Maths support #3251
Maths support #3251
Conversation
… when isMathEnabled
|
closed while refactoring code to be more in keeping with proposals |
|
this has now entered testing at www.pigeon.digital , and the last commit fixes some problems reported by users |
|
Hey all, |
|
i haven't had any time to update this to work with the new message composer, and i'm not sure if much has been done since the last discussion in the room... i'd also love to finish this off but have really lost the swing of things, and can't remember where things were left off |
|
Is anyone able to make this work as a test in the coming weeks as it was done previously above? I am a mathematics PhD at Paris-Saclay. We are all locked in at home now due to COVID-19. I would like to get students and teachers on Riot to continue courses, study groups, etc. Getting this running soon would be super helpful for many Universities worldwide in those days, even if buggy! I am unfortunately unable to do it for sure, but I promise to promote Riot (Matrix) to as many Universities and Professors I have contact to if someone manages to get it done. |
There is a demo build running at https://pigeon.digital that should work fine. I believe it was recently used at a conference that wanted to do some remote interaction. Unless there's a massive rush of traffic, it should continue to work as a temporary solution, but I'm glad to hear that there's still interest in the mathematics community for this! |
|
How much funding would this need to complete? |
|
I haven't looked into this much further, then simply trying to rebase this on the current develop branch, since my JS-fu is basically non-existent, but this branch rebases almost cleanly, with only a trivial conflict in I'd really love to see this land, how can I help to make this happen? |
|
The last I heard was that this was going to be a Google Summer of Code project for somebody, because the team (very understandably) wanted any solution to not be simply for Riot, but for Matrix clients as a whole, which makes things quite a bit more difficult. |
|
@thosgood Thanks for the information! I just read the GSoC proposal and this PR is listed as a partial implementation. That is because it is only client side and there is no server endpoint that would generate a fallback (e.g. an image) or is there anything else missing? I guess most LaTeX users would be happy with the LaTeX code as fallback, as that is what (most of them) they are typing right now already. |
|
@behrmann No server fallback is the main thing missing here, yes. Not to mention that the message sent is just some custom type that mixes together a few different suggestions, whereas the "final" proposal might end up using extensible events or something like this (but this, again, seems to be a bit up in the air still). I agree that, at least as a first goal, LaTeX implementation should really be aimed at the sort of people who use LaTeX often, i.e. not worrying too much if the fallback is just the code itself. Adding better fallbacks/better input methods (some WYSIWYG bar or something) seems like something that should come afterwards, to me. |
|
It seems like the Project unfortunately wasn't picked up by a student for the summer of code, or I am unable to find it on the page. I'd still be happy if this would land, just as a stop gap until a more formal MSC is approved, as this would be quite a killer feature for the natural sciences. |
|
The latest Element.io blog post references this issue, as a feature that will be implemented to ease the feelings of people disagreeing with the name change: https://element.io/blog/welcome-to-element/ So I guess LaTeX math support is on the roadmap after all! |
|
just as a small note, the current version of pigeon.digital uses the latest (pre-Element) build of Riot, i.e. with the new composer, and with all LaTeX rendering options hidden from the user (and all enabled by default), following user feedback |
|
@thosgood Sounds excellent. Do you know if the work is based in this PR or whether it is happening separately (I didn't see any other open PR). Would you be open to updating this PR (and the accompanying riot/element PR) to the version you seem to have in production? I'd love to give it a spin. |
|
@behrmann not too sure: all discussion (that I know of) is taking place in the pigeon.digital room. I could try updating this PR, but I think I'd be lost in git hell for a while in trying to do so... honestly, the changes that I had to make to the latest Riot release were so tiny: include the |
|
@thosgood You can simply rename |
|
I just gave the |
|
yeah, editing messages required a bit more finesse, and was never something that I gave much time to. |
|
as a warning, I think the |
Fair enough. Editing will be important though. I never seem to be able to get long LaTex expressions right on the first try. |
|
@thosgood This has conflicts that need tp be resolved. Would be great to get a status update too. :-) |
|
this has been superseded by #5244 :-) |
|
I guess this should be closed as #5244 has been merged? |
|
Yes, looks like we can close this now. |
|
Are there any plans to bring Maths support to iOS and/or Android? |
For Android, see element-hq/element-android#2133 |
Linked to element-hq/element-web#10428
Added support for mathematics rendering using KaTeX (see element-hq/element-web#1945 (comment)).
Edit: Now enabling maths mode replaces anything wrapped in dollar signs with the same thing wrapped in
<matrix-math>(or<matrix-math-display>) tags in theformatted_body. Enabling KaTeX rendering (as an account option) then uses KaTeX to render anything between such tags.This is an improvement over the previous commit, where the distinction between math mode and arbitrary
org.matrix.custom.htmlcontent was non-existent. Now things really can be changed to fit any proposals very easily (fallbacks, KaTeX vs. MathJax, delimiter choices, etc. etc.).The only remaining 'bug' is that I haven't yet implemented per-room KaTeX rendering settings.
Signed-off-by: Tim Hosgood timhosgood@gmail.com