Skip to content

knime2py — KNIME → Python Workbook

knime2py is a KNIME→Python exporter and code generator. It parses a KNIME workflow, reconstructs its nodes and connections, and emits runnable Python “workbooks” (Jupyter notebook and/or script) by translating supported KNIME nodes into idiomatic pandas / scikit-learn code via a pluggable node registry. In addition to executable code, it writes a machine-readable graph (JSON) and a Graphviz DOT file, preserving port wiring and execution order so the generated Python mirrors the original workflow.

GitHub repo: vitalii-kaplan/knime2py

Features

  • Single-workflow focus – pass either a workflow.knime file or a directory that directly contains workflow.knime.
    Subdirectories are not traversed.
  • Components – nested component workflows are supported if you point directly at the component’s folder (it also has its own workflow.knime).
  • Isolated graphs detection – splits the workflow into unconnected graphs; each becomes its own output set with an ID suffix like __g01, __g02, …
  • Deterministic ordering – a stable depth-first traversal ensures nodes run after all predecessors; cyclic/disconnected regions are handled with a consistent fallback order.
  • Pluggable node registry – KNIME nodes are mapped to emitters in knime2py.nodes.*; unsupported nodes get a clear stub with initialized parameters from settings.xml.

Requirements

  • Python 3.9+ (project is developed/tested on 3.11).
  • Runtime libs used by generated code: pandas, scikit-learn (bundled in Docker/PEX; for source installs see below).
  • Optional: Graphviz CLI (dot) if you want to render .dot files; the tool always writes .dot.

Contributing

  • Emitters live under src/knime2py/nodes/.

  • Each emitter:

  • Parses settings.xml.

  • Generates imports and a code body that reads/writes the shared context.
  • Declares a handle() hook for registry lookup.
  • Tests live under tests/; helpers under tests/support/.

License

MIT

Acknowledgements

KNIME® is a trademark of KNIME AG. This project is an independent community effort and is not affiliated with KNIME AG.