FlowerPower

FlowerPower is a Python framework that builds data pipelines on top of Hamilton. You write ordinary Python functions; FlowerPower handles configuration, execution, CLI, UI, and I/O so your project stays organized and portable.

Choose your path

• New here? Install FlowerPower and follow the Quickstart tutorial to run your first pipeline.
• Looking for task guides? See How-to guides for composing modules, async execution, adapters, and advanced workflows.
• Want the big picture? Read the Architecture concepts overview.
• Need API details? Browse the CLI reference and API reference.
• Prefer examples? Explore the example projects.
• Interested in contributing? Check out the Contributing guide.

What is FlowerPower?

• Modular DAGs via Hamilton: define transformations as plain Python functions and let Hamilton assemble them into a directed acyclic graph.
• Config-driven YAML: separate pipeline logic from parameters, adapters, and runtime settings using layered YAML and environment variables.
• Unified interfaces: interact with the same project through the Python API, the CLI, or a web-based UI.
• Filesystem abstraction: run locally or against remote storage (S3, GCS, etc.) through fsspec-compatible filesystems.
• Extensible I/O: connect to CSV, JSON, Parquet, Delta, DuckDB, SQL databases, and more via optional I/O plugins.

FlowerPower is an orchestration layer

It does not replace Hamilton. You still write pipeline logic with Hamilton's function-based syntax; FlowerPower adds project structure, configuration, and lifecycle tooling.

Source code and issue tracker are on GitHub.