state) - bull_debate_history = current_state["investment_debate_state"]["bull_history"] - - result = self._reflect_on_component( - "BULL", bull_debate_history, situation, returns_losses - ) - bull_memory.add_situations([(situation, result)]) - - def reflect_bear_researcher(self, current_state, returns_losses, bear_memory): - """Reflect on bear researcher's analysis and update memory.""" - situation = self._extract_current_situation(current_state) - bear_debate_history = current_state["investment_debate_state"]["bear_history"] - - result = self._reflect_on_component( - "BEAR", bear_debate_history, situation, returns_losses - ) - bear_memory.add_situations([(situation, result)]) - - def reflect_trader(self, current_state, returns_losses, trader_memory): - """Reflect on trader's decision and update memory.""" - situation = self._extract_current_situation(current_state) - trader_decision = current_state["trader_investment_plan"] - - result = self._reflect_on_component( - "TRADER", trader_decision, situation, returns_losses - ) - trader_memory.add_situations([(situation, result)]) - - def reflect_invest_judge(self, current_state, returns_losses, invest_judge_memory): - """Reflect on investment judge's decision and update memory.""" - situation = self._extract_current_situation(current_state) - judge_decision = current_state["investment_debate_state"]["judge_decision"] - - result = self._reflect_on_component( - "INVEST JUDGE", judge_decision, situation, returns_losses - ) - invest_judge_memory.add_situations([(situation, result)]) - - def reflect_portfolio_manager(self, current_state, returns_losses, portfolio_manager_memory): - """Reflect on portfolio manager's decision and update memory.""" - situation = self._extract_current_situation(current_state) - judge_decision = current_state["risk_debate_state"]["judge_decision"] - - result = self._reflect_on_component( - "PORTFOLIO MANAGER", judge_decision, situation, returns_losses - ) - portfolio_manager_memory.add_situations([(situation, result)]) + return self.quick_thinking_llm.invoke(messages).content diff --git a/tradingagents/graph/setup.py b/tradingagents/graph/setup.py index ae90489c1..1686fc5b0 100644 --- a/tradingagents/graph/setup.py +++ b/tradingagents/graph/setup.py @@ -18,22 +18,12 @@ class GraphSetup: quick_thinking_llm: Any, deep_thinking_llm: Any, tool_nodes: Dict[str, ToolNode], - bull_memory, - bear_memory, - trader_memory, - invest_judge_memory, - portfolio_manager_memory, conditional_logic: ConditionalLogic, ): """Initialize with required components.""" self.quick_thinking_llm = quick_thinking_llm self.deep_thinking_llm = deep_thinking_llm self.tool_nodes = tool_nodes - self.bull_memory = bull_memory - self.bear_memory = bear_memory - self.trader_memory = trader_memory - self.invest_judge_memory = invest_judge_memory - self.portfolio_manager_memory = portfolio_manager_memory self.conditional_logic = conditional_logic def setup_graph( @@ -85,24 +75,16 @@ class GraphSetup: tool_nodes["fundamentals"] = self.tool_nodes["fundamentals"] # Create researcher and manager nodes - bull_researcher_node = create_bull_researcher( - self.quick_thinking_llm, self.bull_memory - ) - bear_researcher_node = create_bear_researcher( - self.quick_thinking_llm, self.bear_memory - ) - research_manager_node = create_research_manager( - self.deep_thinking_llm, self.invest_judge_memory - ) - trader_node = create_trader(self.quick_thinking_llm, self.trader_memory) + bull_researcher_node = create_bull_researcher(self.quick_thinking_llm) + bear_researcher_node = create_bear_researcher(self.quick_thinking_llm) + research_manager_node = create_research_manager(self.deep_thinking_llm) + trader_node = create_trader(self.quick_thinking_llm) # Create risk analysis nodes aggressive_analyst = create_aggressive_debator(self.quick_thinking_llm) neutral_analyst = create_neutral_debator(self.quick_thinking_llm) conservative_analyst = create_conservative_debator(self.quick_thinking_llm) - portfolio_manager_node = create_portfolio_manager( - self.deep_thinking_llm, self.portfolio_manager_memory - ) + portfolio_manager_node = create_portfolio_manager(self.deep_thinking_llm) # Create workflow workflow = StateGraph(AgentState) diff --git a/tradingagents/graph/trading_graph.py b/tradingagents/graph/trading_graph.py index 78bc13e5f..cf5662598 100644 --- a/tradingagents/graph/trading_graph.py +++ b/tradingagents/graph/trading_graph.py @@ -1,18 +1,23 @@ # TradingAgents/graph/trading_graph.py +import logging import os from pathlib import Path import json -from datetime import date +from datetime import datetime, timedelta from typing import Dict, Any, Tuple, List, Optional +import yfinance as yf + +logger = logging.getLogger(__name__) + from langgraph.prebuilt import ToolNode from tradingagents.llm_clients import create_llm_client from tradingagents.agents import * from tradingagents.default_config import DEFAULT_CONFIG -from tradingagents.agents.utils.memory import FinancialSituationMemory +from tradingagents.agents.utils.memory import TradingMemoryLog from tradingagents.agents.utils.agent_states import ( AgentState, InvestDebateState, @@ -92,12 +97,7 @@ class TradingAgentsGraph: self.deep_thinking_llm = deep_client.get_llm() self.quick_thinking_llm = quick_client.get_llm() - # Initialize memories - self.bull_memory = FinancialSituationMemory("bull_memory", self.config) - self.bear_memory = FinancialSituationMemory("bear_memory", self.config) - self.trader_memory = FinancialSituationMemory("trader_memory", self.config) - self.invest_judge_memory = FinancialSituationMemory("invest_judge_memory", self.config) - self.portfolio_manager_memory = FinancialSituationMemory("portfolio_manager_memory", self.config) + self.memory_log = TradingMemoryLog(self.config) # Create tool nodes self.tool_nodes = self._create_tool_nodes() @@ -111,11 +111,6 @@ class TradingAgentsGraph: self.quick_thinking_llm, self.deep_thinking_llm, self.tool_nodes, - self.bull_memory, - self.bear_memory, - self.trader_memory, - self.invest_judge_memory, - self.portfolio_manager_memory, self.conditional_logic, ) @@ -189,14 +184,90 @@ class TradingAgentsGraph: ), } + def _fetch_returns( + self, ticker: str, trade_date: str, holding_days: int = 5 + ) -> Tuple[Optional[float], Optional[float], Optional[int]]: + """Fetch raw and alpha return for ticker over holding_days from trade_date. + + Returns (raw_return, alpha_return, actual_holding_days) or + (None, None, None) if price data is unavailable (too recent, delisted, + or network error). + """ + try: + start = datetime.strptime(trade_date, "%Y-%m-%d") + end = start + timedelta(days=holding_days + 7) # buffer for weekends/holidays + end_str = end.strftime("%Y-%m-%d") + + stock = yf.Ticker(ticker).history(start=trade_date, end=end_str) + spy = yf.Ticker("SPY").history(start=trade_date, end=end_str) + + if len(stock) < 2 or len(spy) < 2: + return None, None, None + + actual_days = min(holding_days, len(stock) - 1, len(spy) - 1) + raw = float( + (stock["Close"].iloc[actual_days] - stock["Close"].iloc[0]) + / stock["Close"].iloc[0] + ) + spy_ret = float( + (spy["Close"].iloc[actual_days] - spy["Close"].iloc[0]) + / spy["Close"].iloc[0] + ) + alpha = raw - spy_ret + return raw, alpha, actual_days + except Exception as e: + logger.debug("_fetch_returns failed for %s@%s: %s", ticker, trade_date, e) + return None, None, None + + def _resolve_pending_entries(self, ticker: str) -> None: + """Resolve pending log entries for ticker at the start of a new run. + + Fetches returns for each same-ticker pending entry, generates reflections, + then writes all updates in a single atomic batch write to avoid redundant I/O. + Skips entries whose price data is not yet available (too recent or delisted). + + Trade-off: only same-ticker entries are resolved per run. Entries for + other tickers accumulate until that ticker is run again. + """ + pending = [e for e in self.memory_log.get_pending_entries() if e["ticker"] == ticker] + if not pending: + return + + updates = [] + for entry in pending: + raw, alpha, days = self._fetch_returns(ticker, entry["date"]) + if raw is None: + continue # price not available yet — try again next run + reflection = self.reflector.reflect_on_final_decision( + final_decision=entry.get("decision", ""), + raw_return=raw, + alpha_return=alpha, + ) + updates.append({ + "ticker": ticker, + "trade_date": entry["date"], + "raw_return": raw, + "alpha_return": alpha, + "holding_days": days, + "reflection": reflection, + }) + + if updates: + self.memory_log.batch_update_with_outcomes(updates) + def propagate(self, company_name, trade_date): """Run the trading agents graph for a company on a specific date.""" self.ticker = company_name - # Initialize state + # Resolve any pending log entries for this ticker before the pipeline runs. + # This adds the outcome + reflection from the previous run at zero latency cost. + self._resolve_pending_entries(company_name) + + # Initialize state — inject memory log context for PM + past_context = self.memory_log.get_past_context(company_name) init_agent_state = self.propagator.create_initial_state( - company_name, trade_date + company_name, trade_date, past_context=past_context ) args = self.propagator.get_graph_args() @@ -221,6 +292,13 @@ class TradingAgentsGraph: # Log state self._log_state(trade_date, final_state) + # Store decision for deferred reflection. + self.memory_log.store_decision( + ticker=company_name, + trade_date=trade_date, + final_trade_decision=final_state["final_trade_decision"], + ) + # Return decision and processed signal return final_state, self.process_signal(final_state["final_trade_decision"]) @@ -264,24 +342,6 @@ class TradingAgentsGraph: with open(log_path, "w", encoding="utf-8") as f: json.dump(self.log_states_dict[str(trade_date)], f, indent=4) - def reflect_and_remember(self, returns_losses): - """Reflect on decisions and update memory based on returns.""" - self.reflector.reflect_bull_researcher( - self.curr_state, returns_losses, self.bull_memory - ) - self.reflector.reflect_bear_researcher( - self.curr_state, returns_losses, self.bear_memory - ) - self.reflector.reflect_trader( - self.curr_state, returns_losses, self.trader_memory - ) - self.reflector.reflect_invest_judge( - self.curr_state, returns_losses, self.invest_judge_memory - ) - self.reflector.reflect_portfolio_manager( - self.curr_state, returns_losses, self.portfolio_manager_memory - ) - def process_signal(self, full_signal): """Process a signal to extract the core decision.""" return self.signal_processor.process_signal(full_signal) From 872b063e6917929335bafc20ef1070fba9e54a69 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 07:25:32 +0000 Subject: [PATCH 03/12] fix: use explicit encoding="utf-8" for all file I/O so Windows users avoid cp1252 crashes (#543, #550, #576) --- cli/main.py | 34 ++++++++++----------- tradingagents/__init__.py | 2 -- tradingagents/dataflows/stockstats_utils.py | 4 +-- tradingagents/dataflows/utils.py | 2 +- 4 files changed, 20 insertions(+), 22 deletions(-) diff --git a/cli/main.py b/cli/main.py index 33d110fb6..6e838fc8b 100644 --- a/cli/main.py +++ b/cli/main.py @@ -463,7 +463,7 @@ def update_display(layout, spinner_text=None, stats_handler=None, start_time=Non def get_user_selections(): """Get all user selections before starting the analysis display.""" # Display ASCII art welcome message - with open(Path(__file__).parent / "static" / "welcome.txt", "r") as f: + with open(Path(__file__).parent / "static" / "welcome.txt", "r", encoding="utf-8") as f: welcome_ascii = f.read() # Create welcome box content @@ -646,19 +646,19 @@ def save_report_to_disk(final_state, ticker: str, save_path: Path): analyst_parts = [] if final_state.get("market_report"): analysts_dir.mkdir(exist_ok=True) - (analysts_dir / "market.md").write_text(final_state["market_report"]) + (analysts_dir / "market.md").write_text(final_state["market_report"], encoding="utf-8") analyst_parts.append(("Market Analyst", final_state["market_report"])) if final_state.get("sentiment_report"): analysts_dir.mkdir(exist_ok=True) - (analysts_dir / "sentiment.md").write_text(final_state["sentiment_report"]) + (analysts_dir / "sentiment.md").write_text(final_state["sentiment_report"], encoding="utf-8") analyst_parts.append(("Social Analyst", final_state["sentiment_report"])) if final_state.get("news_report"): analysts_dir.mkdir(exist_ok=True) - (analysts_dir / "news.md").write_text(final_state["news_report"]) + (analysts_dir / "news.md").write_text(final_state["news_report"], encoding="utf-8") analyst_parts.append(("News Analyst", final_state["news_report"])) if final_state.get("fundamentals_report"): analysts_dir.mkdir(exist_ok=True) - (analysts_dir / "fundamentals.md").write_text(final_state["fundamentals_report"]) + (analysts_dir / "fundamentals.md").write_text(final_state["fundamentals_report"], encoding="utf-8") analyst_parts.append(("Fundamentals Analyst", final_state["fundamentals_report"])) if analyst_parts: content = "\n\n".join(f"### {name}\n{text}" for name, text in analyst_parts) @@ -671,15 +671,15 @@ def save_report_to_disk(final_state, ticker: str, save_path: Path): research_parts = [] if debate.get("bull_history"): research_dir.mkdir(exist_ok=True) - (research_dir / "bull.md").write_text(debate["bull_history"]) + (research_dir / "bull.md").write_text(debate["bull_history"], encoding="utf-8") research_parts.append(("Bull Researcher", debate["bull_history"])) if debate.get("bear_history"): research_dir.mkdir(exist_ok=True) - (research_dir / "bear.md").write_text(debate["bear_history"]) + (research_dir / "bear.md").write_text(debate["bear_history"], encoding="utf-8") research_parts.append(("Bear Researcher", debate["bear_history"])) if debate.get("judge_decision"): research_dir.mkdir(exist_ok=True) - (research_dir / "manager.md").write_text(debate["judge_decision"]) + (research_dir / "manager.md").write_text(debate["judge_decision"], encoding="utf-8") research_parts.append(("Research Manager", debate["judge_decision"])) if research_parts: content = "\n\n".join(f"### {name}\n{text}" for name, text in research_parts) @@ -689,7 +689,7 @@ def save_report_to_disk(final_state, ticker: str, save_path: Path): if final_state.get("trader_investment_plan"): trading_dir = save_path / "3_trading" trading_dir.mkdir(exist_ok=True) - (trading_dir / "trader.md").write_text(final_state["trader_investment_plan"]) + (trading_dir / "trader.md").write_text(final_state["trader_investment_plan"], encoding="utf-8") sections.append(f"## III. Trading Team Plan\n\n### Trader\n{final_state['trader_investment_plan']}") # 4. Risk Management @@ -699,15 +699,15 @@ def save_report_to_disk(final_state, ticker: str, save_path: Path): risk_parts = [] if risk.get("aggressive_history"): risk_dir.mkdir(exist_ok=True) - (risk_dir / "aggressive.md").write_text(risk["aggressive_history"]) + (risk_dir / "aggressive.md").write_text(risk["aggressive_history"], encoding="utf-8") risk_parts.append(("Aggressive Analyst", risk["aggressive_history"])) if risk.get("conservative_history"): risk_dir.mkdir(exist_ok=True) - (risk_dir / "conservative.md").write_text(risk["conservative_history"]) + (risk_dir / "conservative.md").write_text(risk["conservative_history"], encoding="utf-8") risk_parts.append(("Conservative Analyst", risk["conservative_history"])) if risk.get("neutral_history"): risk_dir.mkdir(exist_ok=True) - (risk_dir / "neutral.md").write_text(risk["neutral_history"]) + (risk_dir / "neutral.md").write_text(risk["neutral_history"], encoding="utf-8") risk_parts.append(("Neutral Analyst", risk["neutral_history"])) if risk_parts: content = "\n\n".join(f"### {name}\n{text}" for name, text in risk_parts) @@ -717,12 +717,12 @@ def save_report_to_disk(final_state, ticker: str, save_path: Path): if risk.get("judge_decision"): portfolio_dir = save_path / "5_portfolio" portfolio_dir.mkdir(exist_ok=True) - (portfolio_dir / "decision.md").write_text(risk["judge_decision"]) + (portfolio_dir / "decision.md").write_text(risk["judge_decision"], encoding="utf-8") sections.append(f"## V. Portfolio Manager Decision\n\n### Portfolio Manager\n{risk['judge_decision']}") # Write consolidated report header = f"# Trading Analysis Report: {ticker}\n\nGenerated: {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n\n" - (save_path / "complete_report.md").write_text(header + "\n\n".join(sections)) + (save_path / "complete_report.md").write_text(header + "\n\n".join(sections), encoding="utf-8") return save_path / "complete_report.md" @@ -980,7 +980,7 @@ def run_analysis(): func(*args, **kwargs) timestamp, message_type, content = obj.messages[-1] content = content.replace("\n", " ") # Replace newlines with spaces - with open(log_file, "a") as f: + with open(log_file, "a", encoding="utf-8") as f: f.write(f"{timestamp} [{message_type}] {content}\n") return wrapper @@ -991,7 +991,7 @@ def run_analysis(): func(*args, **kwargs) timestamp, tool_name, args = obj.tool_calls[-1] args_str = ", ".join(f"{k}={v}" for k, v in args.items()) - with open(log_file, "a") as f: + with open(log_file, "a", encoding="utf-8") as f: f.write(f"{timestamp} [Tool Call] {tool_name}({args_str})\n") return wrapper @@ -1005,7 +1005,7 @@ def run_analysis(): if content: file_name = f"{section_name}.md" text = "\n".join(str(item) for item in content) if isinstance(content, list) else content - with open(report_dir / file_name, "w") as f: + with open(report_dir / file_name, "w", encoding="utf-8") as f: f.write(text) return wrapper diff --git a/tradingagents/__init__.py b/tradingagents/__init__.py index 43a2b4398..e69de29bb 100644 --- a/tradingagents/__init__.py +++ b/tradingagents/__init__.py @@ -1,2 +0,0 @@ -import os -os.environ.setdefault("PYTHONUTF8", "1") diff --git a/tradingagents/dataflows/stockstats_utils.py b/tradingagents/dataflows/stockstats_utils.py index 507478830..cb24c5d6a 100644 --- a/tradingagents/dataflows/stockstats_utils.py +++ b/tradingagents/dataflows/stockstats_utils.py @@ -67,7 +67,7 @@ def load_ohlcv(symbol: str, curr_date: str) -> pd.DataFrame: ) if os.path.exists(data_file): - data = pd.read_csv(data_file, on_bad_lines="skip") + data = pd.read_csv(data_file, on_bad_lines="skip", encoding="utf-8") else: data = yf_retry(lambda: yf.download( symbol, @@ -78,7 +78,7 @@ def load_ohlcv(symbol: str, curr_date: str) -> pd.DataFrame: auto_adjust=True, )) data = data.reset_index() - data.to_csv(data_file, index=False) + data.to_csv(data_file, index=False, encoding="utf-8") data = _clean_dataframe(data) diff --git a/tradingagents/dataflows/utils.py b/tradingagents/dataflows/utils.py index 4523de19f..c99b777ab 100644 --- a/tradingagents/dataflows/utils.py +++ b/tradingagents/dataflows/utils.py @@ -8,7 +8,7 @@ SavePathType = Annotated[str, "File path to save data. If None, data is not save def save_output(data: pd.DataFrame, tag: str, save_path: SavePathType = None) -> None: if save_path: - data.to_csv(save_path) + data.to_csv(save_path, encoding="utf-8") print(f"{tag} saved to {save_path}") From 8e7654f0df735e89436de3adb1eab53ba33afeb8 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 07:41:36 +0000 Subject: [PATCH 04/12] fix: drop past-memory directive and placeholder from agent prompts when memory is empty (#572) --- .../agents/managers/portfolio_manager.py | 16 +++++++++++++--- .../agents/managers/research_manager.py | 13 +++++++------ .../agents/researchers/bear_researcher.py | 14 ++++++++++++-- .../agents/researchers/bull_researcher.py | 14 ++++++++++++-- tradingagents/agents/trader/trader.py | 10 +++++++--- 5 files changed, 51 insertions(+), 16 deletions(-) diff --git a/tradingagents/agents/managers/portfolio_manager.py b/tradingagents/agents/managers/portfolio_manager.py index 6c69ae9fd..5d4631b82 100644 --- a/tradingagents/agents/managers/portfolio_manager.py +++ b/tradingagents/agents/managers/portfolio_manager.py @@ -22,6 +22,17 @@ def create_portfolio_manager(llm, memory): for i, rec in enumerate(past_memories, 1): past_memory_str += rec["recommendation"] + "\n\n" + lessons_line = ( + f"- Lessons from past decisions: **{past_memory_str.strip()}**\n" + if past_memories + else "" + ) + thesis_instruction = ( + "3. **Investment Thesis**: Detailed reasoning anchored in the analysts' debate and past reflections." + if past_memories + else "3. **Investment Thesis**: Detailed reasoning anchored in the analysts' debate." + ) + prompt = f"""As the Portfolio Manager, synthesize the risk analysts' debate and deliver the final trading decision. {instrument_context} @@ -38,12 +49,11 @@ def create_portfolio_manager(llm, memory): **Context:** - Research Manager's investment plan: **{research_plan}** - Trader's transaction proposal: **{trader_plan}** -- Lessons from past decisions: **{past_memory_str}** - +{lessons_line} **Required Output Structure:** 1. **Rating**: State one of Buy / Overweight / Hold / Underweight / Sell. 2. **Executive Summary**: A concise action plan covering entry strategy, position sizing, key risk levels, and time horizon. -3. **Investment Thesis**: Detailed reasoning anchored in the analysts' debate and past reflections. +{thesis_instruction} --- diff --git a/tradingagents/agents/managers/research_manager.py b/tradingagents/agents/managers/research_manager.py index 5b4b4fdc5..03e32e492 100644 --- a/tradingagents/agents/managers/research_manager.py +++ b/tradingagents/agents/managers/research_manager.py @@ -20,6 +20,12 @@ def create_research_manager(llm, memory): for i, rec in enumerate(past_memories, 1): past_memory_str += rec["recommendation"] + "\n\n" + past_memory_block = ( + f'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. \n\nHere are your past reflections on mistakes:\n"{past_memory_str.strip()}"\n\n' + if past_memories + else "" + ) + 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 bear analyst, the bull analyst, or choose 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—Buy, Sell, or Hold—must be clear and actionable. Avoid defaulting to Hold simply because both sides have valid points; commit to a stance grounded in the debate's strongest arguments. @@ -29,12 +35,7 @@ Additionally, develop a detailed investment plan for the trader. This should inc 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. - -Here are your past reflections on mistakes: -\"{past_memory_str}\" - -{instrument_context} +{past_memory_block}{instrument_context} Here is the debate: Debate History: diff --git a/tradingagents/agents/researchers/bear_researcher.py b/tradingagents/agents/researchers/bear_researcher.py index a44212dc4..e3922e8c3 100644 --- a/tradingagents/agents/researchers/bear_researcher.py +++ b/tradingagents/agents/researchers/bear_researcher.py @@ -19,6 +19,17 @@ def create_bear_researcher(llm, memory): for i, rec in enumerate(past_memories, 1): past_memory_str += rec["recommendation"] + "\n\n" + memory_section = ( + f"Reflections from similar situations and lessons learned: {past_memory_str.strip()}\n" + if past_memories + else "" + ) + memory_instruction = ( + " You must also address reflections and learn from lessons and mistakes you made in the past." + if past_memories + else "" + ) + prompt = f"""You are a Bear Analyst making the case against investing in the stock. Your goal is to present a well-reasoned argument emphasizing risks, challenges, and negative indicators. Leverage the provided research and data to highlight potential downsides and counter bullish arguments effectively. Key points to focus on: @@ -37,8 +48,7 @@ Latest world affairs news: {news_report} Company fundamentals report: {fundamentals_report} Conversation history of the debate: {history} Last bull argument: {current_response} -Reflections from similar situations and lessons learned: {past_memory_str} -Use this information to deliver a compelling bear argument, refute the bull's claims, and engage in a dynamic debate that demonstrates the risks and weaknesses of investing in the stock. You must also address reflections and learn from lessons and mistakes you made in the past. +{memory_section}Use this information to deliver a compelling bear argument, refute the bull's claims, and engage in a dynamic debate that demonstrates the risks and weaknesses of investing in the stock.{memory_instruction} """ response = llm.invoke(prompt) diff --git a/tradingagents/agents/researchers/bull_researcher.py b/tradingagents/agents/researchers/bull_researcher.py index d23d4d76e..9724415d6 100644 --- a/tradingagents/agents/researchers/bull_researcher.py +++ b/tradingagents/agents/researchers/bull_researcher.py @@ -19,6 +19,17 @@ def create_bull_researcher(llm, memory): for i, rec in enumerate(past_memories, 1): past_memory_str += rec["recommendation"] + "\n\n" + memory_section = ( + f"Reflections from similar situations and lessons learned: {past_memory_str.strip()}\n" + if past_memories + else "" + ) + memory_instruction = ( + " You must also address reflections and learn from lessons and mistakes you made in the past." + if past_memories + else "" + ) + prompt = f"""You are a Bull Analyst advocating for investing in the stock. Your task is to build a strong, evidence-based case emphasizing growth potential, competitive advantages, and positive market indicators. Leverage the provided research and data to address concerns and counter bearish arguments effectively. Key points to focus on: @@ -35,8 +46,7 @@ Latest world affairs news: {news_report} Company fundamentals report: {fundamentals_report} Conversation history of the debate: {history} Last bear argument: {current_response} -Reflections from similar situations and lessons learned: {past_memory_str} -Use this information to deliver a compelling bull argument, refute the bear's concerns, and engage in a dynamic debate that demonstrates the strengths of the bull position. You must also address reflections and learn from lessons and mistakes you made in the past. +{memory_section}Use this information to deliver a compelling bull argument, refute the bear's concerns, and engage in a dynamic debate that demonstrates the strengths of the bull position.{memory_instruction} """ response = llm.invoke(prompt) diff --git a/tradingagents/agents/trader/trader.py b/tradingagents/agents/trader/trader.py index 07e9f262c..964733467 100644 --- a/tradingagents/agents/trader/trader.py +++ b/tradingagents/agents/trader/trader.py @@ -20,8 +20,12 @@ def create_trader(llm, memory): if past_memories: for i, rec in enumerate(past_memories, 1): past_memory_str += rec["recommendation"] + "\n\n" - else: - past_memory_str = "No past memories found." + + memory_instruction = ( + f" Apply lessons from past decisions to strengthen your analysis. Here are reflections from similar situations you traded in and the lessons learned: {past_memory_str.strip()}" + if past_memories + else "" + ) context = { "role": "user", @@ -31,7 +35,7 @@ def create_trader(llm, memory): messages = [ { "role": "system", - "content": f"""You are a trading agent analyzing market data to make investment decisions. Based on your analysis, provide a specific recommendation to buy, sell, or hold. End with a firm decision and always conclude your response with 'FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**' to confirm your recommendation. Apply lessons from past decisions to strengthen your analysis. Here are reflections from similar situations you traded in and the lessons learned: {past_memory_str}""", + "content": f"You are a trading agent analyzing market data to make investment decisions. Based on your analysis, provide a specific recommendation to buy, sell, or hold. End with a firm decision and always conclude your response with 'FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**' to confirm your recommendation.{memory_instruction}", }, context, ] From f85f5d9f5d9243549fe7860be8611698051fda29 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 07:41:36 +0000 Subject: [PATCH 05/12] test: lazy-load LLM provider clients and add API-key fixtures so the test suite runs cleanly without credentials (#588) --- pyproject.toml | 12 ++++++++ tests/conftest.py | 42 ++++++++++++++++++++++++++++ tests/test_google_api_key.py | 3 ++ tests/test_model_validation.py | 3 ++ tests/test_ticker_symbol_handling.py | 3 ++ tradingagents/llm_clients/factory.py | 12 +++++--- 6 files changed, 71 insertions(+), 4 deletions(-) create mode 100644 tests/conftest.py diff --git a/pyproject.toml b/pyproject.toml index 98385e32e..110611630 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -40,3 +40,15 @@ include = ["tradingagents*", "cli*"] [tool.setuptools.package-data] cli = ["static/*"] + +[tool.pytest.ini_options] +testpaths = ["tests"] +addopts = "-ra --strict-markers" +markers = [ + "unit: fast isolated unit tests", + "integration: tests requiring external services", + "smoke: quick sanity-check tests", +] +filterwarnings = [ + "ignore::DeprecationWarning", +] diff --git a/tests/conftest.py b/tests/conftest.py new file mode 100644 index 000000000..504ffb12d --- /dev/null +++ b/tests/conftest.py @@ -0,0 +1,42 @@ +"""Shared pytest fixtures that prevent CI hangs when API keys are absent.""" + +import os +from unittest.mock import MagicMock, patch + +import pytest + + +def pytest_configure(config): + for marker in ("unit", "integration", "smoke"): + config.addinivalue_line("markers", f"{marker}: {marker}-level tests") + + +_API_KEY_ENV_VARS = ( + "OPENAI_API_KEY", + "GOOGLE_API_KEY", + "ANTHROPIC_API_KEY", + "XAI_API_KEY", + "DEEPSEEK_API_KEY", + "DASHSCOPE_API_KEY", + "ZHIPU_API_KEY", + "OPENROUTER_API_KEY", + "AZURE_OPENAI_API_KEY", + "ALPHA_VANTAGE_API_KEY", +) + + +@pytest.fixture(autouse=True) +def _dummy_api_keys(monkeypatch): + for env_var in _API_KEY_ENV_VARS: + monkeypatch.setenv(env_var, os.environ.get(env_var, "placeholder")) + + +@pytest.fixture() +def mock_llm_client(): + client = MagicMock() + client.get_llm.return_value = MagicMock() + with patch( + "tradingagents.llm_clients.factory.create_llm_client", + return_value=client, + ): + yield client diff --git a/tests/test_google_api_key.py b/tests/test_google_api_key.py index e1607c49a..53376ab10 100644 --- a/tests/test_google_api_key.py +++ b/tests/test_google_api_key.py @@ -1,9 +1,12 @@ import unittest from unittest.mock import patch +import pytest + from tradingagents.llm_clients.google_client import GoogleClient +@pytest.mark.unit class TestGoogleApiKeyStandardization(unittest.TestCase): """Verify GoogleClient accepts unified api_key parameter.""" diff --git a/tests/test_model_validation.py b/tests/test_model_validation.py index 50f263182..5392d7cd9 100644 --- a/tests/test_model_validation.py +++ b/tests/test_model_validation.py @@ -1,6 +1,8 @@ import unittest import warnings +import pytest + from tradingagents.llm_clients.base_client import BaseLLMClient from tradingagents.llm_clients.model_catalog import get_known_models from tradingagents.llm_clients.validators import validate_model @@ -19,6 +21,7 @@ class DummyLLMClient(BaseLLMClient): return validate_model(self.provider, self.model) +@pytest.mark.unit class ModelValidationTests(unittest.TestCase): def test_cli_catalog_models_are_all_validator_approved(self): for provider, models in get_known_models().items(): diff --git a/tests/test_ticker_symbol_handling.py b/tests/test_ticker_symbol_handling.py index 858d26cd5..7fbe5315d 100644 --- a/tests/test_ticker_symbol_handling.py +++ b/tests/test_ticker_symbol_handling.py @@ -1,9 +1,12 @@ import unittest +import pytest + from cli.utils import normalize_ticker_symbol from tradingagents.agents.utils.agent_utils import build_instrument_context +@pytest.mark.unit class TickerSymbolHandlingTests(unittest.TestCase): def test_normalize_ticker_symbol_preserves_exchange_suffix(self): self.assertEqual(normalize_ticker_symbol(" cnc.to "), "CNC.TO") diff --git a/tradingagents/llm_clients/factory.py b/tradingagents/llm_clients/factory.py index a9a7e83d8..e1d24557e 100644 --- a/tradingagents/llm_clients/factory.py +++ b/tradingagents/llm_clients/factory.py @@ -1,10 +1,6 @@ from typing import Optional from .base_client import BaseLLMClient -from .openai_client import OpenAIClient -from .anthropic_client import AnthropicClient -from .google_client import GoogleClient -from .azure_client import AzureOpenAIClient # Providers that use the OpenAI-compatible chat completions API _OPENAI_COMPATIBLE = ( @@ -20,6 +16,10 @@ def create_llm_client( ) -> BaseLLMClient: """Create an LLM client for the specified provider. + Provider modules are imported lazily so that simply importing this + factory (e.g. during test collection) does not pull in heavy LLM SDKs + or fail when their API keys are absent. + Args: provider: LLM provider name model: Model name/identifier @@ -35,15 +35,19 @@ def create_llm_client( provider_lower = provider.lower() if provider_lower in _OPENAI_COMPATIBLE: + from .openai_client import OpenAIClient return OpenAIClient(model, base_url, provider=provider_lower, **kwargs) if provider_lower == "anthropic": + from .anthropic_client import AnthropicClient return AnthropicClient(model, base_url, **kwargs) if provider_lower == "google": + from .google_client import GoogleClient return GoogleClient(model, base_url, **kwargs) if provider_lower == "azure": + from .azure_client import AzureOpenAIClient return AzureOpenAIClient(model, base_url, **kwargs) raise ValueError(f"Unsupported LLM provider: {provider}") From 4cbd4b086fd94074d79987c8ca31daec8d33902c Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 08:39:27 +0000 Subject: [PATCH 06/12] feat: add LangGraph checkpoint resume for crash recovery (#594) Long analyses can take many minutes; a crash or interruption forced users to re-run from scratch and re-pay every LLM call. This adds an opt-in checkpoint layer backed by per-ticker SQLite databases so the graph resumes from the last successful node. How to use: - CLI: tradingagents analyze --checkpoint - CLI: tradingagents analyze --clear-checkpoints - Python: config["checkpoint_enabled"] = True Lifecycle: - propagate() recompiles the graph with a SqliteSaver when enabled and injects a deterministic thread_id derived from ticker+date so the same ticker+date resumes while a different date starts fresh. - On successful completion the per-thread checkpoint rows are cleared. - The context manager is closed in a try/finally so a crash never leaks the SQLite connection or leaves the graph in checkpoint mode. Storage: ~/.tradingagents/cache/checkpoints/.db (override via TRADINGAGENTS_CACHE_DIR). The checkpointer module is new (tradingagents/graph/checkpointer.py) and the GraphSetup now returns the uncompiled workflow so it can be recompiled with a saver when needed. Adds langgraph-checkpoint-sqlite>=2.0.0 dependency. 3 new tests verify the crash/resume cycle and that a different date starts fresh. --- README.md | 32 +++++- cli/main.py | 22 +++- pyproject.toml | 1 + tests/test_checkpoint_resume.py | 147 +++++++++++++++++++++++++++ tests/test_memory_log.py | 9 +- tradingagents/default_config.py | 3 + tradingagents/graph/checkpointer.py | 86 ++++++++++++++++ tradingagents/graph/setup.py | 3 +- tradingagents/graph/trading_graph.py | 67 +++++++++--- 9 files changed, 349 insertions(+), 21 deletions(-) create mode 100644 tests/test_checkpoint_resume.py create mode 100644 tradingagents/graph/checkpointer.py diff --git a/README.md b/README.md index 97cbde486..6c8f644ec 100644 --- a/README.md +++ b/README.md @@ -183,7 +183,7 @@ An interface will appear showing results as they load, letting you track the age ### Implementation Details -We built TradingAgents with LangGraph to ensure flexibility and modularity. The framework supports multiple LLM providers: OpenAI, Google, Anthropic, xAI, OpenRouter, and Ollama. +We built TradingAgents with LangGraph to ensure flexibility and modularity. The framework supports multiple LLM providers: OpenAI, Google, Anthropic, xAI, DeepSeek, Qwen (Alibaba DashScope), GLM (Zhipu), OpenRouter, Ollama for local models, and Azure OpenAI for enterprise. ### Python Usage @@ -207,7 +207,7 @@ from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.default_config import DEFAULT_CONFIG config = DEFAULT_CONFIG.copy() -config["llm_provider"] = "openai" # openai, google, anthropic, xai, openrouter, ollama +config["llm_provider"] = "openai" # openai, google, anthropic, xai, deepseek, qwen, glm, openrouter, ollama, azure config["deep_think_llm"] = "gpt-5.4" # Model for complex reasoning config["quick_think_llm"] = "gpt-5.4-mini" # Model for quick tasks config["max_debate_rounds"] = 2 @@ -219,6 +219,34 @@ print(decision) See `tradingagents/default_config.py` for all configuration options. +## Persistence and Recovery + +TradingAgents persists two kinds of state across runs. + +### Decision log + +The decision log is always on. Each completed run appends its decision to `~/.tradingagents/memory/trading_memory.md`. On the next run for the same ticker, TradingAgents fetches the realised return (raw and alpha vs SPY), generates a one-paragraph reflection, and injects the most recent same-ticker decisions plus recent cross-ticker lessons into the Portfolio Manager prompt, so each analysis carries forward what worked and what didn't. + +Override the path with `TRADINGAGENTS_MEMORY_LOG_PATH`. + +### Checkpoint resume + +Checkpoint resume is opt-in via `--checkpoint`. When enabled, LangGraph saves state after each node so a crashed or interrupted run resumes from the last successful step instead of starting over. On a resume run you will see `Resuming from step N for on ` in the logs; on a new run you will see `Starting fresh`. Checkpoints are cleared automatically on successful completion. + +Per-ticker SQLite databases live at `~/.tradingagents/cache/checkpoints/.db` (override the base with `TRADINGAGENTS_CACHE_DIR`). Use `--clear-checkpoints` to reset all of them before a run. + +```bash +tradingagents analyze --checkpoint # enable for this run +tradingagents analyze --clear-checkpoints # reset before running +``` + +```python +config = DEFAULT_CONFIG.copy() +config["checkpoint_enabled"] = True +ta = TradingAgentsGraph(config=config) +_, decision = ta.propagate("NVDA", "2026-01-15") +``` + ## Contributing We welcome contributions from the community! Whether it's fixing a bug, improving documentation, or suggesting a new feature, your input helps make this project better. If you are interested in this line of research, please consider joining our open-source financial AI research community [Tauric Research](https://tauric.ai/). diff --git a/cli/main.py b/cli/main.py index 6e838fc8b..534f50379 100644 --- a/cli/main.py +++ b/cli/main.py @@ -926,7 +926,7 @@ def format_tool_args(args, max_length=80) -> str: return result[:max_length - 3] + "..." return result -def run_analysis(): +def run_analysis(checkpoint: bool = False): # First get all user selections selections = get_user_selections() @@ -943,6 +943,7 @@ def run_analysis(): config["openai_reasoning_effort"] = selections.get("openai_reasoning_effort") config["anthropic_effort"] = selections.get("anthropic_effort") config["output_language"] = selections.get("output_language", "English") + config["checkpoint_enabled"] = checkpoint # Create stats callback handler for tracking LLM/tool calls stats_handler = StatsCallbackHandler() @@ -1197,8 +1198,23 @@ def run_analysis(): @app.command() -def analyze(): - run_analysis() +def analyze( + checkpoint: bool = typer.Option( + False, + "--checkpoint", + help="Enable checkpoint/resume: save state after each node so a crashed run can resume.", + ), + clear_checkpoints: bool = typer.Option( + False, + "--clear-checkpoints", + help="Delete all saved checkpoints before running (force fresh start).", + ), +): + if clear_checkpoints: + from tradingagents.graph.checkpointer import clear_all_checkpoints + n = clear_all_checkpoints(DEFAULT_CONFIG["data_cache_dir"]) + console.print(f"[yellow]Cleared {n} checkpoint(s).[/yellow]") + run_analysis(checkpoint=checkpoint) if __name__ == "__main__": diff --git a/pyproject.toml b/pyproject.toml index b3dbc6feb..b569504ef 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -16,6 +16,7 @@ dependencies = [ "langchain-google-genai>=4.0.0", "langchain-openai>=0.3.23", "langgraph>=0.4.8", + "langgraph-checkpoint-sqlite>=2.0.0", "pandas>=2.3.0", "parsel>=1.10.0", "pytz>=2025.2", diff --git a/tests/test_checkpoint_resume.py b/tests/test_checkpoint_resume.py new file mode 100644 index 000000000..6f2692bd8 --- /dev/null +++ b/tests/test_checkpoint_resume.py @@ -0,0 +1,147 @@ +"""Test checkpoint resume: crash mid-analysis, re-run resumes from last node.""" + +import sqlite3 +import tempfile +import unittest +from pathlib import Path +from typing import TypedDict + +from langgraph.checkpoint.sqlite import SqliteSaver +from langgraph.graph import END, StateGraph + +from tradingagents.graph.checkpointer import ( + checkpoint_step, + clear_checkpoint, + get_checkpointer, + has_checkpoint, + thread_id, +) + +# Mutable flag to simulate crash on first run +_should_crash = False + + +class _SimpleState(TypedDict): + count: int + + +def _node_a(state: _SimpleState) -> dict: + return {"count": state["count"] + 1} + + +def _node_b(state: _SimpleState) -> dict: + if _should_crash: + raise RuntimeError("simulated mid-analysis crash") + return {"count": state["count"] + 10} + + +def _build_graph() -> StateGraph: + builder = StateGraph(_SimpleState) + builder.add_node("analyst", _node_a) + builder.add_node("trader", _node_b) + builder.set_entry_point("analyst") + builder.add_edge("analyst", "trader") + builder.add_edge("trader", END) + return builder + + +class TestCheckpointResume(unittest.TestCase): + def setUp(self): + self.tmpdir = tempfile.mkdtemp() + self.ticker = "TEST" + self.date = "2026-04-20" + + def test_crash_and_resume(self): + """Crash at 'trader' node, then resume from checkpoint.""" + global _should_crash + builder = _build_graph() + tid = thread_id(self.ticker, self.date) + cfg = {"configurable": {"thread_id": tid}} + + # Run 1: crash at trader node + _should_crash = True + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + with self.assertRaises(RuntimeError): + graph.invoke({"count": 0}, config=cfg) + + # Checkpoint should exist at step 1 (analyst completed) + self.assertTrue(has_checkpoint(self.tmpdir, self.ticker, self.date)) + step = checkpoint_step(self.tmpdir, self.ticker, self.date) + self.assertEqual(step, 1) + + # Run 2: resume — trader succeeds this time + _should_crash = False + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + result = graph.invoke(None, config=cfg) + + # analyst added 1, trader added 10 → 11 + self.assertEqual(result["count"], 11) + + def test_clear_checkpoint_allows_fresh_start(self): + """After clearing, the graph starts from scratch.""" + global _should_crash + builder = _build_graph() + tid = thread_id(self.ticker, self.date) + cfg = {"configurable": {"thread_id": tid}} + + # Create a checkpoint by crashing + _should_crash = True + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + with self.assertRaises(RuntimeError): + graph.invoke({"count": 0}, config=cfg) + + self.assertTrue(has_checkpoint(self.tmpdir, self.ticker, self.date)) + + # Clear it + clear_checkpoint(self.tmpdir, self.ticker, self.date) + self.assertFalse(has_checkpoint(self.tmpdir, self.ticker, self.date)) + + # Fresh run succeeds from scratch + _should_crash = False + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + result = graph.invoke({"count": 0}, config=cfg) + + self.assertEqual(result["count"], 11) + + + def test_different_date_starts_fresh(self): + """A different date must NOT resume from an existing checkpoint.""" + global _should_crash + builder = _build_graph() + date2 = "2026-04-21" + + # Run with date1 — crash to leave a checkpoint + _should_crash = True + tid1 = thread_id(self.ticker, self.date) + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + with self.assertRaises(RuntimeError): + graph.invoke({"count": 0}, config={"configurable": {"thread_id": tid1}}) + + self.assertTrue(has_checkpoint(self.tmpdir, self.ticker, self.date)) + + # date2 should have no checkpoint + self.assertFalse(has_checkpoint(self.tmpdir, self.ticker, date2)) + + # Run with date2 — should start fresh and succeed + _should_crash = False + tid2 = thread_id(self.ticker, date2) + self.assertNotEqual(tid1, tid2) + + with get_checkpointer(self.tmpdir, self.ticker) as saver: + graph = builder.compile(checkpointer=saver) + result = graph.invoke({"count": 0}, config={"configurable": {"thread_id": tid2}}) + + # Fresh run: analyst +1, trader +10 = 11 + self.assertEqual(result["count"], 11) + + # Original date checkpoint still exists (untouched) + self.assertTrue(has_checkpoint(self.tmpdir, self.ticker, self.date)) + + +if __name__ == "__main__": + unittest.main() diff --git a/tests/test_memory_log.py b/tests/test_memory_log.py index 915bc3b15..ccd1ca7e7 100644 --- a/tests/test_memory_log.py +++ b/tests/test_memory_log.py @@ -629,7 +629,9 @@ class TestLegacyRemoval: create_portfolio_manager(mock_llm, memory=MagicMock()) def test_full_pipeline_no_regression(self, tmp_path): - """propagate() completes without AttributeError after legacy cleanup.""" + """propagate() completes and stores the decision after the redesign.""" + import functools + fake_state = { "final_trade_decision": "Rating: Buy\nBuy NVDA.", "company_of_interest": "NVDA", @@ -660,6 +662,11 @@ class TestLegacyRemoval: mock_graph.propagator.create_initial_state.return_value = fake_state mock_graph.propagator.get_graph_args.return_value = {} mock_graph.signal_processor.process_signal.return_value = "Buy" + # Bind the real _run_graph so propagate's call to self._run_graph executes + # the actual write path instead of the auto-MagicMock. + mock_graph._run_graph = functools.partial( + TradingAgentsGraph._run_graph, mock_graph + ) TradingAgentsGraph.propagate(mock_graph, "NVDA", "2026-01-10") entries = mock_graph.memory_log.load_entries() assert len(entries) == 1 diff --git a/tradingagents/default_config.py b/tradingagents/default_config.py index 19dbe1c7a..89b517659 100644 --- a/tradingagents/default_config.py +++ b/tradingagents/default_config.py @@ -16,6 +16,9 @@ DEFAULT_CONFIG = { "google_thinking_level": None, # "high", "minimal", etc. "openai_reasoning_effort": None, # "medium", "high", "low" "anthropic_effort": None, # "high", "medium", "low" + # Checkpoint/resume: when True, LangGraph saves state after each node + # so a crashed run can resume from the last successful step. + "checkpoint_enabled": False, # Output language for analyst reports and final decision # Internal agent debate stays in English for reasoning quality "output_language": "English", diff --git a/tradingagents/graph/checkpointer.py b/tradingagents/graph/checkpointer.py new file mode 100644 index 000000000..7a73ee446 --- /dev/null +++ b/tradingagents/graph/checkpointer.py @@ -0,0 +1,86 @@ +"""LangGraph checkpoint support for resumable analysis runs. + +Per-ticker SQLite databases so concurrent tickers don't contend. +""" + +from __future__ import annotations + +import hashlib +import sqlite3 +from contextlib import contextmanager +from pathlib import Path +from typing import Generator + +from langgraph.checkpoint.sqlite import SqliteSaver + + +def _db_path(data_dir: str | Path, ticker: str) -> Path: + """Return the SQLite checkpoint DB path for a ticker.""" + p = Path(data_dir) / "checkpoints" + p.mkdir(parents=True, exist_ok=True) + return p / f"{ticker.upper()}.db" + + +def thread_id(ticker: str, date: str) -> str: + """Deterministic thread ID for a ticker+date pair.""" + return hashlib.sha256(f"{ticker.upper()}:{date}".encode()).hexdigest()[:16] + + +@contextmanager +def get_checkpointer(data_dir: str | Path, ticker: str) -> Generator[SqliteSaver, None, None]: + """Context manager yielding a SqliteSaver backed by a per-ticker DB.""" + db = _db_path(data_dir, ticker) + conn = sqlite3.connect(str(db), check_same_thread=False) + try: + saver = SqliteSaver(conn) + saver.setup() + yield saver + finally: + conn.close() + + +def has_checkpoint(data_dir: str | Path, ticker: str, date: str) -> bool: + """Check whether a resumable checkpoint exists for ticker+date.""" + return checkpoint_step(data_dir, ticker, date) is not None + + +def checkpoint_step(data_dir: str | Path, ticker: str, date: str) -> int | None: + """Return the step number of the latest checkpoint, or None if none exists.""" + db = _db_path(data_dir, ticker) + if not db.exists(): + return None + tid = thread_id(ticker, date) + with get_checkpointer(data_dir, ticker) as saver: + config = {"configurable": {"thread_id": tid}} + cp = saver.get_tuple(config) + if cp is None: + return None + return cp.metadata.get("step") + + +def clear_all_checkpoints(data_dir: str | Path) -> int: + """Remove all checkpoint DBs. Returns number of files deleted.""" + cp_dir = Path(data_dir) / "checkpoints" + if not cp_dir.exists(): + return 0 + dbs = list(cp_dir.glob("*.db")) + for db in dbs: + db.unlink() + return len(dbs) + + +def clear_checkpoint(data_dir: str | Path, ticker: str, date: str) -> None: + """Remove checkpoint for a specific ticker+date by deleting the thread's rows.""" + db = _db_path(data_dir, ticker) + if not db.exists(): + return + tid = thread_id(ticker, date) + conn = sqlite3.connect(str(db)) + try: + for table in ("writes", "checkpoints"): + conn.execute(f"DELETE FROM {table} WHERE thread_id = ?", (tid,)) + conn.commit() + except sqlite3.OperationalError: + pass + finally: + conn.close() diff --git a/tradingagents/graph/setup.py b/tradingagents/graph/setup.py index 1686fc5b0..45d6bfd38 100644 --- a/tradingagents/graph/setup.py +++ b/tradingagents/graph/setup.py @@ -179,5 +179,4 @@ class GraphSetup: workflow.add_edge("Portfolio Manager", END) - # Compile and return - return workflow.compile() + return workflow diff --git a/tradingagents/graph/trading_graph.py b/tradingagents/graph/trading_graph.py index 4f92b1889..bd6f1fc5c 100644 --- a/tradingagents/graph/trading_graph.py +++ b/tradingagents/graph/trading_graph.py @@ -38,6 +38,7 @@ from tradingagents.agents.utils.agent_utils import ( get_global_news ) +from .checkpointer import checkpoint_step, clear_checkpoint, get_checkpointer, thread_id from .conditional_logic import ConditionalLogic from .setup import GraphSetup from .propagation import Propagator @@ -123,8 +124,10 @@ class TradingAgentsGraph: self.ticker = None self.log_states_dict = {} # date to full state dict - # Set up the graph - self.graph = self.graph_setup.setup_graph(selected_analysts) + # Set up the graph: keep the workflow for recompilation with a checkpointer. + self.workflow = self.graph_setup.setup_graph(selected_analysts) + self.graph = self.workflow.compile() + self._checkpointer_ctx = None def _get_provider_kwargs(self) -> Dict[str, Any]: """Get provider-specific kwargs for LLM client creation.""" @@ -259,23 +262,58 @@ class TradingAgentsGraph: self.memory_log.batch_update_with_outcomes(updates) def propagate(self, company_name, trade_date): - """Run the trading agents graph for a company on a specific date.""" + """Run the trading agents graph for a company on a specific date. + When ``checkpoint_enabled`` is set in config, the graph is recompiled + with a per-ticker SqliteSaver so a crashed run can resume from the last + successful node on a subsequent invocation with the same ticker+date. + """ self.ticker = company_name - # Resolve any pending log entries for this ticker before the pipeline runs. - # This adds the outcome + reflection from the previous run at zero latency cost. + # Resolve any pending memory-log entries for this ticker before the pipeline runs. self._resolve_pending_entries(company_name) - # Initialize state — inject memory log context for PM + # Recompile with a checkpointer if the user opted in. + if self.config.get("checkpoint_enabled"): + self._checkpointer_ctx = get_checkpointer( + self.config["data_cache_dir"], company_name + ) + saver = self._checkpointer_ctx.__enter__() + self.graph = self.workflow.compile(checkpointer=saver) + + step = checkpoint_step( + self.config["data_cache_dir"], company_name, str(trade_date) + ) + if step is not None: + logger.info( + "Resuming from step %d for %s on %s", step, company_name, trade_date + ) + else: + logger.info("Starting fresh for %s on %s", company_name, trade_date) + + try: + return self._run_graph(company_name, trade_date) + finally: + if self._checkpointer_ctx is not None: + self._checkpointer_ctx.__exit__(None, None, None) + self._checkpointer_ctx = None + self.graph = self.workflow.compile() + + def _run_graph(self, company_name, trade_date): + """Execute the graph and write the resulting state to disk and memory log.""" + # Initialize state — inject memory log context for PM. past_context = self.memory_log.get_past_context(company_name) init_agent_state = self.propagator.create_initial_state( company_name, trade_date, past_context=past_context ) args = self.propagator.get_graph_args() + # Inject thread_id so same ticker+date resumes, different date starts fresh. + if self.config.get("checkpoint_enabled"): + tid = thread_id(company_name, str(trade_date)) + args.setdefault("config", {}).setdefault("configurable", {})["thread_id"] = tid + if self.debug: - # Debug mode with tracing trace = [] for chunk in self.graph.stream(init_agent_state, **args): if len(chunk["messages"]) == 0: @@ -283,26 +321,29 @@ class TradingAgentsGraph: else: chunk["messages"][-1].pretty_print() trace.append(chunk) - final_state = trace[-1] else: - # Standard mode without tracing final_state = self.graph.invoke(init_agent_state, **args) - # Store current state for reflection + # Store current state for reflection. self.curr_state = final_state - # Log state + # Log state to disk. self._log_state(trade_date, final_state) - # Store decision for deferred reflection. + # Store decision for deferred reflection on the next same-ticker run. self.memory_log.store_decision( ticker=company_name, trade_date=trade_date, final_trade_decision=final_state["final_trade_decision"], ) - # Return decision and processed signal + # Clear checkpoint on successful completion to avoid stale state. + if self.config.get("checkpoint_enabled"): + clear_checkpoint( + self.config["data_cache_dir"], company_name, str(trade_date) + ) + return final_state, self.process_signal(final_state["final_trade_decision"]) def _log_state(self, trade_date, final_state): From 0fda24515f56be44bdf568a19d85b20f69ee60b6 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 19:57:26 +0000 Subject: [PATCH 07/12] feat: structured-output Portfolio Manager + 5-tier rating consistency (#434) Three related changes that take the rating pipeline from heuristic-only to type-safe at the source. 1) Research Manager prompt now uses the same 5-tier scale (Buy / Overweight / Hold / Underweight / Sell) as the Portfolio Manager, signal_processing, and the memory log. The prior 3-tier wording (Buy / Sell / Hold) was the only remaining inconsistency in the pipeline. 2) Centralise the 5-tier vocabulary and the heuristic prose-rating parser into tradingagents/agents/utils/rating.py. Both the memory log and the signal processor now share the same parser instead of duplicating regex and word-walker logic. 3) Make structured output a first-class part of the Portfolio Manager's primary call. The PM uses llm.with_structured_output(PortfolioDecision) so each provider's native structured-output mode (json_schema for OpenAI/xAI, response_schema for Gemini, tool-use for Anthropic, function_calling for OpenAI-compatible providers) yields a typed Pydantic instance directly. A render helper turns that instance back into the same markdown shape downstream consumers (memory log, CLI display, saved reports) already expect, so no other code has to know the PM now produces structured output. Providers without structured support fall back gracefully to free-text + the deterministic heuristic. The previous SignalProcessor had been making a second LLM call to re-extract the rating from the PM's prose; that round-trip is now eliminated. SignalProcessor is a thin adapter over parse_rating(), makes zero LLM calls, and stays for backwards compatibility with process_signal() callers. Schema (PortfolioDecision) captures rating + executive_summary + investment_thesis + optional price_target + time_horizon, with field descriptions doubling as output instructions. Agent prose remains the primary artifact; structured output is layered onto the PM only because it is the one agent whose output has machine-readable downstream consumers. 15 new tests cover the heuristic parser (markdown-bold edge cases that had no coverage before), the structured PM happy path, the free-text fallback path, and that SignalProcessor never invokes the LLM. Full suite: 77 tests pass in ~2s without API keys. --- tests/test_memory_log.py | 70 +++++++++++--- tests/test_signal_processing.py | 90 ++++++++++++++++++ .../agents/managers/portfolio_manager.py | 83 ++++++++++++----- .../agents/managers/research_manager.py | 34 ++++--- tradingagents/agents/schemas.py | 93 +++++++++++++++++++ tradingagents/agents/utils/memory.py | 20 +--- tradingagents/agents/utils/rating.py | 50 ++++++++++ tradingagents/graph/signal_processing.py | 46 +++++---- 8 files changed, 399 insertions(+), 87 deletions(-) create mode 100644 tests/test_signal_processing.py create mode 100644 tradingagents/agents/schemas.py create mode 100644 tradingagents/agents/utils/rating.py diff --git a/tests/test_memory_log.py b/tests/test_memory_log.py index ccd1ca7e7..e0da15efc 100644 --- a/tests/test_memory_log.py +++ b/tests/test_memory_log.py @@ -5,6 +5,7 @@ import pandas as pd from unittest.mock import MagicMock, patch from tradingagents.agents.utils.memory import TradingMemoryLog +from tradingagents.agents.schemas import PortfolioDecision, PortfolioRating from tradingagents.graph.reflection import Reflector from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.graph.propagation import Propagator @@ -82,6 +83,25 @@ def _make_pm_state(past_context=""): } +def _structured_pm_llm(captured: dict, decision: PortfolioDecision | None = None): + """Build a MagicMock LLM whose with_structured_output binding captures the + prompt and returns a real PortfolioDecision (so render_pm_decision works). + """ + if decision is None: + decision = PortfolioDecision( + rating=PortfolioRating.HOLD, + executive_summary="Hold the position; await catalyst.", + investment_thesis="Balanced view; neither side carried the debate.", + ) + structured = MagicMock() + structured.invoke.side_effect = lambda prompt: ( + captured.__setitem__("prompt", prompt) or decision + ) + llm = MagicMock() + llm.with_structured_output.return_value = structured + return llm + + # --------------------------------------------------------------------------- # Core: storage and read path # --------------------------------------------------------------------------- @@ -518,29 +538,55 @@ class TestPortfolioManagerInjection: def test_pm_prompt_includes_past_context(self): captured = {} - mock_llm = MagicMock() - mock_llm.invoke.side_effect = lambda prompt: ( - captured.__setitem__("prompt", prompt) or MagicMock(content="Rating: Hold\nHold.") - ) - pm_node = create_portfolio_manager(mock_llm) + llm = _structured_pm_llm(captured) + pm_node = create_portfolio_manager(llm) state = _make_pm_state(past_context="[2026-01-05 | NVDA | Buy | +5.0% | +2.0% | 5d]\nGreat call.") pm_node(state) assert "Lessons from prior decisions and outcomes" in captured["prompt"] assert "Great call." in captured["prompt"] - assert "and the lessons from prior decisions" in captured["prompt"] def test_pm_no_past_context_no_section(self): """PM prompt omits the lessons section entirely when past_context is empty.""" captured = {} - mock_llm = MagicMock() - mock_llm.invoke.side_effect = lambda prompt: ( - captured.__setitem__("prompt", prompt) or MagicMock(content="Rating: Hold\nHold.") - ) - pm_node = create_portfolio_manager(mock_llm) + llm = _structured_pm_llm(captured) + pm_node = create_portfolio_manager(llm) state = _make_pm_state(past_context="") pm_node(state) assert "Lessons from prior decisions" not in captured["prompt"] - assert "and the lessons from prior decisions" not in captured["prompt"] + + def test_pm_returns_rendered_markdown_with_rating(self): + """The structured PortfolioDecision is rendered to markdown that + downstream consumers (memory log, signal processor, CLI display) + can parse without any extra LLM call.""" + captured = {} + decision = PortfolioDecision( + rating=PortfolioRating.OVERWEIGHT, + executive_summary="Build position gradually over the next two weeks.", + investment_thesis="AI capex cycle remains intact; institutional flows constructive.", + price_target=215.0, + time_horizon="3-6 months", + ) + llm = _structured_pm_llm(captured, decision) + pm_node = create_portfolio_manager(llm) + result = pm_node(_make_pm_state()) + md = result["final_trade_decision"] + assert "**Rating**: Overweight" in md + assert "**Executive Summary**: Build position gradually" in md + assert "**Investment Thesis**: AI capex cycle" in md + assert "**Price Target**: 215.0" in md + assert "**Time Horizon**: 3-6 months" in md + + def test_pm_falls_back_to_freetext_when_structured_unavailable(self): + """If a provider does not support with_structured_output, the agent + falls back to a plain invoke and returns whatever prose the model + produced, so the pipeline never blocks.""" + plain_response = "**Rating**: Sell\n\nExit ahead of guidance." + llm = MagicMock() + llm.with_structured_output.side_effect = NotImplementedError("provider unsupported") + llm.invoke.return_value = MagicMock(content=plain_response) + pm_node = create_portfolio_manager(llm) + result = pm_node(_make_pm_state()) + assert result["final_trade_decision"] == plain_response # get_past_context ordering and limits diff --git a/tests/test_signal_processing.py b/tests/test_signal_processing.py new file mode 100644 index 000000000..4bbfb7475 --- /dev/null +++ b/tests/test_signal_processing.py @@ -0,0 +1,90 @@ +"""Tests for the shared rating heuristic and the SignalProcessor adapter. + +The Portfolio Manager produces a typed PortfolioDecision via structured +output and renders it to markdown that always contains a ``**Rating**: X`` +header. The deterministic heuristic in ``tradingagents.agents.utils.rating`` +is therefore sufficient to extract the rating downstream — no second LLM +call is needed — and SignalProcessor is now a thin adapter that delegates +to it. +""" + +import pytest + +from tradingagents.agents.utils.rating import RATINGS_5_TIER, parse_rating +from tradingagents.graph.signal_processing import SignalProcessor + + +# --------------------------------------------------------------------------- +# Heuristic parser +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestParseRating: + def test_explicit_label_buy(self): + assert parse_rating("Rating: Buy\nReasoning here.") == "Buy" + + def test_explicit_label_overweight(self): + assert parse_rating("Rating: Overweight\nDetails.") == "Overweight" + + def test_explicit_label_with_markdown_bold_value(self): + # Regression: Rating: **Sell** — markdown around the value. + assert parse_rating("Rating: **Sell**\nExit immediately.") == "Sell" + + def test_explicit_label_with_markdown_bold_label(self): + assert parse_rating("**Rating**: Underweight\nTrim exposure.") == "Underweight" + + def test_rendered_pm_markdown_shape(self): + # The exact shape produced by render_pm_decision must always parse. + text = ( + "**Rating**: Buy\n\n" + "**Executive Summary**: Enter at $189-192, 6% portfolio cap.\n\n" + "**Investment Thesis**: AI capex cycle intact; institutional flows constructive." + ) + assert parse_rating(text) == "Buy" + + def test_explicit_label_wins_over_prose_with_markdown(self): + text = ( + "The buy thesis is weakened by guidance.\n" + "Rating: **Sell**\n" + "Exit before earnings." + ) + assert parse_rating(text) == "Sell" + + def test_no_rating_returns_default(self): + assert parse_rating("No clear directional signal at this time.") == "Hold" + + def test_no_rating_custom_default(self): + assert parse_rating("Plain prose.", default="Underweight") == "Underweight" + + def test_all_five_tiers_recognised(self): + for r in RATINGS_5_TIER: + assert parse_rating(f"Rating: {r}") == r + + +# --------------------------------------------------------------------------- +# SignalProcessor: thin adapter over the heuristic +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestSignalProcessor: + def test_returns_rating_from_pm_markdown(self): + sp = SignalProcessor() + md = "**Rating**: Overweight\n\n**Executive Summary**: Build gradually." + assert sp.process_signal(md) == "Overweight" + + def test_makes_no_llm_calls(self): + """SignalProcessor must not invoke the LLM it was constructed with — + the rating is parseable from the rendered PM markdown directly.""" + from unittest.mock import MagicMock + + llm = MagicMock() + sp = SignalProcessor(llm) + sp.process_signal("Rating: Buy\nDetails.") + llm.invoke.assert_not_called() + llm.with_structured_output.assert_not_called() + + def test_default_when_no_rating_present(self): + sp = SignalProcessor() + assert sp.process_signal("Plain prose without a recommendation.") == "Hold" diff --git a/tradingagents/agents/managers/portfolio_manager.py b/tradingagents/agents/managers/portfolio_manager.py index 6780c7dc1..38ab840af 100644 --- a/tradingagents/agents/managers/portfolio_manager.py +++ b/tradingagents/agents/managers/portfolio_manager.py @@ -1,29 +1,53 @@ -from tradingagents.agents.utils.agent_utils import build_instrument_context, get_language_instruction +"""Portfolio Manager: synthesises the risk-analyst debate into the final decision. + +Uses LangChain's ``with_structured_output`` so the LLM produces a typed +``PortfolioDecision`` directly, in a single call. The result is rendered +back to markdown for storage in ``final_trade_decision`` so memory log, +CLI display, and saved reports continue to consume the same shape they do +today. When a provider does not expose structured output, the agent falls +back to a free-text invocation and the existing heuristic rating parser. +""" + +from __future__ import annotations + +import logging + +from tradingagents.agents.schemas import PortfolioDecision, render_pm_decision +from tradingagents.agents.utils.agent_utils import ( + build_instrument_context, + get_language_instruction, +) + +logger = logging.getLogger(__name__) def create_portfolio_manager(llm): - def portfolio_manager_node(state) -> dict: + # Wrap once at agent construction; if the provider does not support + # structured output we keep ``structured_llm`` as None and use the + # free-text fallback for every call. + try: + structured_llm = llm.with_structured_output(PortfolioDecision) + except (NotImplementedError, AttributeError) as exc: + logger.warning( + "Portfolio Manager: provider does not support with_structured_output (%s); " + "falling back to free-text generation", + exc, + ) + structured_llm = None + def portfolio_manager_node(state) -> dict: instrument_context = build_instrument_context(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["fundamentals_report"] - sentiment_report = state["sentiment_report"] research_plan = state["investment_plan"] trader_plan = state["trader_investment_plan"] past_context = state.get("past_context", "") lessons_line = ( f"- Lessons from prior decisions and outcomes:\n{past_context}\n" - if past_context else "" - ) - thesis_instruction = ( - "3. **Investment Thesis**: Detailed reasoning anchored in the analysts' debate and the lessons from prior decisions." if past_context - else "3. **Investment Thesis**: Detailed reasoning anchored in the analysts' debate." + else "" ) prompt = f"""As the Portfolio Manager, synthesize the risk analysts' debate and deliver the final trading decision. @@ -43,14 +67,6 @@ def create_portfolio_manager(llm): - Research Manager's investment plan: **{research_plan}** - Trader's transaction proposal: **{trader_plan}** {lessons_line} - -**Required Output Structure:** -1. **Rating**: State one of Buy / Overweight / Hold / Underweight / Sell. -2. **Executive Summary**: A concise action plan covering entry strategy, position sizing, key risk levels, and time horizon. -{thesis_instruction} - ---- - **Risk Analysts Debate History:** {history} @@ -58,10 +74,10 @@ def create_portfolio_manager(llm): Be decisive and ground every conclusion in specific evidence from the analysts.{get_language_instruction()}""" - response = llm.invoke(prompt) + final_trade_decision = _invoke_pm(structured_llm, llm, prompt) new_risk_debate_state = { - "judge_decision": response.content, + "judge_decision": final_trade_decision, "history": risk_debate_state["history"], "aggressive_history": risk_debate_state["aggressive_history"], "conservative_history": risk_debate_state["conservative_history"], @@ -75,7 +91,30 @@ Be decisive and ground every conclusion in specific evidence from the analysts.{ return { "risk_debate_state": new_risk_debate_state, - "final_trade_decision": response.content, + "final_trade_decision": final_trade_decision, } return portfolio_manager_node + + +def _invoke_pm(structured_llm, plain_llm, prompt: str) -> str: + """Run the PM call and return the markdown-rendered decision. + + Tries the structured-output path first; if it fails for any reason + (provider does not support it, model returns malformed JSON, network + glitch on the structured endpoint), falls back to the plain free-text + invocation so the pipeline still produces a result. + """ + if structured_llm is not None: + try: + decision = structured_llm.invoke(prompt) + return render_pm_decision(decision) + except Exception as exc: + logger.warning( + "Portfolio Manager: structured-output invocation failed (%s); " + "retrying once as free text", + exc, + ) + + response = plain_llm.invoke(prompt) + return response.content diff --git a/tradingagents/agents/managers/research_manager.py b/tradingagents/agents/managers/research_manager.py index 3902a60c4..020b719e4 100644 --- a/tradingagents/agents/managers/research_manager.py +++ b/tradingagents/agents/managers/research_manager.py @@ -9,21 +9,31 @@ def create_research_manager(llm): investment_debate_state = state["investment_debate_state"] - 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 bear analyst, the bull analyst, or choose 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—Buy, Sell, or Hold—must be clear and actionable. Avoid defaulting to 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. -Present your analysis conversationally, as if speaking naturally, without special formatting. + prompt = f"""As the Research Manager and debate facilitator, your role is to critically evaluate this round of debate and deliver a clear, actionable investment plan for the trader. {instrument_context} -Here is the debate: -Debate History: +--- + +**Rating Scale** (use exactly one): +- **Buy**: Strong conviction in the bull thesis; recommend taking or growing the position +- **Overweight**: Constructive view; recommend gradually increasing exposure +- **Hold**: Balanced view; recommend maintaining the current position +- **Underweight**: Cautious view; recommend trimming exposure +- **Sell**: Strong conviction in the bear thesis; recommend exiting or avoiding the position + +Commit to a clear stance whenever the debate's strongest arguments warrant one; reserve Hold for situations where the evidence on both sides is genuinely balanced. + +**Required Output Structure:** +1. **Recommendation**: State one of Buy / Overweight / Hold / Underweight / Sell. +2. **Rationale**: Summarise the key points from both sides and explain which arguments led to this recommendation. +3. **Strategic Actions**: Concrete steps for the trader to implement the recommendation, including position sizing guidance consistent with the rating. + +Present your analysis conversationally, as if speaking naturally to a teammate. + +--- + +**Debate History:** {history}""" response = llm.invoke(prompt) diff --git a/tradingagents/agents/schemas.py b/tradingagents/agents/schemas.py new file mode 100644 index 000000000..aadcaf0b6 --- /dev/null +++ b/tradingagents/agents/schemas.py @@ -0,0 +1,93 @@ +"""Pydantic schemas used by agents that produce structured output. + +The framework's primary artifact is still prose: each agent's natural-language +reasoning is what users read, what gets stored in the memory log, and what +gets saved as markdown reports. Structured output is layered onto agents +whose results have downstream machine-readable consumers (currently only +the Portfolio Manager) so that: + +- The rating is type-safe and never has to be regex-extracted +- Schema field descriptions become the model's output instructions +- Each provider's native structured-output mode is used (json_schema for + OpenAI/xAI, response_schema for Gemini, tool-use for Anthropic) +- A render helper turns the parsed Pydantic instance back into the same + markdown shape the rest of the system already consumes, so display, + memory log, and saved reports keep working unchanged +""" + +from __future__ import annotations + +from enum import Enum +from typing import Optional + +from pydantic import BaseModel, Field + + +class PortfolioRating(str, Enum): + """5-tier portfolio rating used by the Research Manager and Portfolio Manager.""" + + BUY = "Buy" + OVERWEIGHT = "Overweight" + HOLD = "Hold" + UNDERWEIGHT = "Underweight" + SELL = "Sell" + + +class PortfolioDecision(BaseModel): + """Structured output produced by the Portfolio Manager. + + The model fills every field as part of its primary LLM call; no separate + extraction pass is required. Field descriptions double as the model's + output instructions, so the prompt body only needs to convey context and + the rating-scale guidance. + """ + + rating: PortfolioRating = Field( + description=( + "The final position rating. Exactly one of Buy / Overweight / Hold / " + "Underweight / Sell, picked based on the analysts' debate." + ), + ) + executive_summary: str = Field( + description=( + "A concise action plan covering entry strategy, position sizing, " + "key risk levels, and time horizon. Two to four sentences." + ), + ) + investment_thesis: str = Field( + description=( + "Detailed reasoning anchored in specific evidence from the analysts' " + "debate. If prior lessons are referenced in the prompt context, " + "incorporate them; otherwise rely solely on the current analysis." + ), + ) + price_target: Optional[float] = Field( + default=None, + description="Optional target price in the instrument's quote currency.", + ) + time_horizon: Optional[str] = Field( + default=None, + description="Optional recommended holding period, e.g. '3-6 months'.", + ) + + +def render_pm_decision(decision: PortfolioDecision) -> str: + """Render a PortfolioDecision back to the markdown shape the rest of the system expects. + + Memory log, CLI display, and saved report files all read this markdown, + so the rendered output preserves the exact section headers (``**Rating**``, + ``**Executive Summary**``, ``**Investment Thesis**``) that downstream + parsers and the report writers already handle. + """ + parts = [ + f"**Rating**: {decision.rating.value}", + "", + f"**Executive Summary**: {decision.executive_summary}", + "", + f"**Investment Thesis**: {decision.investment_thesis}", + ] + if decision.price_target is not None: + parts.extend(["", f"**Price Target**: {decision.price_target}"]) + if decision.time_horizon: + parts.extend(["", f"**Time Horizon**: {decision.time_horizon}"]) + return "\n".join(parts) diff --git a/tradingagents/agents/utils/memory.py b/tradingagents/agents/utils/memory.py index 199cb8946..fee5ac4a2 100644 --- a/tradingagents/agents/utils/memory.py +++ b/tradingagents/agents/utils/memory.py @@ -4,17 +4,17 @@ from typing import List, Optional from pathlib import Path import re +from tradingagents.agents.utils.rating import parse_rating + class TradingMemoryLog: """Append-only markdown log of trading decisions and reflections.""" - RATINGS = {"buy", "overweight", "hold", "underweight", "sell"} # HTML comment: cannot appear in LLM prose output, safe as a hard delimiter _SEPARATOR = "\n\n\n\n" # Precompiled patterns — avoids re-compilation on every load_entries() call _DECISION_RE = re.compile(r"DECISION:\n(.*?)(?=\nREFLECTION:|\Z)", re.DOTALL) _REFLECTION_RE = re.compile(r"REFLECTION:\n(.*?)$", re.DOTALL) - _RATING_LABEL_RE = re.compile(r"rating.*?[:\-][\s*]*(\w+)", re.IGNORECASE) def __init__(self, config: dict = None): self._log_path = None @@ -40,7 +40,7 @@ class TradingMemoryLog: for line in raw.splitlines(): if line.startswith(f"[{trade_date} | {ticker} |") and line.endswith("| pending]"): return - rating = self._parse_rating(final_trade_decision) + rating = parse_rating(final_trade_decision) tag = f"[{trade_date} | {ticker} | {rating} | pending]" entry = f"{tag}\n\nDECISION:\n{final_trade_decision}{self._SEPARATOR}" with open(self._log_path, "a", encoding="utf-8") as f: @@ -213,20 +213,6 @@ class TradingMemoryLog: # --- Helpers --- - def _parse_rating(self, text: str) -> str: - # First pass: explicit "Rating: X" label — search handles markdown bold/numbered lists - for line in text.splitlines(): - m = self._RATING_LABEL_RE.search(line) - if m and m.group(1).lower() in self.RATINGS: - return m.group(1).capitalize() - # Fallback: first rating word found anywhere in the text - for line in text.splitlines(): - for word in line.lower().split(): - clean = word.strip("*:.,") - if clean in self.RATINGS: - return clean.capitalize() - return "Hold" - def _parse_entry(self, raw: str) -> Optional[dict]: lines = raw.strip().splitlines() if not lines: diff --git a/tradingagents/agents/utils/rating.py b/tradingagents/agents/utils/rating.py new file mode 100644 index 000000000..d5032346a --- /dev/null +++ b/tradingagents/agents/utils/rating.py @@ -0,0 +1,50 @@ +"""Shared 5-tier rating vocabulary and a deterministic heuristic parser. + +The same five-tier scale (Buy, Overweight, Hold, Underweight, Sell) is used by: +- The Research Manager (investment plan recommendation) +- The Portfolio Manager (final position decision) +- The signal processor (rating extracted for downstream consumers) +- The memory log (rating tag stored alongside each decision entry) + +Centralising it here avoids drift between those call sites. +""" + +from __future__ import annotations + +import re +from typing import Tuple + + +# Canonical, ordered 5-tier scale (most bullish to most bearish). +RATINGS_5_TIER: Tuple[str, ...] = ( + "Buy", "Overweight", "Hold", "Underweight", "Sell", +) + +_RATING_SET = {r.lower() for r in RATINGS_5_TIER} + +# Matches "Rating: X" / "rating - X" / "Rating: **X**" — tolerates markdown +# bold wrappers and either a colon or hyphen separator. +_RATING_LABEL_RE = re.compile(r"rating.*?[:\-][\s*]*(\w+)", re.IGNORECASE) + + +def parse_rating(text: str, default: str = "Hold") -> str: + """Heuristically extract a 5-tier rating from prose text. + + Two-pass strategy: + 1. Look for an explicit "Rating: X" label (tolerant of markdown bold). + 2. Fall back to the first 5-tier rating word found anywhere in the text. + + Returns a Title-cased rating string, or ``default`` if no rating word appears. + """ + for line in text.splitlines(): + m = _RATING_LABEL_RE.search(line) + if m and m.group(1).lower() in _RATING_SET: + return m.group(1).capitalize() + + for line in text.splitlines(): + for word in line.lower().split(): + clean = word.strip("*:.,") + if clean in _RATING_SET: + return clean.capitalize() + + return default diff --git a/tradingagents/graph/signal_processing.py b/tradingagents/graph/signal_processing.py index 5ac66c1dd..90fafd04b 100644 --- a/tradingagents/graph/signal_processing.py +++ b/tradingagents/graph/signal_processing.py @@ -1,33 +1,31 @@ -# TradingAgents/graph/signal_processing.py +"""Extract the 5-tier portfolio rating from the Portfolio Manager's decision. + +The Portfolio Manager produces a typed ``PortfolioDecision`` via structured +output and renders it to markdown that always carries a ``**Rating**: X`` +header (see :func:`tradingagents.agents.schemas.render_pm_decision`). The +deterministic heuristic in :mod:`tradingagents.agents.utils.rating` is more +than sufficient to extract that rating; no extra LLM call is needed. + +This module exists for backwards compatibility with callers that expect a +``SignalProcessor.process_signal(text)`` interface. +""" + +from __future__ import annotations from typing import Any +from tradingagents.agents.utils.rating import parse_rating + class SignalProcessor: - """Processes trading signals to extract actionable decisions.""" + """Read the 5-tier rating out of a Portfolio Manager decision.""" - def __init__(self, quick_thinking_llm: Any): - """Initialize with an LLM for processing.""" + def __init__(self, quick_thinking_llm: Any = None): + # The LLM argument is accepted for backwards compatibility but no + # longer used: the PM's structured output guarantees the rating is + # parseable from the rendered markdown without a second LLM call. self.quick_thinking_llm = quick_thinking_llm def process_signal(self, full_signal: str) -> str: - """ - Process a full trading signal to extract the core decision. - - Args: - full_signal: Complete trading signal text - - Returns: - Extracted rating (BUY, OVERWEIGHT, HOLD, UNDERWEIGHT, or SELL) - """ - messages = [ - ( - "system", - "You are an efficient assistant that extracts the trading decision from analyst reports. " - "Extract the rating as exactly one of: BUY, OVERWEIGHT, HOLD, UNDERWEIGHT, SELL. " - "Output only the single rating word, nothing else.", - ), - ("human", full_signal), - ] - - return self.quick_thinking_llm.invoke(messages).content + """Return one of Buy / Overweight / Hold / Underweight / Sell.""" + return parse_rating(full_signal) From bba147798f9b440bcf08fe3f3db3ca8464f9c73a Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 20:27:23 +0000 Subject: [PATCH 08/12] feat: structured-output Trader and Research Manager (#434, finishes the trio) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Extends the canonical structured-output pattern from the Portfolio Manager to the other two decision-making agents. Each of the three agents now returns a typed Pydantic instance via llm.with_structured_output() in a single primary call, and a render helper turns the result into the same markdown shape downstream agents and saved reports already consume. - ResearchPlan: 5-tier recommendation, conversational rationale, concrete strategic actions for the trader. - TraderProposal: 3-tier action (transaction direction is naturally Buy / Hold / Sell — position sizing happens later at the Portfolio Manager), reasoning, and optional entry_price / stop_loss / position_sizing. Rendered output preserves the trailing "FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**" line for backward compatibility with the analyst stop-signal text. - PortfolioDecision: 5-tier rating, executive summary, investment thesis, optional price_target / time_horizon (unchanged). The shared try-structured-then-fallback pattern is extracted into tradingagents/agents/utils/structured.py (bind_structured + invoke_structured_or_freetext) so all three agents go through the same code path and log the same warning when a provider lacks structured output and the agent falls back to free-text generation. Net effect for users: every saved markdown report (research/manager.md, trading/trader.md, portfolio/decision.md) now has consistent section headers across runs and providers, easier to scan. Net effect for the runtime: the rating extraction round-trip is gone — the rating comes from the structured response itself, not a second LLM call. SignalProcessor was already simplified to a heuristic adapter in the previous commit. 11 new tests in tests/test_structured_agents.py cover the Trader and Research Manager render functions, structured-output happy paths, and free-text fallback. Full suite: 88 tests pass in ~2s without API keys. --- tests/test_structured_agents.py | 232 ++++++++++++++++++ .../agents/managers/portfolio_manager.py | 54 +--- .../agents/managers/research_manager.py | 32 ++- tradingagents/agents/schemas.py | 149 ++++++++++- tradingagents/agents/trader/trader.py | 48 +++- tradingagents/agents/utils/structured.py | 73 ++++++ 6 files changed, 519 insertions(+), 69 deletions(-) create mode 100644 tests/test_structured_agents.py create mode 100644 tradingagents/agents/utils/structured.py diff --git a/tests/test_structured_agents.py b/tests/test_structured_agents.py new file mode 100644 index 000000000..ea771a4b0 --- /dev/null +++ b/tests/test_structured_agents.py @@ -0,0 +1,232 @@ +"""Tests for structured-output agents (Trader and Research Manager). + +The Portfolio Manager has its own coverage in tests/test_memory_log.py +(which exercises the full memory-log → PM injection cycle). This file +covers the parallel schemas, render functions, and graceful-fallback +behavior we added for the Trader and Research Manager so all three +decision-making agents share the same shape. +""" + +from unittest.mock import MagicMock + +import pytest + +from tradingagents.agents.managers.research_manager import create_research_manager +from tradingagents.agents.schemas import ( + PortfolioRating, + ResearchPlan, + TraderAction, + TraderProposal, + render_research_plan, + render_trader_proposal, +) +from tradingagents.agents.trader.trader import create_trader + + +# --------------------------------------------------------------------------- +# Render functions +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestRenderTraderProposal: + def test_minimal_required_fields(self): + p = TraderProposal(action=TraderAction.HOLD, reasoning="Balanced setup; no edge.") + md = render_trader_proposal(p) + assert "**Action**: Hold" in md + assert "**Reasoning**: Balanced setup; no edge." in md + # The trailing FINAL TRANSACTION PROPOSAL line is preserved for the + # analyst stop-signal text and any external code that greps for it. + assert "FINAL TRANSACTION PROPOSAL: **HOLD**" in md + + def test_optional_fields_included_when_present(self): + p = TraderProposal( + action=TraderAction.BUY, + reasoning="Strong technicals + fundamentals.", + entry_price=189.5, + stop_loss=178.0, + position_sizing="6% of portfolio", + ) + md = render_trader_proposal(p) + assert "**Action**: Buy" in md + assert "**Entry Price**: 189.5" in md + assert "**Stop Loss**: 178.0" in md + assert "**Position Sizing**: 6% of portfolio" in md + assert "FINAL TRANSACTION PROPOSAL: **BUY**" in md + + def test_optional_fields_omitted_when_absent(self): + p = TraderProposal(action=TraderAction.SELL, reasoning="Guidance cut.") + md = render_trader_proposal(p) + assert "Entry Price" not in md + assert "Stop Loss" not in md + assert "Position Sizing" not in md + assert "FINAL TRANSACTION PROPOSAL: **SELL**" in md + + +@pytest.mark.unit +class TestRenderResearchPlan: + def test_required_fields(self): + p = ResearchPlan( + recommendation=PortfolioRating.OVERWEIGHT, + rationale="Bull case carried; tailwinds intact.", + strategic_actions="Build position over two weeks; cap at 5%.", + ) + md = render_research_plan(p) + assert "**Recommendation**: Overweight" in md + assert "**Rationale**: Bull case carried" in md + assert "**Strategic Actions**: Build position" in md + + def test_all_5_tier_ratings_render(self): + for rating in PortfolioRating: + p = ResearchPlan( + recommendation=rating, + rationale="r", + strategic_actions="s", + ) + md = render_research_plan(p) + assert f"**Recommendation**: {rating.value}" in md + + +# --------------------------------------------------------------------------- +# Trader agent: structured happy path + fallback +# --------------------------------------------------------------------------- + + +def _make_trader_state(): + return { + "company_of_interest": "NVDA", + "investment_plan": "**Recommendation**: Buy\n**Rationale**: ...\n**Strategic Actions**: ...", + } + + +def _structured_trader_llm(captured: dict, proposal: TraderProposal | None = None): + """Build a MagicMock LLM whose with_structured_output binding captures the + prompt and returns a real TraderProposal so render_trader_proposal works. + """ + if proposal is None: + proposal = TraderProposal( + action=TraderAction.BUY, + reasoning="Strong setup.", + ) + structured = MagicMock() + structured.invoke.side_effect = lambda prompt: ( + captured.__setitem__("prompt", prompt) or proposal + ) + llm = MagicMock() + llm.with_structured_output.return_value = structured + return llm + + +@pytest.mark.unit +class TestTraderAgent: + def test_structured_path_produces_rendered_markdown(self): + captured = {} + proposal = TraderProposal( + action=TraderAction.BUY, + reasoning="AI capex cycle intact; institutional flows constructive.", + entry_price=189.5, + stop_loss=178.0, + position_sizing="6% of portfolio", + ) + llm = _structured_trader_llm(captured, proposal) + trader = create_trader(llm) + result = trader(_make_trader_state()) + plan = result["trader_investment_plan"] + assert "**Action**: Buy" in plan + assert "**Entry Price**: 189.5" in plan + assert "FINAL TRANSACTION PROPOSAL: **BUY**" in plan + # The same rendered markdown is also added to messages for downstream agents. + assert plan in result["messages"][0].content + + def test_prompt_includes_investment_plan(self): + captured = {} + llm = _structured_trader_llm(captured) + trader = create_trader(llm) + trader(_make_trader_state()) + # The investment plan is in the user message of the captured prompt. + prompt = captured["prompt"] + assert any("Proposed Investment Plan" in m["content"] for m in prompt) + + def test_falls_back_to_freetext_when_structured_unavailable(self): + plain_response = ( + "**Action**: Sell\n\nGuidance cut hits margins.\n\n" + "FINAL TRANSACTION PROPOSAL: **SELL**" + ) + llm = MagicMock() + llm.with_structured_output.side_effect = NotImplementedError("provider unsupported") + llm.invoke.return_value = MagicMock(content=plain_response) + trader = create_trader(llm) + result = trader(_make_trader_state()) + assert result["trader_investment_plan"] == plain_response + + +# --------------------------------------------------------------------------- +# Research Manager agent: structured happy path + fallback +# --------------------------------------------------------------------------- + + +def _make_rm_state(): + return { + "company_of_interest": "NVDA", + "investment_debate_state": { + "history": "Bull and bear arguments here.", + "bull_history": "Bull says...", + "bear_history": "Bear says...", + "current_response": "", + "judge_decision": "", + "count": 1, + }, + } + + +def _structured_rm_llm(captured: dict, plan: ResearchPlan | None = None): + if plan is None: + plan = ResearchPlan( + recommendation=PortfolioRating.HOLD, + rationale="Balanced view across both sides.", + strategic_actions="Hold current position; reassess after earnings.", + ) + structured = MagicMock() + structured.invoke.side_effect = lambda prompt: ( + captured.__setitem__("prompt", prompt) or plan + ) + llm = MagicMock() + llm.with_structured_output.return_value = structured + return llm + + +@pytest.mark.unit +class TestResearchManagerAgent: + def test_structured_path_produces_rendered_markdown(self): + captured = {} + plan = ResearchPlan( + recommendation=PortfolioRating.OVERWEIGHT, + rationale="Bull case is stronger; AI tailwind intact.", + strategic_actions="Build position gradually over two weeks.", + ) + llm = _structured_rm_llm(captured, plan) + rm = create_research_manager(llm) + result = rm(_make_rm_state()) + ip = result["investment_plan"] + assert "**Recommendation**: Overweight" in ip + assert "**Rationale**: Bull case" in ip + assert "**Strategic Actions**: Build position" in ip + + def test_prompt_uses_5_tier_rating_scale(self): + """The RM prompt must list all five tiers so the schema enum matches user expectations.""" + captured = {} + llm = _structured_rm_llm(captured) + rm = create_research_manager(llm) + rm(_make_rm_state()) + prompt = captured["prompt"] + for tier in ("Buy", "Overweight", "Hold", "Underweight", "Sell"): + assert f"**{tier}**" in prompt, f"missing {tier} in prompt" + + def test_falls_back_to_freetext_when_structured_unavailable(self): + plain_response = "**Recommendation**: Sell\n\n**Rationale**: ...\n\n**Strategic Actions**: ..." + llm = MagicMock() + llm.with_structured_output.side_effect = NotImplementedError("provider unsupported") + llm.invoke.return_value = MagicMock(content=plain_response) + rm = create_research_manager(llm) + result = rm(_make_rm_state()) + assert result["investment_plan"] == plain_response diff --git a/tradingagents/agents/managers/portfolio_manager.py b/tradingagents/agents/managers/portfolio_manager.py index 38ab840af..0e7c18234 100644 --- a/tradingagents/agents/managers/portfolio_manager.py +++ b/tradingagents/agents/managers/portfolio_manager.py @@ -5,35 +5,24 @@ Uses LangChain's ``with_structured_output`` so the LLM produces a typed back to markdown for storage in ``final_trade_decision`` so memory log, CLI display, and saved reports continue to consume the same shape they do today. When a provider does not expose structured output, the agent falls -back to a free-text invocation and the existing heuristic rating parser. +back gracefully to free-text generation. """ from __future__ import annotations -import logging - from tradingagents.agents.schemas import PortfolioDecision, render_pm_decision from tradingagents.agents.utils.agent_utils import ( build_instrument_context, get_language_instruction, ) - -logger = logging.getLogger(__name__) +from tradingagents.agents.utils.structured import ( + bind_structured, + invoke_structured_or_freetext, +) def create_portfolio_manager(llm): - # Wrap once at agent construction; if the provider does not support - # structured output we keep ``structured_llm`` as None and use the - # free-text fallback for every call. - try: - structured_llm = llm.with_structured_output(PortfolioDecision) - except (NotImplementedError, AttributeError) as exc: - logger.warning( - "Portfolio Manager: provider does not support with_structured_output (%s); " - "falling back to free-text generation", - exc, - ) - structured_llm = None + structured_llm = bind_structured(llm, PortfolioDecision, "Portfolio Manager") def portfolio_manager_node(state) -> dict: instrument_context = build_instrument_context(state["company_of_interest"]) @@ -74,7 +63,13 @@ def create_portfolio_manager(llm): Be decisive and ground every conclusion in specific evidence from the analysts.{get_language_instruction()}""" - final_trade_decision = _invoke_pm(structured_llm, llm, prompt) + final_trade_decision = invoke_structured_or_freetext( + structured_llm, + llm, + prompt, + render_pm_decision, + "Portfolio Manager", + ) new_risk_debate_state = { "judge_decision": final_trade_decision, @@ -95,26 +90,3 @@ Be decisive and ground every conclusion in specific evidence from the analysts.{ } return portfolio_manager_node - - -def _invoke_pm(structured_llm, plain_llm, prompt: str) -> str: - """Run the PM call and return the markdown-rendered decision. - - Tries the structured-output path first; if it fails for any reason - (provider does not support it, model returns malformed JSON, network - glitch on the structured endpoint), falls back to the plain free-text - invocation so the pipeline still produces a result. - """ - if structured_llm is not None: - try: - decision = structured_llm.invoke(prompt) - return render_pm_decision(decision) - except Exception as exc: - logger.warning( - "Portfolio Manager: structured-output invocation failed (%s); " - "retrying once as free text", - exc, - ) - - response = plain_llm.invoke(prompt) - return response.content diff --git a/tradingagents/agents/managers/research_manager.py b/tradingagents/agents/managers/research_manager.py index 020b719e4..0e2206b2e 100644 --- a/tradingagents/agents/managers/research_manager.py +++ b/tradingagents/agents/managers/research_manager.py @@ -1,8 +1,18 @@ +"""Research Manager: turns the bull/bear debate into a structured investment plan for the trader.""" +from __future__ import annotations + +from tradingagents.agents.schemas import ResearchPlan, render_research_plan from tradingagents.agents.utils.agent_utils import build_instrument_context +from tradingagents.agents.utils.structured import ( + bind_structured, + invoke_structured_or_freetext, +) def create_research_manager(llm): + structured_llm = bind_structured(llm, ResearchPlan, "Research Manager") + def research_manager_node(state) -> dict: instrument_context = build_instrument_context(state["company_of_interest"]) history = state["investment_debate_state"].get("history", "") @@ -24,31 +34,31 @@ def create_research_manager(llm): Commit to a clear stance whenever the debate's strongest arguments warrant one; reserve Hold for situations where the evidence on both sides is genuinely balanced. -**Required Output Structure:** -1. **Recommendation**: State one of Buy / Overweight / Hold / Underweight / Sell. -2. **Rationale**: Summarise the key points from both sides and explain which arguments led to this recommendation. -3. **Strategic Actions**: Concrete steps for the trader to implement the recommendation, including position sizing guidance consistent with the rating. - -Present your analysis conversationally, as if speaking naturally to a teammate. - --- **Debate History:** {history}""" - response = llm.invoke(prompt) + + investment_plan = invoke_structured_or_freetext( + structured_llm, + llm, + prompt, + render_research_plan, + "Research Manager", + ) new_investment_debate_state = { - "judge_decision": response.content, + "judge_decision": investment_plan, "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, + "current_response": investment_plan, "count": investment_debate_state["count"], } return { "investment_debate_state": new_investment_debate_state, - "investment_plan": response.content, + "investment_plan": investment_plan, } return research_manager_node diff --git a/tradingagents/agents/schemas.py b/tradingagents/agents/schemas.py index aadcaf0b6..55f0e3cfb 100644 --- a/tradingagents/agents/schemas.py +++ b/tradingagents/agents/schemas.py @@ -1,15 +1,16 @@ """Pydantic schemas used by agents that produce structured output. The framework's primary artifact is still prose: each agent's natural-language -reasoning is what users read, what gets stored in the memory log, and what -gets saved as markdown reports. Structured output is layered onto agents -whose results have downstream machine-readable consumers (currently only -the Portfolio Manager) so that: +reasoning is what users read in the saved markdown reports and what the +downstream agents read as context. Structured output is layered onto the +three decision-making agents (Research Manager, Trader, Portfolio Manager) +so that: -- The rating is type-safe and never has to be regex-extracted -- Schema field descriptions become the model's output instructions +- Their outputs follow consistent section headers across runs and providers - Each provider's native structured-output mode is used (json_schema for OpenAI/xAI, response_schema for Gemini, tool-use for Anthropic) +- Schema field descriptions become the model's output instructions, freeing + the prompt body to focus on context and the rating-scale guidance - A render helper turns the parsed Pydantic instance back into the same markdown shape the rest of the system already consumes, so display, memory log, and saved reports keep working unchanged @@ -23,8 +24,13 @@ from typing import Optional from pydantic import BaseModel, Field +# --------------------------------------------------------------------------- +# Shared rating types +# --------------------------------------------------------------------------- + + class PortfolioRating(str, Enum): - """5-tier portfolio rating used by the Research Manager and Portfolio Manager.""" + """5-tier rating used by the Research Manager and Portfolio Manager.""" BUY = "Buy" OVERWEIGHT = "Overweight" @@ -33,6 +39,135 @@ class PortfolioRating(str, Enum): SELL = "Sell" +class TraderAction(str, Enum): + """3-tier transaction direction used by the Trader. + + The Trader's job is to translate the Research Manager's investment plan + into a concrete transaction proposal: should the desk execute a Buy, a + Sell, or sit on Hold this round. Position sizing and the nuanced + Overweight / Underweight calls happen later at the Portfolio Manager. + """ + + BUY = "Buy" + HOLD = "Hold" + SELL = "Sell" + + +# --------------------------------------------------------------------------- +# Research Manager +# --------------------------------------------------------------------------- + + +class ResearchPlan(BaseModel): + """Structured investment plan produced by the Research Manager. + + Hand-off to the Trader: the recommendation pins the directional view, + the rationale captures which side of the bull/bear debate carried the + argument, and the strategic actions translate that into concrete + instructions the trader can execute against. + """ + + recommendation: PortfolioRating = Field( + description=( + "The investment recommendation. Exactly one of Buy / Overweight / " + "Hold / Underweight / Sell. Reserve Hold for situations where the " + "evidence on both sides is genuinely balanced; otherwise commit to " + "the side with the stronger arguments." + ), + ) + rationale: str = Field( + description=( + "Conversational summary of the key points from both sides of the " + "debate, ending with which arguments led to the recommendation. " + "Speak naturally, as if to a teammate." + ), + ) + strategic_actions: str = Field( + description=( + "Concrete steps for the trader to implement the recommendation, " + "including position sizing guidance consistent with the rating." + ), + ) + + +def render_research_plan(plan: ResearchPlan) -> str: + """Render a ResearchPlan to markdown for storage and the trader's prompt context.""" + return "\n".join([ + f"**Recommendation**: {plan.recommendation.value}", + "", + f"**Rationale**: {plan.rationale}", + "", + f"**Strategic Actions**: {plan.strategic_actions}", + ]) + + +# --------------------------------------------------------------------------- +# Trader +# --------------------------------------------------------------------------- + + +class TraderProposal(BaseModel): + """Structured transaction proposal produced by the Trader. + + The trader reads the Research Manager's investment plan and the analyst + reports, then turns them into a concrete transaction: what action to + take, the reasoning that justifies it, and the practical levels for + entry, stop-loss, and sizing. + """ + + action: TraderAction = Field( + description="The transaction direction. Exactly one of Buy / Hold / Sell.", + ) + reasoning: str = Field( + description=( + "The case for this action, anchored in the analysts' reports and " + "the research plan. Two to four sentences." + ), + ) + entry_price: Optional[float] = Field( + default=None, + description="Optional entry price target in the instrument's quote currency.", + ) + stop_loss: Optional[float] = Field( + default=None, + description="Optional stop-loss price in the instrument's quote currency.", + ) + position_sizing: Optional[str] = Field( + default=None, + description="Optional sizing guidance, e.g. '5% of portfolio'.", + ) + + +def render_trader_proposal(proposal: TraderProposal) -> str: + """Render a TraderProposal to markdown. + + The trailing ``FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**`` line is + preserved for backward compatibility with the analyst stop-signal text + and any external code that greps for it. + """ + parts = [ + f"**Action**: {proposal.action.value}", + "", + f"**Reasoning**: {proposal.reasoning}", + ] + if proposal.entry_price is not None: + parts.extend(["", f"**Entry Price**: {proposal.entry_price}"]) + if proposal.stop_loss is not None: + parts.extend(["", f"**Stop Loss**: {proposal.stop_loss}"]) + if proposal.position_sizing: + parts.extend(["", f"**Position Sizing**: {proposal.position_sizing}"]) + parts.extend([ + "", + f"FINAL TRANSACTION PROPOSAL: **{proposal.action.value.upper()}**", + ]) + return "\n".join(parts) + + +# --------------------------------------------------------------------------- +# Portfolio Manager +# --------------------------------------------------------------------------- + + class PortfolioDecision(BaseModel): """Structured output produced by the Portfolio Manager. diff --git a/tradingagents/agents/trader/trader.py b/tradingagents/agents/trader/trader.py index 0ecae8888..ea3f6b232 100644 --- a/tradingagents/agents/trader/trader.py +++ b/tradingagents/agents/trader/trader.py @@ -1,32 +1,60 @@ +"""Trader: turns the Research Manager's investment plan into a concrete transaction proposal.""" + +from __future__ import annotations + import functools +from langchain_core.messages import AIMessage + +from tradingagents.agents.schemas import TraderProposal, render_trader_proposal from tradingagents.agents.utils.agent_utils import build_instrument_context +from tradingagents.agents.utils.structured import ( + bind_structured, + invoke_structured_or_freetext, +) def create_trader(llm): + structured_llm = bind_structured(llm, TraderProposal, "Trader") + def trader_node(state, name): company_name = state["company_of_interest"] instrument_context = build_instrument_context(company_name) investment_plan = state["investment_plan"] - context = { - "role": "user", - "content": f"Based on a comprehensive analysis by a team of analysts, here is an investment plan tailored for {company_name}. {instrument_context} This plan incorporates insights from current technical market trends, macroeconomic indicators, and social media sentiment. Use this plan as a foundation for evaluating your next trading decision.\n\nProposed Investment Plan: {investment_plan}\n\nLeverage these insights to make an informed and strategic decision.", - } - messages = [ { "role": "system", - "content": "You are a trading agent analyzing market data to make investment decisions. Based on your analysis, provide a specific recommendation to buy, sell, or hold. End with a firm decision and always conclude your response with 'FINAL TRANSACTION PROPOSAL: **BUY/HOLD/SELL**' to confirm your recommendation.", + "content": ( + "You are a trading agent analyzing market data to make investment decisions. " + "Based on your analysis, provide a specific recommendation to buy, sell, or hold. " + "Anchor your reasoning in the analysts' reports and the research plan." + ), + }, + { + "role": "user", + "content": ( + f"Based on a comprehensive analysis by a team of analysts, here is an investment " + f"plan tailored for {company_name}. {instrument_context} This plan incorporates " + f"insights from current technical market trends, macroeconomic indicators, and " + f"social media sentiment. Use this plan as a foundation for evaluating your next " + f"trading decision.\n\nProposed Investment Plan: {investment_plan}\n\n" + f"Leverage these insights to make an informed and strategic decision." + ), }, - context, ] - result = llm.invoke(messages) + trader_plan = invoke_structured_or_freetext( + structured_llm, + llm, + messages, + render_trader_proposal, + "Trader", + ) return { - "messages": [result], - "trader_investment_plan": result.content, + "messages": [AIMessage(content=trader_plan)], + "trader_investment_plan": trader_plan, "sender": name, } diff --git a/tradingagents/agents/utils/structured.py b/tradingagents/agents/utils/structured.py new file mode 100644 index 000000000..400e1a82b --- /dev/null +++ b/tradingagents/agents/utils/structured.py @@ -0,0 +1,73 @@ +"""Shared helpers for invoking an agent with structured output and a graceful fallback. + +The Portfolio Manager, Trader, and Research Manager all follow the same +canonical pattern: + +1. At agent creation, wrap the LLM with ``with_structured_output(Schema)`` + so the model returns a typed Pydantic instance. If the provider does + not support structured output (rare; mostly older Ollama models), the + wrap is skipped and the agent uses free-text generation instead. +2. At invocation, run the structured call and render the result back to + markdown. If the structured call itself fails for any reason + (malformed JSON from a weak model, transient provider issue), fall + back to a plain ``llm.invoke`` so the pipeline never blocks. + +Centralising the pattern here keeps the agent factories small and ensures +all three agents log the same warnings when fallback fires. +""" + +from __future__ import annotations + +import logging +from typing import Any, Callable, Optional, TypeVar + +from pydantic import BaseModel + +logger = logging.getLogger(__name__) + +T = TypeVar("T", bound=BaseModel) + + +def bind_structured(llm: Any, schema: type[T], agent_name: str) -> Optional[Any]: + """Return ``llm.with_structured_output(schema)`` or ``None`` if unsupported. + + Logs a warning when the binding fails so the user understands the agent + will use free-text generation for every call instead of one-shot fallback. + """ + try: + return llm.with_structured_output(schema) + except (NotImplementedError, AttributeError) as exc: + logger.warning( + "%s: provider does not support with_structured_output (%s); " + "falling back to free-text generation", + agent_name, exc, + ) + return None + + +def invoke_structured_or_freetext( + structured_llm: Optional[Any], + plain_llm: Any, + prompt: Any, + render: Callable[[T], str], + agent_name: str, +) -> str: + """Run the structured call and render to markdown; fall back to free-text on any failure. + + ``prompt`` is whatever the underlying LLM accepts (a string for chat + invocations, a list of message dicts for chat models that take that + shape). The same value is forwarded to the free-text path so the + fallback sees the same input the structured call did. + """ + if structured_llm is not None: + try: + result = structured_llm.invoke(prompt) + return render(result) + except Exception as exc: + logger.warning( + "%s: structured-output invocation failed (%s); retrying once as free text", + agent_name, exc, + ) + + response = plain_llm.invoke(prompt) + return response.content From 4016fd4efa8d43ddde9d00cc59531490b9ac6a9a Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 20:54:19 +0000 Subject: [PATCH 09/12] fix: stop leaking OpenAI base_url into non-OpenAI provider clients Default config had backend_url='https://api.openai.com/v1' which was forwarded to every provider client, including Google. ChatGoogleGenerativeAI constructed requests against that base, producing malformed URLs like https://api.openai.com/v1/v1beta/models/gemini-2.5-flash:generateContent that 404 with empty body. Discovered while running propagate() against Gemini end-to-end. The structured-output smoke worked because that path constructed the LLM without going through the factory and without forwarding backend_url; propagate() goes through TradingAgentsGraph.__init__ which forwards config['backend_url'] to every provider. Fix: default to None. Each provider client falls back to its own endpoint (api.openai.com for OpenAI via _PROVIDER_CONFIG, Gemini's default for Google, and so on). The CLI flow already sets backend_url explicitly per provider when the user picks one, so that path is unchanged. Verified: full propagate() now passes end-to-end on both OpenAI gpt-5.4-mini and Gemini gemini-3-flash-preview, with all nine structure/log/signal checks green for each. --- tradingagents/default_config.py | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/tradingagents/default_config.py b/tradingagents/default_config.py index 89b517659..7498d1883 100644 --- a/tradingagents/default_config.py +++ b/tradingagents/default_config.py @@ -11,7 +11,12 @@ DEFAULT_CONFIG = { "llm_provider": "openai", "deep_think_llm": "gpt-5.4", "quick_think_llm": "gpt-5.4-mini", - "backend_url": "https://api.openai.com/v1", + # When None, each provider's client falls back to its own default endpoint + # (api.openai.com for OpenAI, generativelanguage.googleapis.com for Gemini, ...). + # The CLI overrides this per provider when the user picks one. Keeping a + # provider-specific URL here would leak (e.g. OpenAI's /v1 was previously + # being forwarded to Gemini, producing malformed request URLs). + "backend_url": None, # Provider-specific thinking configuration "google_thinking_level": None, # "high", "minimal", etc. "openai_reasoning_effort": None, # "medium", "high", "low" From 7c37249f808f9c169ad2198dc384166e7ca7adf9 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Sat, 25 Apr 2026 21:54:30 +0000 Subject: [PATCH 10/12] =?UTF-8?q?chore:=20release=20v0.2.4=20=E2=80=94=20s?= =?UTF-8?q?tructured=20agents,=20checkpoint,=20memory=20log,=20providers?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This release bundles substantial work since v0.2.3: - Structured-output Research Manager, Trader, and Portfolio Manager (canonical with_structured_output pattern, single LLM call per agent, rendered markdown preserves the existing report shape). - LangGraph checkpoint resume for crash recovery (--checkpoint flag). - Persistent decision log replacing the per-agent BM25 memory, with deferred reflection driven by yfinance returns + alpha vs SPY. - DeepSeek, Qwen, GLM, and Azure OpenAI provider support; dynamic OpenRouter model selection. - Docker support; cache and logs moved to ~/.tradingagents/ to fix Docker permission issues. - Windows UTF-8 encoding fix on every file I/O site. - 5-tier rating consistency (Buy / Overweight / Hold / Underweight / Sell) across Research Manager, Portfolio Manager, signal processor, memory log. Plus the small quality items in this commit: 1. Suppress noisy Pydantic serializer warnings from OpenAI Responses-API parse path by defaulting structured-output to method="function_calling" (root-cause fix, not a warnings filter — same typed result, no warnings). 2. Ship scripts/smoke_structured_output.py so contributors can verify their provider's structured-output path with one command. 3. Add opt-in memory_log_max_entries config — when set, oldest resolved memory log entries are pruned once the cap is exceeded; pending entries (unresolved) are never pruned. 4. backend_url default changed from the OpenAI URL to None so the per-provider client falls back to its native endpoint instead of leaking OpenAI's URL into Gemini / other clients. CHANGELOG.md added with the full v0.2.4 entry. 92 tests pass without API keys. --- CHANGELOG.md | 266 +++++++++++++++++++++ README.md | 3 + pyproject.toml | 2 +- scripts/smoke_structured_output.py | 176 ++++++++++++++ tests/test_memory_log.py | 53 ++++ tradingagents/agents/utils/memory.py | 44 +++- tradingagents/default_config.py | 4 + tradingagents/llm_clients/openai_client.py | 16 ++ 8 files changed, 562 insertions(+), 2 deletions(-) create mode 100644 CHANGELOG.md create mode 100644 scripts/smoke_structured_output.py diff --git a/CHANGELOG.md b/CHANGELOG.md new file mode 100644 index 000000000..3fd6afa2b --- /dev/null +++ b/CHANGELOG.md @@ -0,0 +1,266 @@ +# Changelog + +All notable changes to TradingAgents are documented here. + +The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), +and this project follows [Semantic Versioning](https://semver.org/spec/v2.0.0.html). +Breaking changes within the 0.x line are called out explicitly. + +## [0.2.4] — 2026-04-25 + +### Added + +- **Structured-output decision agents.** Research Manager, Trader, and Portfolio + Manager now use `llm.with_structured_output(Schema)` on their primary call + and return typed Pydantic instances. Each provider's native structured-output + mode is used (`json_schema` for OpenAI / xAI, `response_schema` for Gemini, + tool-use for Anthropic, function-calling for OpenAI-compatible providers). + Render helpers preserve the existing markdown shape so memory log, CLI + display, and saved reports keep working unchanged. (#434) +- **LangGraph checkpoint resume** — opt-in via `--checkpoint`. State is saved + after each node so crashed or interrupted runs resume from the last + successful step. Per-ticker SQLite databases under + `~/.tradingagents/cache/checkpoints/`. `--clear-checkpoints` resets them. (#594) +- **Persistent decision log** replacing the per-agent BM25 memory. Decisions + are stored automatically at the end of `propagate()`; the next same-ticker + run resolves prior pending entries with realised return, alpha vs SPY, and + a one-paragraph reflection. Override path with `TRADINGAGENTS_MEMORY_LOG_PATH`. + Optional `memory_log_max_entries` config caps resolved entries; pending + entries are never pruned. (#578, #563, #564, #579) +- **DeepSeek, Qwen (Alibaba DashScope), GLM (Zhipu), and Azure OpenAI** + providers, plus dynamic OpenRouter model selection. +- **Docker support** — multi-stage build with separate dev and runtime images. +- **`scripts/smoke_structured_output.py`** — diagnostic that exercises the + three structured-output agents against any provider so contributors can + verify their setup with one command. +- **5-tier rating scale** (Buy / Overweight / Hold / Underweight / Sell) used + consistently by Research Manager, Portfolio Manager, signal processor, and + the memory log; Trader keeps 3-tier (Buy / Hold / Sell) since transaction + direction is naturally ternary. +- **Pytest fixtures** — lazy LLM client imports plus placeholder API keys so + the test suite runs cleanly without credentials. (#588) + +### Changed + +- **`backend_url` default is now `None`** rather than the OpenAI URL. Each + provider client falls back to its native default. The previous default + leaked the OpenAI URL into non-OpenAI clients (e.g. Gemini), producing + malformed request URLs for Python users who switched providers without + overriding `backend_url`. The CLI flow is unaffected. +- All file I/O passes explicit `encoding="utf-8"` so Windows users no longer + hit `UnicodeEncodeError` with the cp1252 default. (#543, #550, #576) +- Cache and log directories moved to `~/.tradingagents/` to resolve Docker + permission issues. (#519) +- `SignalProcessor` reads the rating from the Portfolio Manager's rendered + markdown via a deterministic heuristic — no extra LLM call. +- OpenAI structured-output calls default to `method="function_calling"` to + avoid noisy `PydanticSerializationUnexpectedValue` warnings emitted by + langchain-openai's Responses-API parse path. Same typed result, no warnings. + +### Fixed + +- Empty memory no longer triggers fabricated past-lessons in agent prompts; + the memory-log redesign makes this structurally impossible since only the + Portfolio Manager consults memory and only when entries exist. (#572) +- Tool-call logging processes every chunk message, not just the last one, and + memory score normalization handles empty score arrays. (#534, #531) + +### Removed + +- `FinancialSituationMemory` (the per-agent BM25 system) and the dead + `reflect_and_remember()` plumbing; subsumed by the persistent decision log. +- Hardcoded Google endpoint that caused 404 when `langchain-google-genai` + changed its API path. (#493, #496) + +### Contributors + +Thanks to everyone who shaped this release through code, design, and reports: + +- [@claytonbrown](https://github.com/claytonbrown) — checkpoint resume (#594), test fixtures (#588), design feedback on cost tracking (#582) and structured validation (#583) +- [@Bcardo](https://github.com/Bcardo) — memory-log redesign (#579), empty-memory hallucination report (#572), encoding fix proposal (#570) +- [@voidborne-d](https://github.com/voidborne-d) — memory persistence design (#564), portfolio manager state fix (#503) +- [@mannubaveja007](https://github.com/mannubaveja007) — structured-output feature request (#434) +- [@kelder66](https://github.com/kelder66) — RAM-only memory issue (#563) +- [@Gujiassh](https://github.com/Gujiassh) — tool-call logging fix (#534), test stub PR (#533) +- [@iuyup](https://github.com/iuyup) — memory score normalization fix (#531) +- [@kaihg](https://github.com/kaihg) — Google base_url fix (#496) +- [@32ryh98yfe](https://github.com/32ryh98yfe) — Gemini 404 report (#493) +- [@uppb](https://github.com/uppb) — OpenRouter dynamic model selection (#482) +- [@guoz14](https://github.com/guoz14) — OpenRouter limited-model report (#337) +- [@samchenku](https://github.com/samchenku) — indicator name normalization (#490) +- [@JasonOA888](https://github.com/JasonOA888) — y_finance pandas import fix (#488) +- [@tiffanychum](https://github.com/tiffanychum) — stale import cleanup (#499) +- [@zaizou](https://github.com/zaizou) — Docker permission issue (#519) +- [@Stosman123](https://github.com/Stosman123), [@mauropuga](https://github.com/mauropuga), [@hotwind2015](https://github.com/hotwind2015) — Windows encoding bug reports (#543, #550, #576) +- [@nnishad](https://github.com/nnishad), [@atharvajoshi01](https://github.com/atharvajoshi01) — encoding fix proposals (#568, #549) + +## [0.2.3] — 2026-03-29 + +### Added + +- **Multi-language output** for analyst reports and final decisions, with a + CLI selector. Internal agent debate stays in English for reasoning quality. (#472) +- **GPT-5.4 family models** in the default catalog, with deep/quick model split. +- **Unified model catalog** as a single source of truth for CLI options and + provider validation. + +### Changed + +- `base_url` is forwarded to Google and Anthropic clients so corporate proxies + work consistently across providers. (#427) +- Standardised the Google `api_key` parameter to the unified `api_key` form. + +### Fixed + +- Backtesting fetchers no longer leak look-ahead data when `curr_date` is in + the middle of a fetched window. (#475) +- Invalid indicator names from the LLM are caught at the tool boundary instead + of crashing the run. (#429) +- yfinance news fetchers respect the same exponential-backoff retry as price + fetchers. (#445) + +### Contributors + +- [@ahmedk20](https://github.com/ahmedk20) — multi-language output (#472) +- [@CadeYu](https://github.com/CadeYu) — model catalog typing (#464) +- [@javierdejesusda](https://github.com/javierdejesusda) — unified Google API key parameter (#453) +- [@voidborne-d](https://github.com/voidborne-d) — yfinance news retry (#445) +- [@kostakost2](https://github.com/kostakost2) — look-ahead bias report (#475) +- [@lu-zhengda](https://github.com/lu-zhengda) — proxy/base_url support request (#427) +- [@VamsiKrishna2021](https://github.com/VamsiKrishna2021) — invalid indicator crash report (#429) + +## [0.2.2] — 2026-03-22 + +### Added + +- **Five-tier rating scale** (Buy / Overweight / Hold / Underweight / Sell) + introduced for the Portfolio Manager. +- **Anthropic effort level** support for Claude models. +- **OpenAI Responses API** path for native OpenAI models. + +### Changed + +- `risk_manager` renamed to `portfolio_manager` to match the role description + shown in the CLI display. +- Exchange-qualified tickers (e.g. `7203.T`, `BRK.B`) preserved across all + agent prompts and tool calls. +- Process-level UTF-8 default attempted for cross-platform consistency + (note: this approach did not actually take effect; replaced in v0.2.4 with + explicit per-call `encoding="utf-8"` arguments). + +### Fixed + +- yfinance rate-limit errors are retried with exponential backoff. (#426) +- HTTP client SSL customisation is supported for environments that need + custom certificate bundles. (#379) +- Report-section writes handle list-of-string content gracefully. + +### Contributors + +- [@CadeYu](https://github.com/CadeYu) — exchange-qualified ticker preservation (#413) +- [@yang1002378395-cmyk](https://github.com/yang1002378395-cmyk) — HTTP client SSL customisation (#379) + +## [0.2.1] — 2026-03-15 + +### Security + +- Patched `langchain-core` vulnerability (LangGrinch). (#335) +- Removed `chainlit` dependency affected by CVE-2026-22218. + +### Added + +- `pyproject.toml` build-system configuration; the project now installs via + modern packaging tooling. + +### Removed + +- `setup.py` — dependencies consolidated to `pyproject.toml`. + +### Fixed + +- Risk manager reads the correct fundamental report source. (#341) +- All `open()` calls receive an explicit UTF-8 encoding (initial pass). +- `get_indicators` tool handles comma-separated indicator names from the LLM. (#368) +- `Propagation` initialises every debate-state field so risk debaters never + see missing keys. +- Stock data parsing tolerates malformed CSVs and NaN values. +- Conditional debate logic respects the configured round count. (#361) + +### Contributors + +- [@RinZ27](https://github.com/RinZ27) — `langchain-core` security patch (#335) +- [@Ljx-007](https://github.com/Ljx-007) — risk manager fundamental-report fix (#341) +- [@makk9](https://github.com/makk9) — debate-rounds config issue (#361) + +## [0.2.0] — 2026-02-04 + +This is the largest release since the initial public version. The framework +moved from single-provider to a multi-provider architecture and grew several +production-ready surfaces. + +### Added + +- **Multi-provider LLM support** (OpenAI, Google, Anthropic, xAI, OpenRouter, + Ollama) via a factory pattern, with provider-specific thinking configurations. +- **Alpha Vantage** integration as a configurable primary data provider, with + yfinance as a community-stability fallback. +- **Footer statistics** in the CLI: real-time tracking of LLM calls, tool + calls, and token usage via LangChain callbacks. +- **Post-analysis report saving** — the framework writes per-section markdown + files (analyst reports, debate transcripts, final decision) when a run + completes. +- **Announcements panel** — fetches updates from `api.tauric.ai/v1/announcements` + for the CLI welcome screen. +- **Tool fallbacks** so a single vendor outage does not stop the pipeline. + +### Changed + +- Risky / Safe risk debaters renamed to **Aggressive / Conservative** for + consistency with the displayed agent labels. +- Default data vendor switched to balance reliability and quota across + community deployments. +- Ollama and OpenRouter model lists updated; default endpoints clarified. + +### Fixed + +- Analyst status tracking and message deduplication in the live display. +- Infinite-loop guard in the agent loop; reflection and logging hardened. +- Various data-vendor implementation bugs and tool-signature mismatches. + +### Contributors + +This release is the first with substantial outside contributions; many community +PRs from late 2025 also landed here. + +- [@luohy15](https://github.com/luohy15) — Alpha Vantage data-vendor integration (#235) +- [@EdwardoSunny](https://github.com/EdwardoSunny) — yfinance fetching optimisations (#245) +- [@Mirza-Samad-Ahmed-Baig](https://github.com/Mirza-Samad-Ahmed-Baig) — infinite-loop guard, reflection, and logging fixes (#89) +- [@ZeroAct](https://github.com/ZeroAct) — saved results path support (#29) +- [@Zhongyi-Lu](https://github.com/Zhongyi-Lu) — `.env` gitignore (#49) +- [@csoboy](https://github.com/csoboy) — local Ollama setup (#53) +- [@chauhang](https://github.com/chauhang) — initial Docker support attempt (#47, later reverted; the merged Docker support shipped in v0.2.4) + +## [0.1.1] — 2025-06-07 + +### Removed + +- Static site assets that had been bundled with v0.1.0; the public site now + lives separately. + +## [0.1.0] — 2025-06-05 + +### Added + +- **Initial public release** of the TradingAgents multi-agent trading + framework: market / sentiment / news / fundamentals analysts; bull and bear + researchers; trader; aggressive, conservative, and neutral risk debaters; + portfolio manager. LangGraph orchestration, yfinance data, per-agent + BM25 memory, single-provider OpenAI integration, interactive CLI. + +[0.2.4]: https://github.com/TauricResearch/TradingAgents/compare/v0.2.3...v0.2.4 +[0.2.3]: https://github.com/TauricResearch/TradingAgents/compare/v0.2.2...v0.2.3 +[0.2.2]: https://github.com/TauricResearch/TradingAgents/compare/v0.2.1...v0.2.2 +[0.2.1]: https://github.com/TauricResearch/TradingAgents/compare/v0.2.0...v0.2.1 +[0.2.0]: https://github.com/TauricResearch/TradingAgents/compare/v0.1.1...v0.2.0 +[0.1.1]: https://github.com/TauricResearch/TradingAgents/compare/v0.1.0...v0.1.1 +[0.1.0]: https://github.com/TauricResearch/TradingAgents/releases/tag/v0.1.0 diff --git a/README.md b/README.md index 6c8f644ec..54af501a9 100644 --- a/README.md +++ b/README.md @@ -28,6 +28,7 @@ # TradingAgents: Multi-Agents LLM Financial Trading Framework ## News +- [2026-04] **TradingAgents v0.2.4** released with structured-output agents (Research Manager, Trader, Portfolio Manager), LangGraph checkpoint resume, persistent decision log, DeepSeek/Qwen/GLM/Azure provider support, Docker, and a Windows UTF-8 encoding fix. See [CHANGELOG.md](CHANGELOG.md) for the full list. - [2026-03] **TradingAgents v0.2.3** released with multi-language support, GPT-5.4 family models, unified model catalog, backtesting date fidelity, and proxy support. - [2026-03] **TradingAgents v0.2.2** released with GPT-5.4/Gemini 3.1/Claude 4.6 model coverage, five-tier rating scale, OpenAI Responses API, Anthropic effort control, and cross-platform stability. - [2026-02] **TradingAgents v0.2.0** released with multi-provider LLM support (GPT-5.x, Gemini 3.x, Claude 4.x, Grok 4.x) and improved system architecture. @@ -251,6 +252,8 @@ _, decision = ta.propagate("NVDA", "2026-01-15") We welcome contributions from the community! Whether it's fixing a bug, improving documentation, or suggesting a new feature, your input helps make this project better. If you are interested in this line of research, please consider joining our open-source financial AI research community [Tauric Research](https://tauric.ai/). +Past contributions, including code, design feedback, and bug reports, are credited per release in [`CHANGELOG.md`](CHANGELOG.md). + ## Citation Please reference our work if you find *TradingAgents* provides you with some help :) diff --git a/pyproject.toml b/pyproject.toml index b569504ef..07cbbd3f7 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta" [project] name = "tradingagents" -version = "0.2.3" +version = "0.2.4" description = "TradingAgents: Multi-Agents LLM Financial Trading Framework" readme = "README.md" requires-python = ">=3.10" diff --git a/scripts/smoke_structured_output.py b/scripts/smoke_structured_output.py new file mode 100644 index 000000000..1d3cf681c --- /dev/null +++ b/scripts/smoke_structured_output.py @@ -0,0 +1,176 @@ +"""End-to-end smoke for structured-output agents against a real LLM provider. + +Runs the three decision-making agents (Research Manager, Trader, Portfolio +Manager) directly with their structured-output bindings and prints the +typed Pydantic instance + the rendered markdown for each. Use this to +verify a provider's native structured-output mode (json_schema for +OpenAI / xAI / DeepSeek / Qwen / GLM, response_schema for Gemini, tool-use +for Anthropic) returns clean instances on the schemas we ship. + +Usage: + OPENAI_API_KEY=... python scripts/smoke_structured_output.py openai + GOOGLE_API_KEY=... python scripts/smoke_structured_output.py google + ANTHROPIC_API_KEY=... python scripts/smoke_structured_output.py anthropic + DEEPSEEK_API_KEY=... python scripts/smoke_structured_output.py deepseek + +The script does NOT call propagate(), to keep the surface tight and the +cost low — it exercises only the three structured-output calls we just +added, plus the heuristic SignalProcessor. +""" + +from __future__ import annotations + +import argparse +import os +import sys + +from tradingagents.agents.managers.portfolio_manager import create_portfolio_manager +from tradingagents.agents.managers.research_manager import create_research_manager +from tradingagents.agents.trader.trader import create_trader +from tradingagents.graph.signal_processing import SignalProcessor +from tradingagents.llm_clients import create_llm_client + + +PROVIDER_DEFAULTS = { + "openai": ("gpt-5.4-mini", None), + "google": ("gemini-2.5-flash", None), + "anthropic": ("claude-sonnet-4-6", None), + "deepseek": ("deepseek-chat", None), + "qwen": ("qwen-plus", None), + "glm": ("glm-5", None), + "xai": ("grok-4", None), +} + + +# Minimal but realistic state for the three agents. +DEBATE_HISTORY = """ +Bull Analyst: NVDA's data-center revenue grew 60% YoY last quarter, driven by +Blackwell ramp; sovereign AI deals with multiple governments add a $40B+ +multi-year tailwind. Margins remain above peer average. + +Bear Analyst: Concentration risk is real — top three customers are >40% of +revenue. Any pause in hyperscaler capex would compress the multiple. China +export restrictions still cap a meaningful portion of demand. +""" + + +def _make_rm_state(): + return { + "company_of_interest": "NVDA", + "investment_debate_state": { + "history": DEBATE_HISTORY, + "bull_history": "Bull Analyst: NVDA's data-center revenue grew 60% YoY...", + "bear_history": "Bear Analyst: Concentration risk is real...", + "current_response": "", + "judge_decision": "", + "count": 1, + }, + } + + +def _make_trader_state(investment_plan: str): + return { + "company_of_interest": "NVDA", + "investment_plan": investment_plan, + } + + +def _make_pm_state(investment_plan: str, trader_plan: str): + return { + "company_of_interest": "NVDA", + "past_context": "", + "risk_debate_state": { + "history": "Aggressive: lean in. Conservative: trim. Neutral: balanced sizing.", + "aggressive_history": "Aggressive: ...", + "conservative_history": "Conservative: ...", + "neutral_history": "Neutral: ...", + "judge_decision": "", + "current_aggressive_response": "", + "current_conservative_response": "", + "current_neutral_response": "", + "count": 1, + }, + "market_report": "Market report.", + "sentiment_report": "Sentiment report.", + "news_report": "News report.", + "fundamentals_report": "Fundamentals report.", + "investment_plan": investment_plan, + "trader_investment_plan": trader_plan, + } + + +def _print_section(title: str, content: str) -> None: + bar = "=" * 70 + print(f"\n{bar}\n{title}\n{bar}\n{content}") + + +def main() -> int: + parser = argparse.ArgumentParser(description=__doc__) + parser.add_argument("provider", choices=list(PROVIDER_DEFAULTS.keys())) + parser.add_argument("--deep-model", default=None, help="Override deep_think_llm") + parser.add_argument("--quick-model", default=None, help="Override quick_think_llm") + args = parser.parse_args() + + default_model, _ = PROVIDER_DEFAULTS[args.provider] + deep_model = args.deep_model or default_model + quick_model = args.quick_model or default_model + + print(f"Provider: {args.provider}") + print(f"Deep model: {deep_model}") + print(f"Quick model: {quick_model}") + + # Build the LLM clients via the framework's factory. + deep_client = create_llm_client(provider=args.provider, model=deep_model) + quick_client = create_llm_client(provider=args.provider, model=quick_model) + deep_llm = deep_client.get_llm() + quick_llm = quick_client.get_llm() + + # 1) Research Manager + rm = create_research_manager(deep_llm) + rm_result = rm(_make_rm_state()) + investment_plan = rm_result["investment_plan"] + _print_section("[1] Research Manager — investment_plan", investment_plan) + + # 2) Trader (consumes RM's plan) + trader = create_trader(quick_llm) + trader_result = trader(_make_trader_state(investment_plan)) + trader_plan = trader_result["trader_investment_plan"] + _print_section("[2] Trader — trader_investment_plan", trader_plan) + + # 3) Portfolio Manager (consumes both) + pm = create_portfolio_manager(deep_llm) + pm_result = pm(_make_pm_state(investment_plan, trader_plan)) + final_decision = pm_result["final_trade_decision"] + _print_section("[3] Portfolio Manager — final_trade_decision", final_decision) + + # 4) SignalProcessor extracts the rating with zero LLM calls. + sp = SignalProcessor() + rating = sp.process_signal(final_decision) + _print_section("[4] SignalProcessor → rating", rating) + + # 5) Lightweight checks: each rendered output should carry the expected + # section headers so downstream consumers (memory log, CLI display, + # saved reports) keep working. + checks = [ + ("Research Manager", investment_plan, ["**Recommendation**:"]), + ("Trader", trader_plan, ["**Action**:", "FINAL TRANSACTION PROPOSAL:"]), + ("Portfolio Manager", final_decision, ["**Rating**:", "**Executive Summary**:", "**Investment Thesis**:"]), + ] + print("\n" + "=" * 70 + "\nStructure checks\n" + "=" * 70) + failures = 0 + for name, text, required in checks: + for marker in required: + ok = marker in text + print(f" {'PASS' if ok else 'FAIL'} {name}: contains {marker!r}") + failures += int(not ok) + + print() + if failures: + print(f"Smoke FAILED: {failures} structure check(s) missing.") + return 1 + print("Smoke PASSED: structured output → rendered markdown chain works for", args.provider) + return 0 + + +if __name__ == "__main__": + sys.exit(main()) diff --git a/tests/test_memory_log.py b/tests/test_memory_log.py index e0da15efc..5d7f7f844 100644 --- a/tests/test_memory_log.py +++ b/tests/test_memory_log.py @@ -291,6 +291,59 @@ class TestTradingMemoryLogCore: assert log.load_entries() == [] assert log.get_past_context("NVDA") == "" + # Rotation: opt-in cap on resolved entries + + def test_rotation_disabled_by_default(self, tmp_path): + """Without max_entries, all resolved entries are kept.""" + log = make_log(tmp_path) + for i in range(7): + _resolve_entry(log, "NVDA", f"2026-01-{i+1:02d}", DECISION_BUY, f"Lesson {i}.") + assert len(log.load_entries()) == 7 + + def test_rotation_prunes_oldest_resolved(self, tmp_path): + """When max_entries is set and exceeded, oldest resolved entries are pruned.""" + log = TradingMemoryLog({ + "memory_log_path": str(tmp_path / "trading_memory.md"), + "memory_log_max_entries": 3, + }) + # Resolve 5 entries; rotation should keep only the 3 most recent. + for i in range(5): + _resolve_entry(log, "NVDA", f"2026-01-{i+1:02d}", DECISION_BUY, f"Lesson {i}.") + entries = log.load_entries() + assert len(entries) == 3 + # Confirm the OLDEST were dropped, not the newest. + dates = [e["date"] for e in entries] + assert dates == ["2026-01-03", "2026-01-04", "2026-01-05"] + + def test_rotation_never_prunes_pending(self, tmp_path): + """Pending entries (unresolved) are kept regardless of the cap.""" + log = TradingMemoryLog({ + "memory_log_path": str(tmp_path / "trading_memory.md"), + "memory_log_max_entries": 2, + }) + # 3 resolved + 2 pending. With cap=2, only 2 resolved survive; both pending stay. + for i in range(3): + _resolve_entry(log, "NVDA", f"2026-01-{i+1:02d}", DECISION_BUY, f"Resolved {i}.") + log.store_decision("NVDA", "2026-02-01", DECISION_BUY) + log.store_decision("NVDA", "2026-02-02", DECISION_OVERWEIGHT) + # Trigger rotation by resolving one more entry — pending entries must stay. + _resolve_entry(log, "NVDA", "2026-01-04", DECISION_BUY, "Resolved 3.") + entries = log.load_entries() + pending = [e for e in entries if e["pending"]] + resolved = [e for e in entries if not e["pending"]] + assert len(pending) == 2, "pending entries must never be pruned" + assert len(resolved) == 2, f"expected 2 resolved after rotation, got {len(resolved)}" + + def test_rotation_under_cap_is_noop(self, tmp_path): + """No rotation when resolved count <= max_entries.""" + log = TradingMemoryLog({ + "memory_log_path": str(tmp_path / "trading_memory.md"), + "memory_log_max_entries": 10, + }) + for i in range(3): + _resolve_entry(log, "NVDA", f"2026-01-{i+1:02d}", DECISION_BUY, f"Lesson {i}.") + assert len(log.load_entries()) == 3 + # Rating parsing: markdown bold and numbered list formats def test_rating_parsed_from_bold_markdown(self, tmp_path): diff --git a/tradingagents/agents/utils/memory.py b/tradingagents/agents/utils/memory.py index fee5ac4a2..c94717556 100644 --- a/tradingagents/agents/utils/memory.py +++ b/tradingagents/agents/utils/memory.py @@ -17,11 +17,14 @@ class TradingMemoryLog: _REFLECTION_RE = re.compile(r"REFLECTION:\n(.*?)$", re.DOTALL) def __init__(self, config: dict = None): + cfg = config or {} self._log_path = None - path = (config or {}).get("memory_log_path") + path = cfg.get("memory_log_path") if path: self._log_path = Path(path).expanduser() self._log_path.parent.mkdir(parents=True, exist_ok=True) + # Optional cap on resolved entries. None disables rotation. + self._max_entries = cfg.get("memory_log_max_entries") # --- Write path (Phase A) --- @@ -153,6 +156,7 @@ class TradingMemoryLog: if not updated: return + new_blocks = self._apply_rotation(new_blocks) new_text = self._SEPARATOR.join(new_blocks) tmp_path = self._log_path.with_suffix(".tmp") tmp_path.write_text(new_text, encoding="utf-8") @@ -206,6 +210,7 @@ class TradingMemoryLog: if not matched: new_blocks.append(block) + new_blocks = self._apply_rotation(new_blocks) new_text = self._SEPARATOR.join(new_blocks) tmp_path = self._log_path.with_suffix(".tmp") tmp_path.write_text(new_text, encoding="utf-8") @@ -213,6 +218,43 @@ class TradingMemoryLog: # --- Helpers --- + def _apply_rotation(self, blocks: List[str]) -> List[str]: + """Drop oldest resolved blocks when their count exceeds max_entries. + + Pending blocks are always kept (they represent unprocessed work). + Returns ``blocks`` unchanged when rotation is disabled or under cap. + """ + if not self._max_entries or self._max_entries <= 0: + return blocks + + # Tag each block with (kept, is_resolved) by parsing tag-line markers. + decisions = [] + for block in blocks: + stripped = block.strip() + if not stripped: + decisions.append((block, False)) + continue + tag_line = stripped.splitlines()[0].strip() + is_resolved = ( + tag_line.startswith("[") + and tag_line.endswith("]") + and not tag_line.endswith("| pending]") + ) + decisions.append((block, is_resolved)) + + resolved_count = sum(1 for _, r in decisions if r) + if resolved_count <= self._max_entries: + return blocks + + to_drop = resolved_count - self._max_entries + kept: List[str] = [] + for block, is_resolved in decisions: + if is_resolved and to_drop > 0: + to_drop -= 1 + continue + kept.append(block) + return kept + def _parse_entry(self, raw: str) -> Optional[dict]: lines = raw.strip().splitlines() if not lines: diff --git a/tradingagents/default_config.py b/tradingagents/default_config.py index 7498d1883..fa6d5742c 100644 --- a/tradingagents/default_config.py +++ b/tradingagents/default_config.py @@ -7,6 +7,10 @@ DEFAULT_CONFIG = { "results_dir": os.getenv("TRADINGAGENTS_RESULTS_DIR", os.path.join(_TRADINGAGENTS_HOME, "logs")), "data_cache_dir": os.getenv("TRADINGAGENTS_CACHE_DIR", os.path.join(_TRADINGAGENTS_HOME, "cache")), "memory_log_path": os.getenv("TRADINGAGENTS_MEMORY_LOG_PATH", os.path.join(_TRADINGAGENTS_HOME, "memory", "trading_memory.md")), + # Optional cap on the number of resolved memory log entries. When set, + # the oldest resolved entries are pruned once this limit is exceeded. + # Pending entries are never pruned. None disables rotation entirely. + "memory_log_max_entries": None, # LLM settings "llm_provider": "openai", "deep_think_llm": "gpt-5.4", diff --git a/tradingagents/llm_clients/openai_client.py b/tradingagents/llm_clients/openai_client.py index f943124a9..bbfcd39e3 100644 --- a/tradingagents/llm_clients/openai_client.py +++ b/tradingagents/llm_clients/openai_client.py @@ -18,6 +18,22 @@ class NormalizedChatOpenAI(ChatOpenAI): def invoke(self, input, config=None, **kwargs): return normalize_content(super().invoke(input, config, **kwargs)) + def with_structured_output(self, schema, *, method=None, **kwargs): + """Wrap with structured output, defaulting to function_calling for OpenAI. + + langchain-openai's Responses-API-parse path (the default for json_schema + when use_responses_api=True) calls response.model_dump(...) on the OpenAI + SDK's union-typed parsed response, which makes Pydantic emit ~20 + PydanticSerializationUnexpectedValue warnings per call. The function-calling + path returns a plain tool-call shape that does not trigger that + serialization, so it is the cleaner choice for our combination of + use_responses_api=True + with_structured_output. Both paths use OpenAI's + strict mode and produce the same typed Pydantic instance. + """ + if method is None: + method = "function_calling" + return super().with_structured_output(schema, method=method, **kwargs) + # Kwargs forwarded from user config to ChatOpenAI _PASSTHROUGH_KWARGS = ( "timeout", "max_retries", "reasoning_effort", From 2c97bad45c773760db3ef7cb787a9fd1fbf7ac67 Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Fri, 1 May 2026 18:56:36 +0000 Subject: [PATCH 11/12] fix(security): validate ticker before using as path component (#618) The ticker symbol reaches three filesystem-path construction sites (load_ohlcv cache filename, checkpointer DB path, _log_state results directory) without validation. A value containing path separators or "../" escapes the configured cache / checkpoints / results directory. Two attack vectors: - Programmatic callers passing arbitrary ticker to propagate() - Prompt injection via fetched news content steering the LLM into tool calls with attacker-chosen ticker Fix: new safe_ticker_component() validator in tradingagents/dataflows/ utils.py applied at all three sites. Allows the standard ticker character set ([A-Za-z0-9._\-\^], up to 32 chars) and explicitly rejects dot-only values like "." and ".." which would otherwise pass the regex but traverse parent directories. Seven test cases cover the accepted formats (BRK-B, 7203.T, ^GSPC, etc.) and the rejected inputs (path separators, null bytes, whitespace, empty values, overlong strings, dot-only values). Closes #618. --- tests/test_safe_ticker_component.py | 52 +++++++++++++++++++++ tradingagents/dataflows/stockstats_utils.py | 7 ++- tradingagents/dataflows/utils.py | 35 ++++++++++++++ tradingagents/graph/checkpointer.py | 6 ++- tradingagents/graph/trading_graph.py | 7 ++- 5 files changed, 103 insertions(+), 4 deletions(-) create mode 100644 tests/test_safe_ticker_component.py diff --git a/tests/test_safe_ticker_component.py b/tests/test_safe_ticker_component.py new file mode 100644 index 000000000..3bdc02234 --- /dev/null +++ b/tests/test_safe_ticker_component.py @@ -0,0 +1,52 @@ +"""Tests for the ticker path-component validator that blocks directory traversal.""" + +import os +import unittest + +import pytest + +from tradingagents.dataflows.utils import safe_ticker_component + + +@pytest.mark.unit +class TestSafeTickerComponent(unittest.TestCase): + def test_accepts_common_ticker_formats(self): + for ticker in ("AAPL", "BRK-B", "BRK.A", "0700.HK", "7203.T", "BHP.AX", "^GSPC"): + self.assertEqual(safe_ticker_component(ticker), ticker) + + def test_rejects_path_separators(self): + for bad in (".", "..", "../etc", "a/b", "a\\b", "/abs", "..\\..\\x"): + with self.assertRaises(ValueError): + safe_ticker_component(bad) + + def test_rejects_null_byte_and_whitespace(self): + for bad in ("AAP L", "AAPL\x00", "AAPL\n", "\tAAPL"): + with self.assertRaises(ValueError): + safe_ticker_component(bad) + + def test_rejects_empty_or_non_string(self): + for bad in ("", None, 123, b"AAPL"): + with self.assertRaises(ValueError): + safe_ticker_component(bad) + + def test_rejects_overlong_input(self): + with self.assertRaises(ValueError): + safe_ticker_component("A" * 33) + + def test_rejects_dot_only_values(self): + # '.' and '..' pass the regex but traverse when used as a path + # component (e.g. ``Path(results_dir) / ticker / "logs"``). + for bad in (".", "..", "...", "...."): + with self.assertRaises(ValueError): + safe_ticker_component(bad) + + def test_traversal_string_does_not_escape_join(self): + """Sanity: sanitized values stay within base when joined.""" + base = os.path.realpath("/tmp/cache") + ticker = safe_ticker_component("AAPL") + joined = os.path.realpath(os.path.join(base, f"{ticker}.csv")) + self.assertTrue(joined.startswith(base + os.sep)) + + +if __name__ == "__main__": + unittest.main() diff --git a/tradingagents/dataflows/stockstats_utils.py b/tradingagents/dataflows/stockstats_utils.py index cb24c5d6a..260ef73cd 100644 --- a/tradingagents/dataflows/stockstats_utils.py +++ b/tradingagents/dataflows/stockstats_utils.py @@ -8,6 +8,7 @@ from stockstats import wrap from typing import Annotated import os from .config import get_config +from .utils import safe_ticker_component logger = logging.getLogger(__name__) @@ -51,6 +52,10 @@ def load_ohlcv(symbol: str, curr_date: str) -> pd.DataFrame: subsequent calls the cache is reused. Rows after curr_date are filtered out so backtests never see future prices. """ + # Reject ticker values that would escape the cache directory when + # interpolated into the cache filename (e.g. ``../../tmp/x``). + safe_symbol = safe_ticker_component(symbol) + config = get_config() curr_date_dt = pd.to_datetime(curr_date) @@ -63,7 +68,7 @@ def load_ohlcv(symbol: str, curr_date: str) -> pd.DataFrame: os.makedirs(config["data_cache_dir"], exist_ok=True) data_file = os.path.join( config["data_cache_dir"], - f"{symbol}-YFin-data-{start_str}-{end_str}.csv", + f"{safe_symbol}-YFin-data-{start_str}-{end_str}.csv", ) if os.path.exists(data_file): diff --git a/tradingagents/dataflows/utils.py b/tradingagents/dataflows/utils.py index c99b777ab..3d8a45d81 100644 --- a/tradingagents/dataflows/utils.py +++ b/tradingagents/dataflows/utils.py @@ -1,4 +1,5 @@ import os +import re import json import pandas as pd from datetime import date, timedelta, datetime @@ -6,6 +7,40 @@ from typing import Annotated SavePathType = Annotated[str, "File path to save data. If None, data is not saved."] +# Tickers can contain letters, digits, dot, dash, underscore, and caret +# (for index symbols like ^GSPC). Anything else is rejected so the value +# never escapes a containing directory when interpolated into a path. +_TICKER_PATH_RE = re.compile(r"^[A-Za-z0-9._\-\^]+$") + + +def safe_ticker_component(value: str, *, max_len: int = 32) -> str: + """Validate ``value`` is safe to interpolate into a filesystem path. + + Tickers come from user CLI input or from LLM tool calls, both of which + can be influenced by attacker-controlled content (e.g. prompt injection + embedded in fetched news). Without validation, a value like + ``"../../../etc/foo"`` flows into ``os.path.join`` / ``Path /`` and + escapes the configured cache, checkpoint, or results directory. + + Returns ``value`` unchanged when it matches the allowed pattern; raises + ``ValueError`` otherwise. + """ + if not isinstance(value, str) or not value: + raise ValueError(f"ticker must be a non-empty string, got {value!r}") + if len(value) > max_len: + raise ValueError(f"ticker exceeds {max_len} chars: {value!r}") + if not _TICKER_PATH_RE.fullmatch(value): + raise ValueError( + f"ticker contains characters not allowed in a filesystem path: {value!r}" + ) + # The regex above allows '.', so values like '.', '..', '...' would pass, + # and as a path component they traverse the parent directory. Reject any + # value that's only dots. + if set(value) == {"."}: + raise ValueError(f"ticker cannot consist solely of dots: {value!r}") + return value + + def save_output(data: pd.DataFrame, tag: str, save_path: SavePathType = None) -> None: if save_path: data.to_csv(save_path, encoding="utf-8") diff --git a/tradingagents/graph/checkpointer.py b/tradingagents/graph/checkpointer.py index 7a73ee446..3ba19726d 100644 --- a/tradingagents/graph/checkpointer.py +++ b/tradingagents/graph/checkpointer.py @@ -13,12 +13,16 @@ from typing import Generator from langgraph.checkpoint.sqlite import SqliteSaver +from tradingagents.dataflows.utils import safe_ticker_component + def _db_path(data_dir: str | Path, ticker: str) -> Path: """Return the SQLite checkpoint DB path for a ticker.""" + # Reject ticker values that would escape the checkpoints directory. + safe = safe_ticker_component(ticker).upper() p = Path(data_dir) / "checkpoints" p.mkdir(parents=True, exist_ok=True) - return p / f"{ticker.upper()}.db" + return p / f"{safe}.db" def thread_id(ticker: str, date: str) -> str: diff --git a/tradingagents/graph/trading_graph.py b/tradingagents/graph/trading_graph.py index bd6f1fc5c..d7e8b5731 100644 --- a/tradingagents/graph/trading_graph.py +++ b/tradingagents/graph/trading_graph.py @@ -18,6 +18,7 @@ from tradingagents.llm_clients import create_llm_client from tradingagents.agents import * from tradingagents.default_config import DEFAULT_CONFIG from tradingagents.agents.utils.memory import TradingMemoryLog +from tradingagents.dataflows.utils import safe_ticker_component from tradingagents.agents.utils.agent_states import ( AgentState, InvestDebateState, @@ -378,8 +379,10 @@ class TradingAgentsGraph: "final_trade_decision": final_state["final_trade_decision"], } - # Save to file - directory = Path(self.config["results_dir"]) / self.ticker / "TradingAgentsStrategy_logs" + # Save to file. Reject ticker values that would escape the + # results directory when joined as a path component. + safe_ticker = safe_ticker_component(self.ticker) + directory = Path(self.config["results_dir"]) / safe_ticker / "TradingAgentsStrategy_logs" directory.mkdir(parents=True, exist_ok=True) log_path = directory / f"full_states_log_{trade_date}.json" From 7e9e7b83c7fcc18d941300b253c6ed24d985788d Mon Sep 17 00:00:00 2001 From: Yijia-Xiao Date: Fri, 1 May 2026 19:23:23 +0000 Subject: [PATCH 12/12] feat: DeepSeek V4 thinking-mode round-trip via DeepSeekChatOpenAI subclass Resolves #599: thinking-mode models require reasoning_content to be echoed back across turns; multi-turn agent runs failed with HTTP 400. The fix isolates DeepSeek's quirks (reasoning_content round-trip and the deepseek-reasoner no-tool_choice limitation) into a subclass so the general OpenAI-compatible client stays untouched. Adds DeepSeek V4 Pro/Flash to the catalog. 9 new tests; rationale documented in the class docstrings. Design adapted from #600; #611 closed in favour of this approach. --- tests/test_deepseek_reasoning.py | 169 +++++++++++++++++++++ tradingagents/llm_clients/model_catalog.py | 2 + tradingagents/llm_clients/openai_client.py | 108 ++++++++++--- 3 files changed, 262 insertions(+), 17 deletions(-) create mode 100644 tests/test_deepseek_reasoning.py diff --git a/tests/test_deepseek_reasoning.py b/tests/test_deepseek_reasoning.py new file mode 100644 index 000000000..fb300336d --- /dev/null +++ b/tests/test_deepseek_reasoning.py @@ -0,0 +1,169 @@ +"""Tests for DeepSeekChatOpenAI thinking-mode behaviour. + +Two pieces verified: + +1. ``reasoning_content`` is captured on receive into the AIMessage's + ``additional_kwargs`` and re-attached on send so DeepSeek's API + sees the same value across turns. +2. ``with_structured_output`` raises NotImplementedError for + ``deepseek-reasoner`` so the agent factories' free-text fallback + handles the request instead of failing at runtime. +""" + +import os + +import pytest +from langchain_core.messages import AIMessage, HumanMessage +from langchain_core.prompt_values import ChatPromptValue + +from tradingagents.llm_clients.openai_client import ( + DeepSeekChatOpenAI, + NormalizedChatOpenAI, + _input_to_messages, +) + + +# --------------------------------------------------------------------------- +# _input_to_messages — the helper that handles list / ChatPromptValue / other +# (Gemini bot review note: non-list inputs must also work) +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestInputToMessages: + def test_list_input_returned_as_is(self): + msgs = [HumanMessage(content="hi")] + assert _input_to_messages(msgs) is msgs + + def test_chat_prompt_value_unwrapped(self): + msgs = [HumanMessage(content="hi")] + prompt_value = ChatPromptValue(messages=msgs) + assert _input_to_messages(prompt_value) == msgs + + def test_string_input_yields_empty_list(self): + # A bare string isn't a message-bearing input; the caller's normal + # langchain conversion happens upstream of _get_request_payload. + assert _input_to_messages("hello") == [] + + +# --------------------------------------------------------------------------- +# Reasoning content propagation across turns +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestDeepSeekReasoningContent: + def _client(self): + os.environ.setdefault("DEEPSEEK_API_KEY", "placeholder") + return DeepSeekChatOpenAI( + model="deepseek-v4-flash", + api_key="placeholder", + base_url="https://api.deepseek.com", + ) + + def test_capture_on_receive(self): + """When the response carries reasoning_content, it lands on the + AIMessage's additional_kwargs so the next turn can echo it back.""" + client = self._client() + result = client._create_chat_result( + { + "model": "deepseek-v4-flash", + "choices": [ + { + "index": 0, + "message": { + "role": "assistant", + "content": "Plan: buy NVDA.", + "reasoning_content": "Step 1: trend is up. Step 2: ...", + }, + "finish_reason": "stop", + } + ], + "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}, + } + ) + ai = result.generations[0].message + assert ai.additional_kwargs["reasoning_content"] == "Step 1: trend is up. Step 2: ..." + + def test_propagate_on_send(self): + """When an outgoing AIMessage carries reasoning_content, the request + payload echoes it on the corresponding message dict.""" + client = self._client() + prior = AIMessage( + content="Plan", + additional_kwargs={"reasoning_content": "weighed bull case"}, + ) + new_user = HumanMessage(content="Refine.") + payload = client._get_request_payload([prior, new_user]) + # Find the assistant message in the payload + assistant_dicts = [m for m in payload["messages"] if m.get("role") == "assistant"] + assert assistant_dicts, "assistant message missing from outgoing payload" + assert assistant_dicts[0]["reasoning_content"] == "weighed bull case" + + def test_propagate_through_chat_prompt_value(self): + """Gemini bot review note: non-list inputs (ChatPromptValue) must + also propagate reasoning_content.""" + client = self._client() + prior = AIMessage( + content="Plan", + additional_kwargs={"reasoning_content": "weighed bull case"}, + ) + prompt_value = ChatPromptValue(messages=[prior, HumanMessage(content="Refine.")]) + payload = client._get_request_payload(prompt_value) + assistant_dicts = [m for m in payload["messages"] if m.get("role") == "assistant"] + assert assistant_dicts[0]["reasoning_content"] == "weighed bull case" + + +# --------------------------------------------------------------------------- +# deepseek-reasoner: structured output unavailable, falls through to free-text +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestDeepSeekReasonerStructuredOutput: + def test_with_structured_output_raises_for_reasoner(self): + client = DeepSeekChatOpenAI( + model="deepseek-reasoner", + api_key="placeholder", + base_url="https://api.deepseek.com", + ) + from pydantic import BaseModel + + class _Sample(BaseModel): + answer: str + + with pytest.raises(NotImplementedError): + client.with_structured_output(_Sample) + + def test_with_structured_output_works_for_v4(self): + """V4 models (non-reasoner) accept tool_choice; structured output works.""" + client = DeepSeekChatOpenAI( + model="deepseek-v4-flash", + api_key="placeholder", + base_url="https://api.deepseek.com", + ) + from pydantic import BaseModel + + class _Sample(BaseModel): + answer: str + + # Should return a Runnable, not raise. (The actual API call would + # require a real key; we only assert binding succeeds.) + wrapped = client.with_structured_output(_Sample) + assert wrapped is not None + + +# --------------------------------------------------------------------------- +# Base class isolation: NormalizedChatOpenAI does NOT have DeepSeek behaviour +# --------------------------------------------------------------------------- + + +@pytest.mark.unit +class TestBaseClassIsolation: + def test_normalized_does_not_propagate_reasoning_content(self): + """The general-purpose NormalizedChatOpenAI must not carry + DeepSeek-specific behaviour. Only the subclass does.""" + assert not hasattr(NormalizedChatOpenAI, "_get_request_payload") or ( + NormalizedChatOpenAI._get_request_payload + is NormalizedChatOpenAI.__bases__[0]._get_request_payload + ) diff --git a/tradingagents/llm_clients/model_catalog.py b/tradingagents/llm_clients/model_catalog.py index a2c57ed89..9a723a8b9 100644 --- a/tradingagents/llm_clients/model_catalog.py +++ b/tradingagents/llm_clients/model_catalog.py @@ -65,10 +65,12 @@ MODEL_OPTIONS: ProviderModeOptions = { }, "deepseek": { "quick": [ + ("DeepSeek V4 Flash - Latest V4 fast model", "deepseek-v4-flash"), ("DeepSeek V3.2", "deepseek-chat"), ("Custom model ID", "custom"), ], "deep": [ + ("DeepSeek V4 Pro - Latest V4 flagship model", "deepseek-v4-pro"), ("DeepSeek V3.2 (thinking)", "deepseek-reasoner"), ("DeepSeek V3.2", "deepseek-chat"), ("Custom model ID", "custom"), diff --git a/tradingagents/llm_clients/openai_client.py b/tradingagents/llm_clients/openai_client.py index bbfcd39e3..b74e26ef4 100644 --- a/tradingagents/llm_clients/openai_client.py +++ b/tradingagents/llm_clients/openai_client.py @@ -1,6 +1,7 @@ import os from typing import Any, Optional +from langchain_core.messages import AIMessage from langchain_openai import ChatOpenAI from .base_client import BaseLLMClient, normalize_content @@ -11,29 +12,97 @@ class NormalizedChatOpenAI(ChatOpenAI): """ChatOpenAI with normalized content output. The Responses API returns content as a list of typed blocks - (reasoning, text, etc.). This normalizes to string for consistent - downstream handling. + (reasoning, text, etc.). ``invoke`` normalizes to string for + consistent downstream handling. ``with_structured_output`` defaults + to function-calling so the Responses-API parse path is avoided + (langchain-openai's parse path emits noisy + PydanticSerializationUnexpectedValue warnings per call without + affecting correctness). + + Provider-specific quirks (e.g. DeepSeek's thinking mode) live in + purpose-built subclasses below so this base class stays small. """ def invoke(self, input, config=None, **kwargs): return normalize_content(super().invoke(input, config, **kwargs)) def with_structured_output(self, schema, *, method=None, **kwargs): - """Wrap with structured output, defaulting to function_calling for OpenAI. - - langchain-openai's Responses-API-parse path (the default for json_schema - when use_responses_api=True) calls response.model_dump(...) on the OpenAI - SDK's union-typed parsed response, which makes Pydantic emit ~20 - PydanticSerializationUnexpectedValue warnings per call. The function-calling - path returns a plain tool-call shape that does not trigger that - serialization, so it is the cleaner choice for our combination of - use_responses_api=True + with_structured_output. Both paths use OpenAI's - strict mode and produce the same typed Pydantic instance. - """ if method is None: method = "function_calling" return super().with_structured_output(schema, method=method, **kwargs) + +def _input_to_messages(input_: Any) -> list: + """Normalise a langchain LLM input to a list of message objects. + + Accepts a list of messages, a ``ChatPromptValue`` (from a + ChatPromptTemplate), or anything else (treated as no messages). + Used by providers that need to walk the outgoing message history; + in particular DeepSeek thinking-mode propagation must work for + both bare-list invocations and ChatPromptTemplate-driven ones, so + treating only ``list`` here would silently skip half the call sites. + """ + if isinstance(input_, list): + return input_ + if hasattr(input_, "to_messages"): + return input_.to_messages() + return [] + + +class DeepSeekChatOpenAI(NormalizedChatOpenAI): + """DeepSeek-specific overrides on top of the OpenAI-compatible client. + + Two quirks that don't apply to other OpenAI-compatible providers: + + 1. **Thinking-mode round-trip.** When DeepSeek's thinking models return + a response with ``reasoning_content``, that field must be echoed + back as part of the assistant message on the next turn or the API + fails with HTTP 400. ``_create_chat_result`` captures the field on + receive and ``_get_request_payload`` re-attaches it on send. + + 2. **deepseek-reasoner has no tool_choice.** Structured output via + function-calling is unavailable, so we raise NotImplementedError + and let the agent factories fall back to free-text generation + (see ``tradingagents/agents/utils/structured.py``). + """ + + def _get_request_payload(self, input_, *, stop=None, **kwargs): + payload = super()._get_request_payload(input_, stop=stop, **kwargs) + outgoing = payload.get("messages", []) + for message_dict, message in zip(outgoing, _input_to_messages(input_)): + if not isinstance(message, AIMessage): + continue + reasoning = message.additional_kwargs.get("reasoning_content") + if reasoning is not None: + message_dict["reasoning_content"] = reasoning + return payload + + def _create_chat_result(self, response, generation_info=None): + chat_result = super()._create_chat_result(response, generation_info) + response_dict = ( + response + if isinstance(response, dict) + else response.model_dump( + exclude={"choices": {"__all__": {"message": {"parsed"}}}} + ) + ) + for generation, choice in zip( + chat_result.generations, response_dict.get("choices", []) + ): + reasoning = choice.get("message", {}).get("reasoning_content") + if reasoning is not None: + generation.message.additional_kwargs["reasoning_content"] = reasoning + return chat_result + + def with_structured_output(self, schema, *, method=None, **kwargs): + if self.model_name == "deepseek-reasoner": + raise NotImplementedError( + "deepseek-reasoner does not support tool_choice; structured " + "output is unavailable. Agent factories fall back to " + "free-text generation automatically." + ) + return super().with_structured_output(schema, method=method, **kwargs) + # Kwargs forwarded from user config to ChatOpenAI _PASSTHROUGH_KWARGS = ( "timeout", "max_retries", "reasoning_effort", @@ -75,10 +144,12 @@ class OpenAIClient(BaseLLMClient): self.warn_if_unknown_model() llm_kwargs = {"model": self.model} - # Provider-specific base URL and auth + # Provider-specific base URL and auth. An explicit base_url on the + # client (e.g. a corporate proxy) takes precedence over the + # provider default so users can route through their own gateway. if self.provider in _PROVIDER_CONFIG: - base_url, api_key_env = _PROVIDER_CONFIG[self.provider] - llm_kwargs["base_url"] = base_url + default_base, api_key_env = _PROVIDER_CONFIG[self.provider] + llm_kwargs["base_url"] = self.base_url or default_base if api_key_env: api_key = os.environ.get(api_key_env) if api_key: @@ -98,7 +169,10 @@ class OpenAIClient(BaseLLMClient): if self.provider == "openai": llm_kwargs["use_responses_api"] = True - return NormalizedChatOpenAI(**llm_kwargs) + # DeepSeek's thinking-mode quirks live in their own subclass so the + # base NormalizedChatOpenAI stays free of provider-specific branches. + chat_cls = DeepSeekChatOpenAI if self.provider == "deepseek" else NormalizedChatOpenAI + return chat_cls(**llm_kwargs) def validate_model(self) -> bool: """Validate model for the provider."""