Related update:
Skrutable's word splitting with the Dharmamitra model now preserves compounds with hyphens, whereas before it only had spaces between everything. That means that with one click you can now go from
tava karakamalasthāṃ sphāṭikīmakṣamālāṃ nakhakiraṇavibhinnāṃ dāḍimībījabuddhyā |
pratikalamanukarṣanyena kīro niṣiddhaḥ sa bhavatu mama bhūtyai vāṇi te mandahāsaḥ ||
to
tava kara-kamala-sthām sphāṭikīm akṣamālām nakha-kiraṇa-vibhinnām dāḍimī-bīja-buddhyā |
pratikalam anukarṣan yena kīraḥ niṣiddhaḥ sa bhavatu mama bhūtyai vāṇi te manda-hāsaḥ ||
(See screenshot for Devanāgarī equivalent.)
This is a little dream come true for me, personally. Hope you and your students like it, too! Again, all scientific credit goes to the Dharmamitra team.
Kind regards,
Tyler
Hi all. Yet another update here from Skrutable Land:
With the blessing of the Dharmamitra team, I’ve now added their powerful recent model as the new default option on
skrutable.info for sandhi and compound splitting. Just click the Split button as before and enjoy improved results. Also as before, Skrutable’s custom interface to the underlying API supports arbitrarily long inputs and punctuation preservation. Caveats: Such model output is never perfect, and there may still be a few bugs on Skrutable’s side. That said, I'm sharing now because I think it’s already quite useful and hope some of you will, too.
For this and some other functionality, some new controls can be found at the new page
skrutable.info/settings. For example, users can switch back to the old 2018 model if desired. More settings to come here soon.
Kind regards
Tyler