docs: Add v2.0.0 release changelog

2026-07-27 23:47:46 +00:00 · 2026-04-10 12:05:35 +02:00
parent 1cdea7db26
commit e22c209533
1 changed files with 107 additions and 0 deletions
@@ -0,0 +1,107 @@
+# Predix v2.0.0 Release Notes
+
+**Release Date:** 2026-04-10
+
+**Tag:** v2.0.0-2026.04.10
+
+---
+
+## 🎉 Overview
+
+Major update adding AI-powered strategy generation, realistic backtesting, and comprehensive CLI tooling. Predix now autonomously generates, evaluates, and optimizes trading strategies using local LLMs.
+
+---
+
+## ✨ Added
+
+### LLM-Powered Strategy Generation
+- **StrategyOrchestrator**: Generate trading strategies by combining factors with LLM
+- **Local llama.cpp Support**: Run strategy generation locally (Qwen3.5-35B)
+- **OpenRouter Support**: Optional cloud model fallback
+- **Improved Prompts (v3)**: IC-sign-aware factor combination instructions
+- **Diverse Factor Selection**: Automatic selection by type (momentum, divergence, volatility, session)
+
+### Realistic Backtesting
+- **OHLCV-Based Returns**: Real price returns instead of factor proxies
+- **Spread Costs**: 1.5 bps per trade deducted from returns
+- **Forward-Fill Support**: Daily factors → 1-min frequency
+- **Proper Annualization**: sqrt(252*1440) for 1-min data
+
+### CLI Commands (6 New)
+- `rdagent start_llama` - Start llama.cpp server
+- `rdagent start_loop` - Start strategy generator loop with auto-restart
+- `rdagent generate_strategies` - Generate strategies from factors
+- `rdagent optimize_portfolio` - Portfolio optimization
+- `rdagent strategies_report` - Generate performance reports
+- `rdagent eval_all` - Evaluate factors with full data
+- `rdagent batch_backtest` - Batch backtest existing factors
+- `rdagent report` - Generate PDF performance reports
+- `rdagent rebacktest` - Re-backtest existing strategies
+
+### Code Quality
+- **282+ Integration Tests**: All features tested
+- **Security Hardening**: All Dependabot/CodeQL alerts resolved
+- **Pre-commit Hooks**: Automated tests + security scanning
+
+---
+
+## 📊 Performance
+
+- **2,800+ Factors**: Generated and evaluated
+- **14 Qualified Strategies**: After strict re-evaluation
+- **Top Strategy**: Sharpe 6.04, 21.88% annualized return, -1.57% max drawdown
+
+---
+
+## 🔧 Changed
+
+- Utility scripts organized in `scripts/` directory
+- Generated data moved to `results/`
+- Config files moved to `constraints/`
+- Root directory cleaned (872 → 29 files)
+
+---
+
+## 🐛 Fixed
+
+- JSON strategy files no longer committed to root
+- LICENSE badge link corrected (main → master)
+- Security vulnerabilities resolved (bandit, path traversal)
+
+---
+
+## 📦 Installation
+
+```bash
+git clone https://github.com/TPTBusiness/Predix
+cd Predix
+pip install -e .
+```
+
+## 🚀 Quick Start
+
+```bash
+# Start LLM server
+rdagent start_llama
+
+# Run trading loop
+rdagent fin_quant --auto-strategies
+
+# Generate strategies manually
+rdagent generate_strategies --count 5 --optuna
+```
+
+---
+
+## 🔒 Security
+
+- All known vulnerabilities resolved
+- Bandit security scanning integrated
+- Pre-commit hooks for automated checks
+- Path traversal prevention hardened
+
+---
+
+## 📄 License
+
+MIT License - see [LICENSE](../LICENSE) for details.