[Feature Request] Support for non-Devnagari transliteration

[ma08]

Hi,

Great project!! Really interesting and inspiring.

Are you interested in adding support for non-Devnagari transliteration, maybe one or all of IAST, ITRANS, SLP1? Will be useful for folks with difficulty in reading Devnagari.

I would love to contribute to developing this feature. Maybe for a start, we could implement a small transient popup that shows the transliteration(s) near the pointer on extended mouse hover?

[akprasad]

Thanks for filing this issue!

Yes, I'm very interested in having a fast and standard transliteration engine. If it's implemented in Rust, we can have a single engine for both backend and frontend work. Ideally we can have a port of https://github.com/virtualvinodh/aksharamukha-python, which is generally the most sophisticated transliterator available for Sanskrit.

As for the transliteration itself, there is a PR in progress at #39 that @mudgebiscuit is working on. @mudgebiscuit -- as you can see, @ma08 is eager to help with this work. Is there something related to your PR that he can get involved with?

(As a reminder, Vidyut is a backend engine -- its goal is to be infrastructure for other projects, and it doesn't have a frontend of its own. Our Vidyullekha demo https://ambuda-org.github.io/vidyullekha/ , which you might have had in mind, is simply a demo/debugger for the backend code. But I'm happy to receive a PR there to switch the script if that's something you're excited about -- could be a dropdown that, when changed, just changes the script used on the frontend.)

[ma08]

Thanks for the quick response and the info @akprasad! I realize I should learn more about the ecosystem of ambuda. I will go through the docs to understand it more. I was a bit eager to get started and posted this issue!

So the Vidyullekha demo is hosted using the https://github.com/ambuda-org/ambuda-org.github.io repo which has the static content?

[akprasad]

No problem -- I love the enthusiasm! :)

So the Vidyullekha demo is hosted using the https://github.com/ambuda-org/ambuda-org.github.io repo which has the static content?

Yes -- it's a copy of the files we generate here.

Basic project layout and terms:

• Ambuda: the project responsible for https://ambuda.org/ and all content there, including the reader, the proofing interface, text and dictionary data, etc.
• Vidyut: the backend analysis layer that we plan to use extensively on Ambuda. It's all Rust code, though for ease of use we also have Python bindings available. These bindings are how we plan to use Vidyut on Ambuda.

Long-term, Ambuda will be a comprehensive library of Sanskrit literature and associated tools, like corpus search etc. Long-term, Vidyut will be the stable, fast, and reliable backend layer that supports all of the difficult computational work that projects like Ambuda need.

[ma08]

Got it! Thanks for the links and the summary, the reason I am partial to this particular feature is because it would let me appreciate the demo/backend more. Currently I am having a hard time consuming the text on the demo for Vidyullekha.

As for the dropdown for the script, I can probably use one similar to what's implemented here https://en.amarahasa.com/start-here/. Where can I find the source code for that?

[akprasad]

Amarahasa is using some older code. Instead, I suggest that you update app.js to have a new script variable in const App. Then, update dev and devaNoSvara to translate to this.script. Finally, add your dropdown in index.html with x-model=script. That should broadly do the trick -- when the dropdown is changed, script will be updated, and that will cause deva and devaNoSvara to re-run automatically.

Useful docs:

• https://alpinejs.dev/ for our JS code
• https://github.com/indic-transliteration/sanscript.js for values you can pass to Sanscript. We don't need all of these -- devanagari, iast, and a few Indian scripts that support Sanskrit (kannada, telugu) are enough.

[ma08]

Thanks @akprasad for the detailed pointers!! I will get started on these.