jupyter_firefly_extensions

JupyterLab extensions for rendering FITS and displaying tables, images, & charts with Firefly.

Overview

These extensions add the following features to JupyterLab:

• Open a FITS file from the file browser and see it in a tab.
• Start the full Firefly viewer in a tab (either through launcher or command palette).
• Use FireflyClient in notebook to start Firefly in a tab and send data (tables, images, charts) to it using the FireflyClient python API.

Prerequisites

• JupyterLab ^4.0.0 - where these extensions will run. Check past releases if you are using JupyterLab<4.

• firefly_client ^2.1.1 - can be installed with pip install firefly_client.

• Firefly server - you can run it locally via a Firefly Docker image obtained from https://hub.docker.com/r/ipac/firefly.

• astropy ^3.0.0 - (optional) used for convenience in example notebooks.

• nodejs ^18.0.0 - only needed if you're doing development install

Very Important: first setup the Firefly URL - 3 ways

• Add the following line to your ~/.jupyter/jupyter_notebook_config.py

  c.Firefly.url = 'http://localhost:8080/firefly'
  

Or

• Add the following line to your ~/.jupyter/jupyter_notebook_config.json under the root object.

  "Firefly": {
    "url": "http://localhost:8080/firefly"
  }
  

Or

• Use the environment variable in the shell where you start jupyter lab

  in bash:

  export FIREFLY_URL=http://localhost:8080/firefly
  

  in tcsh:

  setenv FIREFLY_URL http://localhost:8080/firefly
  

where the URL points to a Firefly server.

Installation

pip install jupyter_firefly_extensions

Open JupyterLab (with jupyter lab) to start using these extensions - see examples to learn how.

Development install

First:

If developing firefly_client, be sure to clone the firefly_client repository (https://github.com/Caltech-IPAC/firefly_client) and then do pip install -e . from inside its directory.

Make sure you have nodejs >=18.0.0 installed on your system or virtual environment. It's required for building TS/JS source. In conda environment, you can install it using conda install nodejs

Then:

git clone https://github.com/Caltech-IPAC/jupyter_firefly_extensions
cd jupyter_firefly_extensions

# Install package in development mode (changes in python source will reflect automatically)
pip install -e .

# Link your development version of the extension with JupyterLab
jupyter labextension develop . --overwrite

# Enable the server extension
jupyter server extension enable jupyter_firefly_extensions

# Rebuild extension TS/JS source each time you make a change
jlpm run build

Helpful commands

• jupyter server extension list - show a list of server extensions
• jupyter lab extension list - show a list of lab extensions
• jupyter lab - run jupyter lab
• jlpm <any yarn command> - JupyterLab-provided, locked version of the yarn

To remove extension:

pip uninstall jupyter_firefly_extensions

Examples

The examples directory has several example notebooks to demonstrate the extension features. When using the examples you should copy the directory and contents to another place or jupyter lab will and to keep rebuilding

• slate-demo-explicit.ipynb, slate-demo-explicit2.ipynb - demonstrates opening a Firefly tab and sending data to it with the FireflyClient python API
• Other notebooks demonstrate capabilites of widgets which are no longer supported, so they won't work.

Besides this, you can also use this extension to display fits images. In the file browser of jupyter lab, simply clicking on a .fits file will show the image in a new tab.

Troubleshooting

If you are using a local Firefly server and facing issues with rendering images, check the console for an error message about being unable to load 'firefly-thread.worker.js'. If that's the case, you can clean your existing Firefly build using gradle clean and then build and deploy it in the development environment (instead of the local one, i.e., the default) by using gradle -Penv=dev firefly:bAD. Then, reload the Jupyter Lab browser tab (and empty the cache). You shouldn't see that console error anymore and the images should render correctly.