Integration with JupyterLab

[psychemedia]

I just watched a demo video for the JupyterLab Chart Editor which wraps the plotly chart editor app in a JupyterLab panel and lets you open a plotly chart JSON file in that editor. Essentially, it pops an HTML app into a panel in JupyterLab, and I think registers the app as a file viewer for a particular file type. (I'm not completely taken by it, tbh, because it means you can do irreproducible things to the chart definition file, but that's another issue).

JupyterLab extensions can also open files from a dialogue as the iframe/html previewer shows: https://github.com/timkpaine/jupyterlab_iframe.

This made me wonder about what datasette integration with JupyterLab might do.

For example, by right-clicking on a CSV file (for which there is already a CSV table view) in the file browser, offer a View / Run as datasette file viewer option that will:

• run the CSV file through csvs-to-sqlite;
• launch the datasette server and display the datasette view in a JupyterLab panel.

(? Create a new SQLite db for each CSV file and launch each datasette view on a new port? Or have a JupyterLab (session?) SQLite db that stores all datasette viewed CSVs and runs on a single port?)

As a freebie, the datasette API would allow you to run efficient SQL queries against the file eg using using pandas.read_sql() queries in a notebook in the same space.

Related:

• JupyterLab extensions docs
• a cookiecutter for wrting JupyterLab extensions using Javascript
• a cookiecutter for writing JupyterLab extensions using Typescript
• tutorial: Let’s Make an xkcd JupyterLab Extension

[simonw]

I've been thinking a bit about ways of using Jupyter Notebook more effectively with Datasette (thinks like a publish_dataframes(df1, df2, df3) function which publishes some Pandas dataframes and returns you a URL to a new hosted Datasette instance) but you're right, Jupyter Lab is potentially a much more interesting fit.

[psychemedia]

In terms of integration with pandas, I was pondering two different ways datasette/csvs_to_sqlite integration may work:

• like pandasql, to provide a SQL query layer either by a direct connection to the sqlite db or via datasette API;
• as an improvement of pandas.to_sql(), which is a bit ropey (e.g. pandas.to_sql_from_csvs(), routing the dataframe to sqlite via csvs_tosqlite rather than the dodgy mapping that pandas supports).

The pandas.publish_* idea could be quite interesting though... Would it be useful/fruitful to think about publish_ as a complement to pandas.to_?

[psychemedia]

Another route would be something like creating a datasette IPython magic for notebooks to take a dataframe and easily render it as a datasette. You'd need to run the app in the background rather than block execution in the notebook. Related to that, or to publishing a dataframe in notebook cell for use in other cells in a non-blocking way, there may be cribs in something like https://github.com/micahscopes/nbmultitask .

[MichaelTiemannOSC]

Just watched this video which demonstrates the integration of any webapp into JupyterLab: https://youtu.be/FH1dKKmvFtc

Maybe this is the answer?