Closed bollwyvl closed 2 years ago
I don't have strong opinions about this but I'd love to get the input from the upstream devs of nbconvert. (I'm guessing you are one of them, right?)
Can you bring this up with them and get their opinion on this?
From the nbconvert issue, we've got one :+1: from the most active maintainer... we can let it cook for a while, but i can knock together a wip PR...
If the most active maintainer is happy, go for it!
Welp, here we go: #47
Hey folks!
Over on https://github.com/conda-forge/staged-recipes/pull/12516#discussion_r502814168 we've been chatting about the tradeoffs of offering a
jupyter-book-webpdf
, which would be somewhat simpler if there was annbconvert-webpdf
which already carried thepyppeteer
pin, beyond whatrun_constrained
would do.However we solve it over there, we could take the opportunity to offer, a la
matplotlib[-base]
nbconvert-core
(which could benoarch: python
) that still carried all therun_constrained
pieces, but was lighter-weight and -featured (#24)nbconvert
that actually requiredpandoc
, and therefore could not benoarch: python
, for nownbconvert-webpdf
that acutally requiredpyppeteer
, and could actually also benoarch: python
It would take a while (or maybe a migration) to get everything that didn't specifically need the
pandoc
capability to use-core
, but eventually, maybe, (more) people could be happy(er).I'm happy to work up a PR if people think this is a good idea.