Some details and questions about this PR:

• First pass to remove the front end chat removes the frontend extension and components only used by this extension

• Remove code-toolbar by using a simplified markdown renderer in settings simplifies the rendermimeMarkdown (still used in the settings panel), to remove the code toolbar (useless in settings panel)

  • Should @jupyter/chat export its rendermimeMarkdown (basically a copy of this one) and the setting panel use it ?

• Remove chat-message-menu (should be ported in jupyter-chat) and Follow up 'Remove chat-message-menu (should be ported in jupyter-chat… remove the ChatMessageMenu and the associated contexts (selection and active-cell)

  • This is not yet supported in @jupyter/chat. Should we port this feature ?

• Remove chat handler removes the remaining ChatHandler in the frontend.

• Remove the generative AI menu removes the Generative AI menu from notebook cell

  • I think this feature should be fixed and restored

• Some components are kept, for possible future usage, but can be removed:

  • TelemetryContext, I don't know if it's ever been used
  • TooltippedButton, has been ported in @jupyter/chat, so it could be exported from there if necessary
  • ExpandableTextField, I don't know if it's ever been used

@brichet Thanks for the info! Let me answer your questions:

Should @jupyter/chat export its rendermimeMarkdown (basically a copy of this one) and the setting panel use it ?

Sure. If @jupyter/chat provides the RendermimeMarkdown component from Jupyter AI, then it's fine to delete our copy to use yours. BTW, we should probably rename that component to something more obvious, like RenderedMarkdown.

This [chat message menu] is not yet supported in @jupyter/chat. Should we port this feature ?

No need IMO. The existing feature is a bit ugly, and the replace/insert actions are unnecessary. We will probably want some kind of chat message menu in @jupyter/chat, but this component is simple enough to be re-implemented from scratch.

Remove the generative AI menu removes the Generative AI menu from notebook cell

I think this feature should be fixed and restored

Yes, deleting this is fine. Brian and I don't see the need to include this feature as part of v3 given that it will likely be superseded by some upcoming features.

Some components are kept, for possible future usage, but can be removed:
TelemetryContext, I don't know if it's ever been used
TooltippedButton, has been ported in @jupyter/chat, so it could be exported from there if necessary
ExpandableTextField, I don't know if it's ever been used

Keep TelemetryContext as a reference, but feel free to delete the other 2.

@dlqqq thanks for your answers.

I removed the remaining components, it should be ready now.

Note that the Python 3.9 tests are failing because of the kw_only argument being passed to the dataclass decorators, new to jupyterlab-chat:

@jupyter-ai/core:   warnings.warn("An error occurred.")
@jupyter-ai/core: /opt/hostedtoolcache/Python/3.9.20/x64/lib/python3.9/site-packages/jupyterlab/debuglog.py:55: UserWarning: ModuleNotFoundError: There is no labextension at .. Errors encountered: [TypeError("the 'package' argument is required to perform a relative import for '.'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'"), TypeError("dataclass() got an unexpected keyword argument 'kw_only'")]

kw_only is only available in Python 3.10+: https://docs.python.org/3/library/dataclasses.html#dataclasses.dataclass

Opened a new issue upstream to track this: jupyterlab/jupyter-chat#134