Commit Graph

17 Commits

Author SHA1 Message Date
swj.premkumar e7784f2b99 - Guarantees clean string outputs for downstream consumers (CLI, WebUI).
- Prevents crashes when LLMs return complex JSON/List structures.
2026-01-15 02:01:54 -06:00
swj.premkumar 05ce55125f ### Added
- **Dynamic Parameter Tuning (The Learning Loop)**: Implemented full self-reflection cycle. The Reflector agent now parses its own advice into JSON (`rsi_period`, `stop_loss_pct`), persists it to `data_cache/runtime_config.json`, and the Market Analyst loads it to tune the Regime Detector in real-time.
- **Audit Archival**: Every tuning event is now archived to `results/{TICKER}/{DATE}/runtime_config.json` for historical auditing, ensuring we can reproduce why parameters changed on any given day.
- **Atomic Persistence**: Implemented `agent_utils.write_json_atomic` to prevent race conditions during config saves.
- **Centralized Config**: Moved hardcoded paths to `default_config.py` (DRY principle).

### Fixed
- **Reflector Logic Gap**: The Reflector was previously "shouting into the void"—making suggestions but having no mechanism to apply them. This circuit is now closed.
2026-01-13 06:40:07 -06:00
swj.premkumar 1f279a9df2 - **Insider Veto Protocol (Rule B)**: Hard-coded safety gate in `trading_graph.py` that blocks ALL buy signals if Net Insider Selling exceeds $50M while the stock is in a technical downtrend (Price < 50 SMA). This prevents "Falling Knife" catches.
- **Relative Strength Determinism**: Upgraded `market_analyst.py` to calculate a mathematical `risk_multiplier` (0.0x - 1.5x) based on the Asset Regime vs. SPY Regime correlation, removing LLM "confidence" hallucinations from position sizing.
- **Portfolio Awareness (Rule 72)**: Implemented State Persistence (`portfolio`, `cash_balance`) and a hard-coded Stop Loss check in `trading_graph.py`. If a position's unrealized PnL drops below -10%, the system forces a "LIQUIDATE" order, bypassing all AI debate.
- **Self-Tuning Architecture**: Updated `reflection.py` to output a structured JSON block (`UPDATE_PARAMETERS`) instead of prose advice, enabling future automated parameter optimization.
2026-01-13 05:27:24 -06:00
swj.premkumar e88a01d0ea Fixed
Override Logic Mismatches: Fixed critical Enum-to-String type mismatch in
apply_trend_override
 that was silencing the "Safety Valve" logic.
Data Pipeline Failures: Injected robust error handling and type checking in
market_analyst.py
 to identify why RegimeDetector receives invalid data (causing "UNKNOWN" regimes).
Gemini 404 Errors: Removed invalid/deprecated model names causing 404s
2026-01-11 20:13:01 -06:00
swj.premkumar a6e4c9b770 feat: implement trend override, harden regime detection, and organize tests
**Core Logic (Safety Valve & Regime Detection):**

  *   Added "Momentum Override": `Overall Return > 30%` now forces `TRENDING_UP` (Bull) regime to capture volatile winners.
    *   Prioritized Trend Strength (ADX) over Volatility for single stocks.
    *   Fixed `Hurst Exponent` calculation to handle non-positive inputs safely.
*   **Data Reliability ([market_analyst.py]
2026-01-11 11:18:46 -06:00
swj.premkumar d0f229a444 Dataflow: Robustified
route_to_vendor
 to treat empty/whitespace-only results as failures, ensuring the fallback chain continues to the next vendor.
Analyst Nodes: Removed destructive logic that was overwriting real company names with ticker symbols in the anonymizer state.
Fixed
News Vendors: Resolved a TypeError in Google news results and removed the non-functional OpenAI news implementations.
Report Generator: Fixed a NameError related to sys and restored path-parsing logic in
2026-01-11 08:32:08 -06:00
swj.premkumar 54a3395b37 Phase 2 changes 2026-01-10 05:17:58 -06:00
swj.premkumar d2ebd6d587 ickerAnonymizer` into all analyst agents (`Market`, `News`, `Fundamentals`, `Social`) and data tools. The LLM now only sees "ASSET_XXX" in prompts, preventing data contamination. 2026-01-09 21:40:45 -06:00
swj.premkumar a0ab1a9b3e The **TradingAgents** system is a risk-managed, LLM-driven trading engine designed to execute trades based on validated truth, not hallucinations. It connects hierarchical LLM agents with deterministic safety gates to ensure that every trade is architecturally sound, factually correct, and risk-compliant. 2026-01-09 19:28:49 -06:00
luohy15 86bc0e793f minor fix 2025-09-27 00:04:59 +08:00
luohy15 8b04ec307f minor fix 2025-09-26 23:25:33 +08:00
luohy15 0ab323c2c6 Add Alpha Vantage API integration as primary data provider
- Replace FinnHub with Alpha Vantage API in README documentation
- Implement comprehensive Alpha Vantage modules:
  - Stock data (daily OHLCV with date filtering)
  - Technical indicators (SMA, EMA, MACD, RSI, Bollinger Bands, ATR)
  - Fundamental data (overview, balance sheet, cashflow, income statement)
  - News and sentiment data with insider transactions
- Update news analyst tools to use ticker-based news search
- Integrate Alpha Vantage vendor methods into interface routing
- Maintain backward compatibility with existing vendor system

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-09-26 22:57:50 +08:00
luohy15 a6734d71bc WIP 2025-09-26 16:17:50 +08:00
Yijia Xiao b575bd0941
Merge pull request #52 from TauricResearch/dev
Merge dev into main. Add support for Anthropic and OpenRouter.
2025-06-23 20:38:14 -04:00
Edward Sun 52284ce13c fixed anthropic support. Anthropic has different format of response when it has tool calls. Explicit handling added 2025-06-21 12:51:34 -07:00
Atharv Sabde 2af7ef3d79
fundamentals_analyst.py(spelling mistake.markdown) 2025-06-19 21:48:16 +05:30
Yijia-Xiao cc97cb6d5d chore(release): v0.1.0 – initial public release of TradingAgents 2025-06-05 04:27:57 -07:00