Add support for 4 plainb formats

[Yahiewi]

This PR addresses issue #46 by adding support for the 4 formats supported by jupyter-plainb:

Supported formats

Parser	Menu label	File type	Format
parsePy	Percent format (.py)	.py	Jupytext percent
parseSphinxGallery	Sphinx Gallery (.py)	.py	Sphinx-Gallery
parseClassicMd	Classic Markdown (.md)	.md	Standard markdown — fenced code blocks become code cells
parseMystMd	MyST Notebook (.md)	.md	MyST-NB {code-cell} directives

I also added a plainb subdirectory under demo/files which contains the examples from jupyter-plainb.
This needs jupyter-plainb to be installed in the testing environment.
Similar to jupyter-plainb, Specta options are added to the "Open With" context menu.
Here is a screencast:

Screencast.from.2026-06-01.14-18-26.mp4

I'm not sure why the output cells in the specta view (dashboard mode I believe) are aligned to the left and I'm currently investigating it.
EDIT: This seems to be a general specta problem for which I opened issue #49

[trungleduc]

you need to update this https://github.com/notebook-link/specta/actions/runs/26754522649/job/78853130650#step:4:1211

[Yahiewi]

Thanks @trungleduc CI is green now.

[trungleduc]

does this mean specta needs to depend on jupyter-plainb?

[Yahiewi]

does this mean specta needs to depend on jupyter-plainb?

Only for this feature to work, otherwise if jupyter-plainb isn't installed nothing happens and specta works as usual.
It's also not a dependency as I didn't add it to the package.json, I only installed it using pip

[trungleduc]

hmm it does not sound practical since you need to know about jupyter-plainb to use this feature. How about adding plainb to packages.json and using the parser to convert the content directly in specta? So specta work without any external deps

[trungleduc]

and please verify if the stanalone specta app also works

[Yahiewi]

Hey @trungleduc I've made some changes to the plainb library by adding format detection in this PR and I used it in jupyter-plainb in this PR and now I also updated this PR to use it. The reason I made this change, is because in the standalone specta app, files open directly when clicking on them, so having the extension know what format a plaintext file is in, would allow it to use the correct parser.
Also as we no longer have jupyter-plainb installed like you asked, I created a new plainb_factory.ts which has code very similar to model.ts from jupyter-plainb.
Here is a demo in the standalone app:

Screencast.from.2026-06-03.14-27-06.mp4

[trungleduc]

The specta metadata still works with the plainb file?

[Yahiewi]

The specta metadata still works with the plainb file?

Yes, I've added different specta metadata configurations for the 5 plain demo examples.

Here is a screencast of the examples being viewed in the Specta standalone app:

Screencast.from.2026-06-05.15-25-59.mp4

[trungleduc]

looks copy-pasted from previous block, could we refactor it?

[Yahiewi]

Thanks for the suggestion, I refactored duplicated code into helper functions getSpectaDocInfo and openDocument in tool.ts