jppype 0.2.0a4

JupyterLab extension adding data visualisation widgets.

Jupyter ProtoPype: jppype

Jupyter extensions adding ipywidgets specialized for data-view. It is designed as a frontend for ProtoPype (to be released) but can be used as a standalone.

The widgets work both with Jupyter Notebook and Jupyter Lab.

Installation

To install use pip:

pip install jppype==0.2.0a3

Cv2 is required by jppype but is not listed in the package dependencies (to prevent version conflicts). You can install it with:

pip install opencv-python-headless

Usage

Import jppype in your notebook and use the widgets:

import jppype as jpp

2D image viewer

viewer = jpp.View2D()

# Supported image formats: numpy array, torch tensor
image = np.random.rand(100, 100) 
viewer.add_image(image)
viewer

Features coming...

• Widgets

  • 2D image viewer
    • Color images layer
    • Label layer
    • Graph layer
    • Vector Field layer
  • Table viewer

• Python interactivity

  • Add/Update/Remove layers
  • Callback on click
  • Async request of user selection (point, area, node, edge...)

Development Installation

Pull the latest version of the repository

git clone https://github.com/gabriel-lepetitaimon/jppype.git
cd jppype

For a development installation (requires Node.js and Yarn version 1):

conda create -n jppype python
conda install -c conda-forge yarn nodejs

Download the npm dependencies:

cd ts-src
npm install
cd ..

Upgrade jupyter-lab to version 4.0.0 if necessary:

pip install -U jupyterlab

Install the python package and the extensions in development mode (this stage may take a while as it will build the jupyterlab extension):

pip install -e .[dev,build]
jupyter nbextension install --py --symlink --overwrite --sys-prefix jppype
jupyter nbextension enable --py --sys-prefix jppype

If you which to develop for JupyterLab, also run the command:

jupyter labextension develop --overwrite jppype

Then you need to rebuild the JS when you make a code change:

cd ts-src
jlpm clean
jlpm build

or automatically rebuild when a file changes:

cd ts-src
npm watch

You then need to refresh the JupyterLab page when your javascript changes.