stacking-sats-pipeline 0.2.0

Hypertrial's Stacking Sats Library - Optimized Bitcoin DCA

Stacking Sats Pipeline

A Bitcoin DCA strategy backtesting framework for testing strategies against historical price data.

Requirements

• Python 3.11 or 3.12
• pip

Installation

pip install stacking-sats-pipeline

Quick Start

Library Interface

from stacking_sats_pipeline import backtest, strategy

# Simple function approach
def my_strategy(df):
    """Calculate weights based on price data"""
    # Your strategy logic here
    return weights

results = backtest(my_strategy)
results.summary()
results.plot()

# Or use decorator approach
@strategy(name="My Strategy", auto_backtest=True)
def my_strategy(df):
    return weights

Note: Data is now loaded directly into memory from CoinMetrics (no CSV files needed). For legacy file-based loading, use load_data(use_memory=False).

Data Extraction

Extract all data sources to local files for offline analysis:

CLI Usage

# Extract all data to CSV format
stacking-sats --extract-data csv

# Extract all data to Parquet format (smaller files, better compression)
stacking-sats --extract-data parquet

# Extract to specific directory
stacking-sats --extract-data csv --output-dir data/
stacking-sats --extract-data parquet -o exports/

Python API

from stacking_sats_pipeline import extract_all_data

# Extract all data to CSV in current directory
extract_all_data("csv")

# Extract all data to Parquet in specific directory
extract_all_data("parquet", "data/exports/")

What gets extracted:

• 📈 Bitcoin Price Data (CoinMetrics) → btc_coinmetrics.csv/parquet
• 😨 Fear & Greed Index (Alternative.me) → fear_greed.csv/parquet
• 💵 U.S. Dollar Index (FRED) → dxy_fred.csv/parquet*

*Requires FRED_API_KEY environment variable. Get a free key at FRED API

File Format Benefits:

• CSV: Human-readable, universally compatible
• Parquet: ~50% smaller files, faster loading, preserves data types

Interactive Tutorial

pip install marimo
marimo edit tutorials/examples.py

Command Line

stacking-sats --strategy path/to/your_strategy.py

Usage Examples

Basic Strategy

def simple_ma_strategy(df):
    """Buy more when price is below 200-day moving average"""
    df = df.copy()
    past_price = df["PriceUSD"].shift(1)
    df["ma200"] = past_price.rolling(window=200, min_periods=1).mean()
    
    base_weight = 1.0 / len(df)
    weights = pd.Series(base_weight, index=df.index)
    
    # Buy 50% more when below MA
    below_ma = df["PriceUSD"] < df["ma200"]
    weights[below_ma] *= 1.5
    
    return weights / weights.sum()

results = backtest(simple_ma_strategy)

Quick Comparison

strategy1_perf = quick_backtest(strategy1)
strategy2_perf = quick_backtest(strategy2)

Custom Parameters

results = backtest(
    my_strategy,
    start_date="2021-01-01",
    end_date="2023-12-31",
    cycle_years=2
)

Strategy Requirements

Your strategy function must:

def your_strategy(df: pd.DataFrame) -> pd.Series:
    """
    Args:
        df: DataFrame with 'PriceUSD' column and datetime index
        
    Returns:
        pd.Series with weights that sum to 1.0 per cycle
    """
    # Your logic here
    return weights

Validation Rules:

• Weights sum to 1.0 within each cycle
• All weights positive (≥ 1e-5)
• No forward-looking data
• Return pandas Series indexed by date

Development

For development and testing:

Requirements: Python 3.11 or 3.12

# Clone the repository
git clone https://github.com/hypertrial/stacking_sats_pipeline.git
cd stacking_sats_pipeline

# Set up development environment (installs dependencies + pre-commit hooks)
make setup-dev

# OR manually:
pip install -e ".[dev]"
pre-commit install

# Run tests
make test
# OR: pytest

# Code quality (MANDATORY - CI will fail if not clean)
make lint          # Fix linting issues
make format        # Format code
make check         # Check without fixing (CI-style)

# Run specific test categories
pytest -m "not integration"  # Skip integration tests
pytest -m integration        # Run only integration tests

Code Quality Standards

⚠️ MANDATORY: All code must pass ruff linting and formatting checks.

• Linting/Formatting: We use ruff for both linting and code formatting
• Pre-commit hooks: Automatically run on every commit to catch issues early
• CI enforcement: Pull requests will fail if code doesn't meet standards

Quick commands:

make help          # Show all available commands
make lint          # Fix ALL issues (autopep8 + ruff + format)
make autopep8      # Fix line length issues specifically  
make format        # Format code with ruff only
make format-all    # Comprehensive formatting (autopep8 + ruff)
make check         # Check code quality (what CI runs)

For detailed testing documentation, see TESTS.md.

Contributing Data Sources

The data loading system is designed to be modular and extensible. To add new data sources (exchanges, APIs, etc.), see the Data Loader Contribution Guide which provides step-by-step instructions for implementing new data loaders.

Command Line Options

# Basic usage
stacking-sats --strategy your_strategy.py

# Skip plots
stacking-sats --strategy your_strategy.py --no-plot

# Run simulation
stacking-sats --strategy your_strategy.py --simulate --budget 1000000

# Extract data
stacking-sats --extract-data csv --output-dir data/
stacking-sats --extract-data parquet -o exports/

# Show help
stacking-sats --help

Project Structure

├── stacking_sats_pipeline/
│   ├── main.py          # Pipeline orchestrator
│   ├── backtest/        # Validation & simulation
│   ├── data/            # Modular data loading system
│   │   ├── coinmetrics_loader.py  # CoinMetrics data source
│   │   ├── data_loader.py         # Multi-source data loader
│   │   └── CONTRIBUTE.md          # Guide for adding data sources
│   ├── plot/            # Visualization
│   ├── strategy/        # Strategy templates
│   └── weights/         # Historical allocation calculator
├── tutorials/examples.py # Interactive notebook
└── tests/               # Comprehensive test suite

Output

The pipeline provides:

• Validation Report: Strategy compliance
• Performance Metrics: SPD (Sats Per Dollar) statistics
• Comparative Analysis: vs Uniform DCA and Static DCA
• Visualizations: Weight distribution plots

Example Output

============================================================
COMPREHENSIVE STRATEGY VALIDATION
============================================================
✅ ALL VALIDATION CHECKS PASSED

Your Strategy Performance:
Dynamic SPD: mean=4510.21, median=2804.03
Dynamic SPD Percentile: mean=39.35%, median=43.80%

Mean Excess vs Uniform DCA: -0.40%
Mean Excess vs Static DCA: 9.35%