विद्वस्

[vvasuki]

विद्वोभ्याम्

विद्वस् पुं
विद्वाः; विद्वसौ; विद्वसः
विद्वसम्; विद्वसौ; विद्वसः
विद्वसा; विद्वोभ्याम्; विद्वोभिः
विद्वसे; विद्वोभ्याम्; विद्वोभ्यः
विद्वसः; विद्वोभ्याम्; विद्वोभ्यः
विद्वसः; विद्वसोः; विद्वसाम्
विद्वसि; विद्वसोः; विद्वस्सु, विद्वःसु
विद्वः; विद्वसौ; विद्वसः

इति दोषाय।

[gasyoun]

Why do you continue to use this software @vvasuki ?

[vvasuki]

Why do you continue to use this software @vvasuki ?

Es super.

[akprasad]

This behavior is actually correct, though it's admittedly very confusing.

If you pass vidvas as a plain pratipadika, it will be treated like manas, tejas, etc. What you should do instead is pass the krdanta vid + vasu~. But vasu~ isn't exposed as a Krt (this is my error), so we should instead use vid + Satf~, where Satf~ is replaced with vasu~ (https://ashtadhyayi.com/sutraani/7/1/36):

from vidyut.prakriya import *

vid = Dhatu.mula("vida~", Gana.Adadi)
vidvas = Pratipadika.krdanta(vid, Krt.Satf)
vidvan = Pada.Subanta(vidvas, Linga.Pum, Vibhakti.Prathama, Vacana.Eka)

v = Vyakarana()
for p in v.derive(vidvan):
    print(p.text)

[vvasuki]

This behavior is actually correct, though it's admittedly very confusing.

If you pass vidvas as a plain pratipadika, it will be treated like manas, tejas, etc.

तर्हि kosha.pratipadikas() इत्यस्मिन्न् अपि "विद्वस्" इत्य् अविशिष्टम् प्रातिपदिकं न स्याद् इत्य् आशयः।
अन्यथैवं भ्रमो जायेत यन्त्रस्यापि मनुष्याणाम् अपि।

[akprasad]

Agreed.

The root cause is that I'm pulling pratipadikas from a scrape of the Monier-Williams dictionary and it contains many entries that are simply krdantas. The simple solution is to delete every entry that is also a krdanta, but this is too general; for example, this approach will remove राम (रम् + घञ्). And doing nothing causes the problem you mention.

So there needs to be some in-between solution that cleans up most of these. I am hoping to avoid manual annotation but that might ultimately be necessary.