Skip to content

A development environment for Ballet collaborations on top of Jupyter Lab

License

Notifications You must be signed in to change notification settings

ballet/ballet-assemble

Repository files navigation

Github Actions Status PyPI Shield npm version

Assemblé

A development environment for Ballet collaborations on top of Jupyter Lab

Using Assemblé, you can develop patches for Ballet projects within Jupyter Lab and then easily submit them as GitHub Pull Requests (PRs) in one click without leaving your notebook.

Assemblé (pronounced "assam blay") is a ballet move that involves lifting off the floor on one leg and landing on two.

Assemblé is composed of (1) a Python package named ballet_assemble for the server extension (2) a NPM package also named ballet-assemble for the frontend extension and (3) tight integration with Binder for each Ballet project.

Requirements

  • JupyterLab >= 2.0
  • Python>=3.6 (64-bit Version)

Install

Installation can be done completely using pip, which installs both the server and the frontend extensions. The frontend extension only can be installed using jupyter labextension install but will not function properly without the corresponding server extension.

pip install ballet_assemble
jupyter lab build

Note: You will need NodeJS to install the extension; the installation process will complain if it is not found.

Authenticate with GitHub

The extension provides an in-Lab experience for authenticating with GitHub. When you open a notebook, you should see the GitHub icon to the right on the Notebook toolbar. The icon should be grey at first, indicating you are not authenticated. Click the icon to open a login window, in which you can enter your GitHub username and password. These will be exchanged by the extension for an OAuth token and will be used to propose changes to the upstream Ballet project on your behalf (if you attempt to submit features).

Authenticate with GitHub

Alternately, you can provide a personal access token directly using the configuration approaches below.

Configure

The extension ties into the same configuration system as Jupyter [Lab] itself. You can configure the extension with command line arguments or via the config file, just like you configure Jupyter Notebook or Jupyter Lab.

All configuration options

The following configuration options are available:

$ python -c 'from ballet_assemble.app import print_help;print_help()'

AssembleApp options
-----------------
--AssembleApp.access_token_timeout=<Int>
    Default: 60
    timeout to receive access token from server via polling
--AssembleApp.ballet_yml_path=<Unicode>
    Default: ''
    path to ballet.yml file of Ballet project (if Lab is not run from project
    directory)
--AssembleApp.debug=<Bool>
    Default: False
    enable debug mode (no changes made on GitHub), will read from
    $ASSEMBLE_DEBUG if present
--AssembleApp.github_token=<Unicode>
    Default: ''
    github access token, will read from $GITHUB_TOKEN if present
--AssembleApp.oauth_gateway_url=<Unicode>
    Default: 'https://github-oauth-gateway.herokuapp.com/'
    url to github-oauth-gateway server

Command line arguments

Invoke Jupyter Lab with command line arguments providing config to the ballet extension, for example:

jupyter lab --AssembleApp.debug=True

Config file

  1. Determine the path to your jupyter config file (you may have to create it if it does not exist):

    touch "$(jupyter --config-dir)/jupyter_notebook_config.py"
  2. Append desired config to the end of the file, for example:

    c.AssembleApp.debug = True

Troubleshoot

If you are see the frontend extension but it is not working, check that the server extension is enabled:

jupyter serverextension list

If the server extension is installed and enabled but your not seeing the frontend, check the frontend is installed:

jupyter labextension list

If it is installed, try:

jupyter lab clean
jupyter lab build

Contributing

Development Install

Running make install-develop will install necessary dependencies and set up the development environment. Alternatively, these steps can be be taken manually with the instructions below

The jlpm command is JupyterLab's pinned version of yarn that is installed with JupyterLab. You may use yarn or npm in lieu of jlpm below.

# Clone the repo to your local environment
# Move to ballet-assemble directory
# Install server extension with dev dependencies
pip install -e .[dev]
# Register server extension
jupyter serverextension enable --py ballet_assemble
# Install dependencies
jlpm
# Build Typescript source
jlpm build
# Link your development version of the extension with JupyterLab
jupyter labextension link .

# Rebuild Typescript source after making changes
jlpm build
# Rebuild JupyterLab after making any changes
jupyter lab build

You can watch the source directory and run JupyterLab in watch mode to watch for changes in the extension's source and automatically rebuild the extension and application.

# Watch the source directory in another terminal tab
jlpm watch
# Run jupyterlab in watch mode in one terminal tab
jupyter lab --watch

Uninstall

pip uninstall ballet_assemble
jupyter labextension uninstall ballet-assemble

Run

jupyter lab

Release process

bumpversion <part>
make release

###Feature: Code Slicing This feature allows to backtrace all code dependencies of a selected cell an extract them from a notebook.

The user receives an executable subset of code lines (the "slice") which only contains code that is needed for the computation of the selected cell. Note that code will be collected even if it does not define a feature definition.

The slice can then be submitted to an upstream repository.

The code-slicing feature can be activated upon selecting a certain cell and clicking the "SLICE" button in the toolbar.

Limitations:

  • If the cells were executed out of order (i.e. cell 1 is a dependency of cell 2, but cell 2 dragged above cell 1) the code cannot be collected.

  • Unlike the gather tool, code cells that have been deleted are no candidates for slicing. For example, if a user creates these two cells but then deletes cell 1, its code content will not be considered for the slice.

Credits

Contributors