ibis-vega-transform

A JupyterLab extension for performing Vega transforms lazily using Ibis.

Python evaluation of Vega transforms using Ibis expressions.

For inspiration, see https://github.com/jakevdp/altair-transform

This extension is composed of a Python package named ibis-vega-transform for the server extension and a NPM package named ibis-vega-transform for the frontend extension.

Requirements

• JupyterLab >= 3.0

Getting started

pip install ibis-vega-transform

Then in a notebook, import the Python package and pass in an ibis expression to a Altair chart:

import altair as alt
import ibis_vega_transform
import ibis
import pandas as pd


source = pd.DataFrame({
    'a': ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I'],
    'b': [28, 55, 43, 91, 81, 53, 19, 87, 52]
})

# or ibis.pandas if ibis version < 1.4
connection = ibis.backends.pandas.connect({'source': source })
table = connection.table('source')

alt.Chart(table).mark_bar().encode(
    x='a',
    y='b'
)

Check out the notebooks in the ./examples/ directory to see some options using interactive charts and the OmniSci backend.

Usage

Importing ibis_vega_transform sets the altair renderer and data transformer to "ibis". It also monkeypatches the Ibis chart constructor to handle ibis expressions.

Now, whenever you pass an ibis expression to a chart constructor, it will use the custom ibis renderer, which pushes all data aggregates to ibis, instead of in the browser.

You can also set a debug flag, to have it instead pull in the first N rows of the ibis expression and use the default renderer. This is useful to see how the default pipeline would have rendered your chart. If you are getting some error, I reccomend setting this first to see if the error was on the Altair side or on the ibis-vega-transform side. If the fallback chart rendered correctly, it means the error is in this codebase. If it's wrong, then the error is in your code or in altair or in Vega.

# enable fallback mode
ibis_vega_transform.set_fallback(True)
# disable fallback mode (the default)
ibis_vega_transform.set_fallback(False)

Tracing

If you want to see traces of the interactions for debugging and performance analysis, install the jaeger-all-in-one binary and the jupyterlab-server-proxy lab extension to see the Jaeger icon in the launcher.

conda install jaeger -c conda-forge
jupyter labextension install jupyterlab-server-proxy-saulshanabrook

The Jaeger server won't actually be started until a HTTP request is sent to it, so before you run your visualization, click the "Jaeger" icon in the JupyterLab launcher or go to /jaeger to open the UI. Then run your visualization and you should see the traces appear in Jaeger.

You also will likely have to increase the max UDP packet size on your OS to accomdate for the large logs:

Mac

# Edit now
sudo sysctl net.inet.udp.maxdgram=200000
# Edit on restart
echo net.inet.udp.maxdgram=200000 | sudo tee -a /etc/sysctl.conf

Troubleshoot

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

jupyter server extension list

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

jupyter labextension list

Contributing

Development install

Note: You will need NodeJS to build the extension package.

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
git clone git@github.com:Quansight/ibis-vega-transform.git

# Change directory to the ibis-vega-transform directory and
# Create a conda environment
cd ibis-vega-transform
conda env create -f binder/environment.yml
conda activate ibis-vega-transform

# Install package in development mode
pip install -e .
# Link your development version of the extension with JupyterLab
jupyter labextension develop . --overwrite
# Rebuild extension Typescript source after making changes
jlpm run build

You can watch the source directory and run JupyterLab at the same time in different terminals to watch for changes in the extension's source and automatically rebuild the extension.

# Watch the source directory in one terminal, automatically rebuilding when needed
jlpm run watch
# Run JupyterLab in another terminal
jupyter lab

With the watch command running, every saved change will immediately be built locally and available in your running JupyterLab. Refresh JupyterLab to load the change in your browser (you may need to wait several seconds for the extension to be rebuilt).

By default, the jlpm run build command generates the source maps for this extension to make it easier to debug using the browser dev tools. To also generate source maps for the JupyterLab core extensions, you can run the following command:

jupyter lab build --minimize=False

A pre-commit hook is installed usig Husky (Git > 2.13 is required!) to format files.

Run the formatting tools at any time using:

black ibis_vega_transform
jlpm run prettier

Tracing

We are using jupyter-jaeger to trace each interaction for benchmarking.

Uninstall

pip uninstall ibis-vega-transform