bigmlcom/python

Provides local prediction capabilities using models created on BigML's platform, supporting decision trees, topic models, and image processing without requiring API calls. The library uses requests-based HTTP transport with numpy/scipy for numerical computations, and offers optional extras for specialized model types (topics, images) that have platform-specific dependencies. Integrates with BigML.io's REST API for resource management (sources, datasets, models, predictions) while maintaining compatibility across Linux, Windows, and Google App Engine environments.

292 stars. No commits in the last 6 months.