Add Markdown rendering to AI Chat

[Yubo-Cao]

Closes #12234 (partially)

See this: https://drive.google.com/file/d/1VOvWBLP5E9I_sLXd7RpShfXoP0ttg0JZ/view?usp=sharing for a quick demo. Ctrl+C is supported as a part of this PR as well.

Before:

After:

Mandatory checks

• I own the copyright of the code submitted and I license it under the MIT license
• Change in CHANGELOG.md described in a way that is understandable for the average user (if change is visible to the user)
• Tests created for changes (if applicable) (not applicable)
• Manually tested changed features in running JabRef (always required)
• Screenshots added in PR description (if change is visible to the user)
• Checked developer's documentation: Is the information available and up to date? If not, I outlined it in this pull request.
• Checked documentation: Is the information available and up to date? If not, I created an issue at https://github.com/JabRef/user-documentation/issues or, even better, I submitted a pull request to the documentation repository.

[subhramit]

Code reads really good.
Did not try out, but demo is attached to the description, so green from me.
Thanks @Yubo-Cao!

[koppor]

See this: drive.google.com/file/d/1VOvWBLP5E9I_sLXd7RpShfXoP0ttg0JZ/view?usp=sharing for a quick demo.

Looks very nice.

Side track: "Summary" is a very different AI function, therefore @InAnYan put another tab for "Summary". -- For instance, for large PDFs, JabRef splits into chuncks and summarizes chunks.

[koppor]

Only a small comment.

Are spaces correctly handled for the getTextFlowContent()?

I assume, adding test cases is too hard here?

[InAnYan]

This is wonderful, Yubo!

I can copy text! And you also solved a very very naughty problem, when you could click on past messages and chat scrolled down! That problem make me feel desperate about JavaFX.

I could approve this PR without reviewing code 😄

Though, I remembered I have one small addition

[InAnYan]

Unfortunately, I had to request one small change

[InAnYan]

OMG, Yubo, is it possible to submit this code to some library? GemsFX maybe? This is a very useful piece of code

[koppor]

+1 for GemsFX. If the maintainer Dirk doesn't want it, then ControlsFX

[Yubo-Cao]

Only a small comment.

Are spaces correctly handled for the getTextFlowContent()?

I assume, adding test cases is too hard here?

So after actually switch to Arch Linux and run testcases over there (of course that was not needed at all to come up with this conclusion looking at how I butchered the Markdown AST), is that we currently don't handle whitespace perfectly, if at all.

1. White space inside a single markdown node is always preserved, so *italic with space* that kind of thing is OK
2. White space between nodes is always not preserved.
3. Bulleted indentation always get destroyed (if you nest).
4. Original markdown style marker would always disappear (think * and all those special characters)

I think my new attempt would be trying to keep track a bit more metadata in MarkdownTextflow and basically override getTextFlowContent.

[subhramit]

@Yubo-Cao Please resolve conversations after you push the commits resolving them, not before

[subhramit]

If things get too convoluted, git merge upstream/main, git reset upstream/main and then make a fresh commit. you may have to force push.

Note: gitk --all to have the commit ids ready in gitk in case something goes wrong. Your local changes should not be lost, but it is always a good idea to have a trace/backup.

[subhramit]

If things get too convoluted, git merge upstream/main, git reset upstream/main and then make a fresh commit. you may have to force push.

Note: gitk --all to have the commit ids ready in gitk in case something goes wrong. Your local changes should not be lost, but it is always a good idea to have a trace/backup.

Another handy resource: https://ohshitgit.com/

[koppor]

If things get too convoluted, git merge upstream/main, git reset upstream/main and then make a fresh commit. you may have to force push.

However, JabRef mainatinaers don't like force pushes - only in rare emergency cases ^^

I always direct to https://lostechies.com/joshuaflanagan/2010/09/03/use-gitk-to-understand-git/ -- then, the git commands should be more understandable.

[koppor]

The IOB looks like it's coming from Flexmark.
Debug and see what token(s) produce this and then file an issue at their repo

Maybe MWE using a JUnit test...

[subhramit]

Green from me.
To summarize, TODOs:

1. Submission of useful jfx snippet to GemsFX/ControlsFX
2. MWE reproduction and reporting to Flexmark
3. File known issues in JabRef

Current:
Address Ruslan's comment on the changelog entry.

Rest seems good to go, thank you for such wonderful work, Yubo.

[Yubo-Cao]

I cannot reproduce the errors that Ruslan encountered. However, I added a check for string boundaries to ensure that this won't happen. Regarding the color contrast, I simply derived -jr-theme up a bit, and I think that would be sufficient.

[InAnYan]

OMG, maybe it is something wrong with my machine, if Yubo cannot reproduce my issues 🤣

Nice color change

[calixtus]

Can you do a screenshot with dark theme too please?

[Yubo-Cao]

Can you do a screenshot with dark theme too please?

Thank you so much for the catch! I honestly forget about the dark theme users ...

[subhramit]

Failing tests are related to CSL, so I am queuing this. Nvm there's conflicts and failing build tests too.

[koppor]

Failing tests are related to CSL, so I am queuing this. Nvm there's conflicts and failing build tests too.

I reverted CSL so that the assigned contributor has time to work on it.

[trag-bot]

@trag-bot didn't find any issues in the code! ✅✨