8000 Stop generating pdf docs? · Issue #17051 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]

Skip to content

Stop generating pdf docs? #17051

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
NicolasHug opened this issue Apr 26, 2020 · 15 comments · Fixed by #17564
Closed

Stop generating pdf docs? #17051

NicolasHug opened this issue Apr 26, 2020 · 15 comments · Fixed by #17564

Comments

@NicolasHug
Copy link
Member

Is anyone strongly against stopping the generation of pdf docs?

CI is currently broken because of a LaTeX error. It happens every now and then, and it blocks other PRs. LaTeX error messages being what they are, these issues can take time and effot to fix.

Should we just stop generating the pdfs? It makes us waste time, and I personally doubt these are really useful anyway. We could just provide an archive of the html docs instead, which would be more readable.

@rth
Copy link
Member
rth commented Apr 26, 2020

+1 to stop generating it at each commit on master but I think it's still useful to make them once per release. We had issues in the past about people using it, and it's nice to have something off-line.

@NicolasHug
Copy link
Member Author

how about providing a zip of the html docs for off-line browsing? I feel like this would be more usable than a big pdf

@rth
Copy link
Member
rth commented Apr 26, 2020

Let's discuss it in the meeting tomorrow?

@jnothman
Copy link
Member
jnothman commented Apr 27, 2020 via email

@adrinjalali
Copy link
Member

I think this needs to be solved for the release. @rth would you be able to put together a PR for this one maybe?

@NicolasHug
8000
Copy link
Member Author

The actual blocker is the LaTeX issue right? It would be fine to still ship a pdf for this version

@adrinjalali
Copy link
Member

The actual blocker is the LaTeX issue right? It would be fine to still ship a pdf for this version

Yes.

@thomasjpfan
Copy link
Member

I'll look into it today. (Currently, I am not able to reproduce)

@thomasjpfan
Copy link
Member

It fixed itself on master.

@NicolasHug
Copy link
Member Author

love it.

Apparently the error comes from a hyperref being split across 2 pages (details here, and as far as I understand there's pretty much no fix). It might happen again if we get unlucky with the docs layout.

@alfaro96
Copy link
Member
alfaro96 commented Jun 10, 2020

Hey!

Have you reached a consensus about this issue (+1 about providing a zip of HTML docs instead of a PDF). I may work on this issue if so.

@thomasjpfan
Copy link
Member
thomasjpfan commented Jun 10, 2020

I haven't thought this out yet, but this maybe blocked by case insensitive file systems: #13022

For example the generated html for DBSCAN and dbscan:

https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html

https://scikit-learn.org/stable/modules/generated/sklearn.cluster.dbscan.html

may overwrite each other in a case insensitive file system.

@rth
Copy link
Member
rth commented Jun 10, 2020

Has anyone tried generating epub for the docs? It is supported by sphinx and it uses HTLM internally so we won't have to deal with LaTeX.

@alfaro96
Copy link
Member

I have a preliminary version for the CI with a ZIP of the HTML. I will try to generate an EPUB to know how it looks.

@alfaro96
Copy link
Member
alfaro96 commented Jun 10, 2020

This link provides the ZIP and EPUB version for the offline documentation.

I can submit a PR to include any of these versions in the CI. IMHO, I consider that the ZIP of the HTML is more usable.

WDYT?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

6 participants
0