Skip to content

KNIME UI Component

Example

This page explains how to run knime2py from a simple KNIME Component UI: you point it at a KNIME workflow, choose which artifacts to emit, and it generates Python workbooks plus graph files into a target folder.

If you prefer the CLI, see Quickstart in the docs. The UI component is just a thin front-end around the same k2p entrypoint.

What the component does

  • Accepts a workflow path (workflow.knime or a folder containing exactly one workflow.knime).
  • Writes outputs into an output directory you choose.
  • Lets you select which artifacts to emit:
  • Python script (.py)
  • Jupyter notebook (.ipynb)
  • Graphviz DOT (.dot)
  • Machine-readable graph JSON (.json)

The generated Python mirrors the KNIME workflow’s node order and port wiring and uses idiomatic pandas / scikit-learn in a shared context dictionary.

Prerequisites

  • KNIME Analytics Platform 5.x
  • A working knime2py executable PEX: a local k2p-*.pex (Python 3.11 required to run)

Verify the binary once on your machine:

python3 k2p-<YOUR_OS>.pex --help

Install the component

Open the KNIME Hub page for the component. Drag-and-drop it into your workflow (or “Open in KNIME”, then copy into your project).

Component on Hub: Component on Hub

You will see a single Component node in your canvas:

Component on canvas: Component on canvas

Configure

Double-click the component (or right-click → Configure). Fill in:

  • k2p_bin: Full path to the k2p executable you will use. Examples:
  • /Users/you/bin/k2p-macos-arm64.pex
  • C:\Users\you\bin\k2p-windows.pex
  • k2p_workflow: Path to the KNIME workflow you want to export.
  • Either /path/to/workflow.knime
  • Or a directory that directly contains workflow.knime
  • output_dir: Directory where artifacts will be written (will be created if missing).
  • Selection table: One column with any of .py, .ipynb, .dot, .json (one row per item). Leave blank to default to a safe set (script + JSON + DOT).

Configuration dialog: Configuration dialog

Notes:

  • Graph rendering is not automatic; knime2py writes a .dot file. Render with Graphviz, e.g.:

bash dot -Tpng out/COMPONENT__g01.dot -o out/COMPONENT__g01.png

Run

Click Execute. When it finishes:

  • Open the output folder to see generated files:

<base>__g01.json <base>__g01.dot <base>__g01_workbook.py <base>__g01_workbook.ipynb <base>__g02.json …

where <base> is the workflow directory name and __gNN is the component index.

Troubleshooting

  • Nothing is generated

  • Re-check k2p_bin points to a real executable; run it in a terminal to confirm.

  • Verify k2p_workflow is a file named workflow.knime or a directory containing exactly one workflow.knime.

  • PEX won’t run

  • Use Python 3.11 to execute the .pex file: python3 k2p-*.pex --help

  • DOT won’t preview

  • Install Graphviz and render manually with dot -Tpng ….

  • Permissions / paths

  • On macOS, if running from protected folders (Desktop/Documents), grant KNIME disk access or copy the workflow elsewhere.

How it works (internals)

The component is a thin wrapper around the CLI:

  • Reads your selections (ports/paths) from the dialog.
  • Constructs the k2p invocation.
  • Runs it and reports success/failure.
  • It does not alter your KNIME workflow; files are written only to output_dir.

If you need more control (e.g., Docker runs, custom flags), use the CLI directly as described in Quickstart.

See also