elyra-ai / streamlit-extension Goto Github PK

We should decide on how we want to license this then add the license headers as needed.

Currently its under the license that the cookie cutter included, we should find out if this means the files generated by the cookie cutter are licensed or if we can change the license as desired.

I used the API to dynamically add the context menu items. We should switch to using settings like seen in this example https://github.com/jupyterlab/extension-examples/tree/master/context-menu

Soon after this repo was generated by the cookie cutter, it was updated to include tests. We should add tests for this extension, if needed we can leverage the latest cookie cutter to generate the initial tests and copy them into the repo.

We could potentially add some basic integration with ELyra by adding a streamlit button to the toolbar on the python editor. Which side this should be done from and how closely we integrate should be decided based on the results of #5

I think I've got this working but I can't seem to integerate this link:

https://docs.streamlit.io/get-started/tutorials/create-a-multipage-app

With the extension, I assume the page should open up in JupyterLab notebook?

Is it possible to add a guide showing how to get the multi app page demo from the Streamlit tutorial
up and running?

Or some example on how to start a Streamlit app within a JupyterLab Notebook?

When demoing the extension last month I hit an error when attempting to drag and drop a file into the drag and drop a file section in a streamlit app. I need to follow up on this and figure out what happened, why, and how it can be addressed.

This extension is currently specific to Streamlit. We should find out if we can make it more generic to any web app, and if so refactor and split out that generic code into a lib package for use elsewhere.

This extension repo should be donated to an open source org, we should decide which org it best belongs to and figure out the steps to take.

Potential orgs (in order of probability)

1. elyra-ai
2. CODAIT
3. jupyterlab

Streamlit has a built in file upload widget that supports drag and drop from the file system in addition to a browse option. The file widget will allow upload of local files, but not ones in the jupyterlab filesystem (important for cloud instances of jupyterlab).

We should add a hook/widget that enables users to add files from the jupyterlab filesystem to the streamlit file widget. The most straight forward method would be to add drag and drop support from the jupyterlab file browser.

The cookie cutter used to initialize this repo included a lot of generated files. We should take stock of which of these files are necessary for our implementation and remove any extraneous files or code.