chaldene 0.2.0

Visual Programming in JupyterLab for Image Processing

Chaldene

Notebook-Embedded Visual Workflow Authoring for Scientific Image Processing

Key Features

• Low-barrier, drag-and-drop node-based authoring with an explicit, reusable workflow representation
• Image-specific inspection and comparison support, including branching, stepwise views, synchronized viewing, cursor linking, and difference overlays, with intermediate images captured as a visual trace
• Co-locates workflow structure, parameter settings, outputs, and narrative context within a single notebook artifact during authoring

Installation

Local Installation

The only external requirement is OpenJDK 11 (required for PyImageJ), for example with Conda:

conda install -c conda-forge "openjdk=11"

Then install Chaldene and launch JupyterLab:

pip install chaldene
jupyter lab

Use Without Local Installation

Chaldene targets practitioners with limited programming experience, so it is not intended to require manual installation and configuration by each end user. Instead, it can be provided through preconfigured environments, such as Binder, managed JupyterHub/JupyterLab servers, or lab-maintained software images, allowing practitioners to focus on workflow authoring rather than software administration.

Quick Start

1. Create a new notebook

   • Click "+" to create a new notebook
   • Add a Visual Code cell from the cell toolbar

2. Start building workflows

   • Drag and drop nodes to create your image processing workflows
   • Connect nodes to build workflows
   • Adjust parameters and inspect the outputs to refine the workflows

Examples

📂 Examples are available in the use_cases/ folder

Below are two representative workflows created by users, demonstrating Chaldene's capabilities for interactive image processing:

More Features

Publish to Zenodo

Archive your workflow files or results to Zenodo and mint a citable DOI, directly from JupyterLab.

1. Create a Zenodo personal access token with the deposit:write and deposit:actions scopes.
2. In the JupyterLab file browser, right-click the empty area and choose Publish to Zenodo.
3. In the dialog, select one or more files/folders to publish (folders are zipped automatically).
4. Enter your access token, then a title, author(s), and description.
5. All selected items are uploaded to a single Zenodo record and the minted DOI is shown when publishing completes.

Python API

ChaldeneClient lets you control VP cells from Python — useful for driving parameters with ipywidgets, running parameter sweeps, or integrating VP cells into larger notebook workflows.

from chaldene import ChaldeneClient

client = ChaldeneClient()

Once a VP cell is visible in the notebook, client.get_ready_cell_ids() returns its ID. From there you can update node inputs and re-run the cell:

cell_id = client.get_ready_cell_ids()[-1]
client.set_input(cell_id, node_id='1', handle_id='in1', value=[0.2, 0.8])
client.run(cell_id)

Combine with ipywidgets.interact for live parameter control:

import ipywidgets as widgets

@widgets.interact(threshold=widgets.FloatSlider(min=0.0, max=1.0, step=0.05))
def update(threshold):
    ids = client.get_ready_cell_ids()
    if ids:
        client.set_input(ids[-1], '1', 'in1', [threshold, 0.9])
        client.run(ids[-1])

See api_tutorials/ for full working examples.

Development

• 📖 Developer Guide - Setup and development instructions
• 🚀 Release Guide - Package Build and Release