AlphaPy - Modernized Edition

AlphaPy is a machine learning framework for both speculators and data scientists, modernized for Python 3.11+ with the latest package versions while preserving all original functionality.

Fork Information: This is a modernized fork of ScottfreeLLC/AlphaPy with the primary goal of updating the codebase for Python 3.13 support and modernizing the build system while maintaining 100% backward compatibility.

This modernized fork maintains 100% compatibility with the original API and behavior while updating the infrastructure for modern Python development:

• ✅ Python 3.11-3.13 support (tested and confirmed working)
• ✅ Modern build system using pyproject.toml (replaced setup.py)
• ✅ Updated dependencies to latest stable versions
• ✅ Replaced pyfolio with actively maintained pyfolio-reloaded
• ✅ Package manager support for uv, pip, poetry, etc.
• ✅ Optional TensorFlow - now in a separate dependency group
• ✅ Code compliance improvements with ruff linting and formatting
• ✅ Fixed deprecations for Python 3.10+ compatibility

• ✅ All original functionality intact
• ✅ API compatibility - no breaking changes
• ✅ Algorithm support - all original ML algorithms
• ✅ Pipeline behavior - identical processing logic
• ✅ Configuration format - same YAML structure

# Clone the repository
git clone https://github.com/yourusername/AlphaPy.git
cd AlphaPy

# Install with uv (recommended - excludes TensorFlow)
uv sync --all-groups --no-group tensorflow

# Or install specific groups
uv sync --group dev --group test --group ml-extras

# Add TensorFlow if needed (Linux/Windows only)
uv sync --group tensorflow

# Install in development mode
pip install -e .

# Install with all ML libraries
pip install -e ".[ml-extras]"

# Install with TensorFlow support
pip install -e ".[tensorflow]"

XGBoost requires OpenMP to be installed on macOS. Without it, you'll get an error when importing XGBoost:

Library not loaded: /usr/local/opt/libomp/lib/libomp.dylib

Solution:

brew install libomp

If you still get errors after installing, you may need to:

# Reinstall XGBoost
uv pip uninstall xgboost
uv sync --group ml-extras

# Or set the library path
export DYLD_LIBRARY_PATH=/usr/local/opt/libomp/lib:$DYLD_LIBRARY_PATH

Note: This is a system dependency that cannot be installed via Python packages. It must be installed separately using Homebrew.

The modernized version organizes dependencies into logical groups:

AlphaPy provides a comprehensive ML framework for:

• Machine Learning Models using scikit-learn, XGBoost, LightGBM, CatBoost, and optionally Keras/TensorFlow
• Ensemble Methods with blending and stacking
• Market Analysis with MarketFlow for financial markets
• Sports Prediction with SportFlow for sporting events
• Trading Systems and portfolio analysis
• Feature Engineering with extensive transformations
• AutoML capabilities with hyperparameter optimization

# Traditional usage remains unchanged
from alphapy.model import Model
from alphapy.data import get_data

# Create and run model exactly as before
model = Model(specs)

# Analyze markets
mflow

# Predict sporting events
sflow

If you're migrating from the original AlphaPy:

1. No code changes required - Your existing code will work as-is
2. Configuration files remain the same (YAML format unchanged)
3. Import statements are identical
4. Model outputs are compatible

*Python 3.13 support depends on all dependencies being compatible

# Run tests
uv run pytest

# Run with coverage
uv run pytest --cov=alphapy

# Run linting
uv run ruff check alphapy/

# Run formatting
uv run ruff format alphapy/

# Type checking
uv run mypy alphapy/

Original documentation is available at:

• http://alphapy.readthedocs.io/en/latest/

Build documentation locally:

cd docs/
make html

This modernized version maintains:

• ✅ Identical API - All functions, classes, and methods unchanged
• ✅ Same behavior - Algorithms produce identical results
• ✅ Configuration compatibility - Existing YAML files work
• ✅ Output format - Same file formats and structures

# Install all development dependencies
uv sync --all-groups

# Run code quality checks
uv run ruff check alphapy/
uv run ruff format alphapy/ --check
uv run mypy alphapy/

# Run tests
uv run pytest

# Main pipeline
alphapy [--train | --predict]

# Market analysis
mflow

# Sports prediction  
sflow

• Python 3.11+
• NumPy < 2.0 (for compatibility)
• scikit-learn >= 1.3.0
• pandas >= 2.1.0

• XGBoost >= 2.0.0 (requires brew install libomp on macOS)
• LightGBM >= 4.1.0
• CatBoost >= 1.2.0
• TensorFlow >= 2.15.0 (optional, heavy dependency)

We aim to support Python 3.13 as soon as all critical dependencies are compatible. Current blockers:

• NumPy (currently supports up to 3.12)
• TensorFlow (currently supports up to 3.12)
• Some scientific computing libraries

Track Python 3.13 compatibility in issue #XXX.

1. XGBoost on macOS: Requires OpenMP (brew install libomp)
2. TensorFlow on Apple Silicon: May require special installation
3. Deprecation warnings: From pandas-datareader (upstream issue)

This project maintains the original Apache License, Version 2.0. See LICENSE file for details.

Original AlphaPy created by ScottFree Analytics LLC (Mark Conway & Robert D. Scott II)

This modernization preserves their excellent work while updating it for contemporary Python development.

• Original AlphaPy Repository
• Documentation
• PyPI Package

Note: This is a modernization effort that maintains full backward compatibility. The core algorithms, logic, and behavior of AlphaPy remain unchanged. Only the build system, dependencies, and Python compatibility have been updated.