Ajasra b71faa9758 refactor for modularity

2025-05-23 20:37:14 +08:00

10 KiB

Raw Blame History

Cycles - Advanced Trading Strategy Backtesting Framework

A Python framework for backtesting cryptocurrency trading strategies with multi-timeframe analysis, strategy combination, and advanced signal processing. Recently refactored for improved modularity, reusability, and maintainability.

✨ Key Features

🏗️ Modular Architecture: Clean separation of concerns with reusable components
🔧 Multi-Strategy System: Combine multiple trading strategies with configurable weights and rules
⏱️ Multi-Timeframe Analysis: Strategies operate on different timeframes (1min, 5min, 15min, 1h, etc.)
📊 Advanced Strategies:
- Default Strategy: Meta-trend analysis using multiple Supertrend indicators
- BBRS Strategy: Bollinger Bands + RSI with market regime detection
🎯 Flexible Signal Combination: Weighted consensus, majority voting, any/all combinations
⚡ Precise Execution: 1-minute precision for accurate risk management
📈 Comprehensive Analysis: Detailed performance metrics and trade analysis
📱 Enhanced CLI: Improved command-line interface with better error handling
🔍 Debug Mode: Sequential execution with interactive plotting

🚀 Quick Start

Prerequisites

Python 3.8+
uv package manager (recommended) or pip

Installation

# Clone the repository
git clone <repository-url>
cd Cycles

# Install dependencies with uv (recommended)
uv sync

# Or install with pip
pip install -r requirements.txt

Running Backtests

The new CLI provides a clean interface with automatic config discovery:

# Use default configuration
python main.py

# Use specific config (searches configs/ directory automatically)
python main.py config_bbrs.json

# Debug mode with interactive plotting
python main.py --debug config_combined.json

# Full path also supported
python main.py configs/config_default_5min.json

# Get help
python main.py --help

Available Configurations

config_default.json: Default meta-trend strategy (15min)
config_default_5min.json: Default strategy on 5-minute timeframe
config_bbrs.json: BBRS strategy with market regime detection
config_bbrs_multi_timeframe.json: BBRS with multiple timeframes
config_combined.json: Multi-strategy combination with weighted consensus

🏛️ Architecture Overview

The framework follows a clean, modular architecture for maximum reusability:

cycles/
├── application.py              # 🎯 Main application orchestration
├── utils/
│   ├── config_manager.py       # ⚙️ Centralized configuration management
│   ├── results_processor.py    # 📊 Results processing & metrics calculation
│   ├── backtest_runner.py      # 🚀 Backtest execution logic
│   ├── storage.py              # 💾 Data storage utilities
│   └── system.py               # 🖥️ System utilities
├── strategies/                 # 📈 Strategy implementations
│   ├── base.py                 # 🏗️ Base strategy framework
│   ├── default_strategy.py     # 📊 Meta-trend strategy
│   ├── bbrs_strategy.py        # 🎯 BBRS strategy
│   └── manager.py              # 🎛️ Strategy orchestration
├── backtest.py                 # ⚡ Core backtesting engine
└── charts.py                   # 📈 Visualization components

📱 Enhanced CLI Interface

python main.py [-h] [--debug] [config]

# Examples:
python main.py                              # Use default config
python main.py config_bbrs.json             # Use specific config
python main.py --debug config_combined.json # Debug mode with plotting

# Available configs:
# - config_default.json: Default meta-trend strategy
# - config_bbrs.json: BBRS strategy
# - config_combined.json: Multi-strategy combination

⚙️ Configuration

Strategies are configured using JSON files in the configs/ directory:

{
    "start_date": "2024-01-01",
    "stop_date": "2024-01-31",
    "initial_usd": 10000,
    "timeframes": ["15min"],
    "strategies": [
        {
            "name": "default",
            "weight": 1.0,
            "params": {
                "timeframe": "15min",
                "stop_loss_pct": 0.03
            }
        }
    ],
    "combination_rules": {
        "entry": "weighted_consensus",
        "exit": "any",
        "min_confidence": 0.6
    }
}

Strategy Configuration

Single Strategy

{
    "strategies": [
        {
            "name": "default",
            "weight": 1.0,
            "params": {
                "timeframe": "15min",
                "stop_loss_pct": 0.03
            }
        }
    ]
}

Multi-Strategy Combination

{
    "strategies": [
        {
            "name": "default",
            "weight": 0.6,
            "params": {"timeframe": "15min"}
        },
        {
            "name": "bbrs",
            "weight": 0.4,
            "params": {"bb_width": 0.05}
        }
    ],
    "combination_rules": {
        "entry": "weighted_consensus",
        "exit": "any",
        "min_confidence": 0.6
    }
}

Combination Rules

Entry Methods: any, all, majority, weighted_consensus
Exit Methods: any, all, priority (prioritizes stop-loss signals)
Min Confidence: Threshold for signal acceptance (0.0 - 1.0)

🛠️ Programmatic Usage

The new modular architecture enables programmatic usage:

from cycles.application import BacktestApplication

# Simple usage
app = BacktestApplication("configs/config_bbrs.json")
app.run(debug=False)

# Custom workflow
from cycles.utils import ConfigManager, BacktestRunner

config_manager = ConfigManager("my_config.json")
runner = BacktestRunner()

for timeframe in config_manager.timeframes:
    config = config_manager.get_timeframe_task_config(timeframe)
    results = runner.run_single_timeframe(data, timeframe, config)
    # Process results...

📈 Available Strategies

1. Default Strategy (Meta-Trend Analysis)

Type: Meta-trend analysis using multiple Supertrend indicators
Timeframes: Configurable (5min, 15min, 1h, etc.)
Features: Triple Supertrend confirmation, precise stop-loss

{
    "name": "default",
    "params": {
        "timeframe": "15min",
        "stop_loss_pct": 0.03
    }
}

2. BBRS Strategy (Bollinger Bands + RSI)

Type: Bollinger Bands + RSI with market regime detection
Features: Adaptive parameters, volume confirmation, multi-timeframe analysis
Market Regimes: Trending vs sideways market detection

{
    "name": "bbrs",
    "params": {
        "bb_width": 0.05,
        "strategy_name": "MarketRegimeStrategy",
        "stop_loss_pct": 0.05
    }
}

📊 Output & Results

Generated Files

YYYY_MM_DD_HH_MM_backtest.csv: Performance summary per timeframe
YYYY_MM_DD_HH_MM_trades.csv: Individual trade records
backtest.log: Detailed execution logs

Debug Mode

Interactive Charts: Real-time plotting of strategy signals and trades
Sequential Execution: Step-by-step analysis
Enhanced Logging: Detailed strategy information

Performance Metrics

Trade Statistics: Win rate, profit ratio, drawdown analysis
Portfolio Metrics: Initial/final USD, total returns, fees
Risk Metrics: Maximum drawdown, stop-loss frequency

📚 Documentation

Comprehensive documentation available in the docs/ directory:

🏗️ Refactoring Summary - Overview of new architecture improvements
🎛️ Strategy Manager - Multi-strategy orchestration
📈 Strategies - Individual strategy implementations
⏱️ Timeframe System - Multi-timeframe management
📊 Analysis - Technical analysis components
💾 Storage Utils - Data management
🖥️ System Utils - System utilities

🎯 Migration & Compatibility

The refactored framework is 100% backward compatible:

✅ Same command-line interface
✅ Same configuration file format
✅ Same output file format
✅ Same functionality

Zero breaking changes while providing a much cleaner, more maintainable architecture.

🔧 Advanced Usage

Custom Strategy Development

from cycles.strategies.base import StrategyBase, StrategySignal

class MyStrategy(StrategyBase):
    def get_timeframes(self):
        return ["1h"]
    
    def initialize(self, backtester):
        self._resample_data(backtester.original_df)
        # Setup indicators...
        self.initialized = True
    
    def get_entry_signal(self, backtester, df_index):
        # Your entry logic...
        return StrategySignal("ENTRY", confidence=0.8)

Custom Results Processing

from cycles.utils import BacktestMetrics, ResultsProcessor

class CustomProcessor(ResultsProcessor):
    def process_backtest_results(self, results, timeframe, config, summary):
        # Custom processing logic...
        return custom_summary, custom_trades

🚀 Future Enhancements

The new architecture enables easy addition of:

🌐 Web API Interface: REST API for remote backtesting
⚡ Real-time Trading: Live trading integration
📊 Advanced Analytics: Enhanced performance analysis
🔌 Strategy Marketplace: Plugin system for custom strategies
📈 Multiple Asset Classes: Beyond cryptocurrency support

🤝 Contributing

We welcome contributions! The new modular architecture makes it easier to:

Add new strategies
Enhance analysis capabilities
Improve visualization
Extend data sources

📄 License

[Add your license information here]

Recent Updates: The framework has been significantly refactored for improved modularity and reusability while maintaining full backward compatibility. See Refactoring Summary for details.

10 KiB Raw Blame History