[prakriya] लेट् implementation based on लट् लिट् लुङ् stems

[yajnadevam]

Fixes #239

Most Panini reachable लेट् forms are now supported:

Stem	Testable	Matched	Non-compliant	Pass rate
लट्-stem	755	558	197	73.9%
लुङ्-stem	587	239	348	40.7%
लिट्-stem	160	55	105	34.4%
Total	1,502	852	650	56.7%

Vedic archaic forms cannot be derived purely from Paninian rules. This would require a system of "vedic rules" and vidyut may not be the right platform for supporting all such forms.

Types of non-compliance are:

Reason	Example	%
Vedic only forms	Weak stem retention(vidAH),vedic alternation(vas->ucCAt) etc	72.6%
guṇa/vriddhi mismatch in Panini vs vedic	Pāṇini: cicitat, Veda: ciketat; Pāṇini: vAkzati, Veda: vakzati	23.9%
Vedic intensive keeps ā	rāraṇat, vāvṛdhāti	1.8%
Vedic drops final -e in ātmanepada	aśnavanta, kṛṇavanta	1.7%

Bug due to T::Nit from #236 is also resolved
Both forms now derive correctly:

Form	Corpus	Vidyut output (current)
brū 1du atm leṬ	bravāvahai	bravAvahE, bravAvahe, bravAva, bravAvaH ✓
han 2sg paras leṬ	hanaḥ	hanaH, hanasi ✓

brū → vac fix: The vac-substitution (2.4.53) is gated on ārdhadhātuka context (ardhadhatuka::run_before_vikarana). Since leṬ is sārvadhātuka per 3.4.113, the gate's early-return fires for leṬ and 2.4.53 never runs. brū is preserved → bravāvahai.

han → ghn fix: The old T::Nit hack triggered 6.4.98. Replaced that with the real Pāṇinian path -> leṬ pratyayas are tagged Sarvadhātuka (3.4.117) + pit (3.4.94) early in samjna.rs.

[tbdasap]

Seems to produce parasmaipadi forms also for एध

Derivation for	After	Before
eDa~, BvAdi, 2, , Kartari, Let, Prathama, Eka	एधाद् , एधात् , एधातै , एधाते , एधाति	एधाते
eDa~, BvAdi, 2, , Kartari, Let, Prathama, Dvi	एधातः , एधैते	एधैते
eDa~, BvAdi, 2, , Kartari, Let, Prathama, Bahu	एधान् , एधान्ति , एधातै , एधाते	एधाते
eDa~, BvAdi, 2, , Kartari, Let, Madhyama, Eka	एधाः , एधासै , एधासे , एधासि	एधासे
eDa~, BvAdi, 2, , Kartari, Let, Madhyama, Dvi	एधाथः , एधैथे	एधैथे
eDa~, BvAdi, 2, , Kartari, Let, Madhyama, Bahu	एधाध्वै , एधाध्वे , एधाथ	एधाध्वे
eDa~, BvAdi, 2, , Kartari, Let, Uttama, Eka	एधा , एधानि , एधै	एधै
eDa~, BvAdi, 2, , Kartari, Let, Uttama, Dvi	एधाव , एधावः , एधावहै , एधावहे	एधावहे
eDa~, BvAdi, 2, , Kartari, Let, Uttama, Bahu	एधाम , एधामः , एधामहै , एधामहे	एधामहे
eDa~, BvAdi, 2, , Karmani, Let, Prathama, Eka	एध्यातै , एध्याते	एध्याते
eDa~, BvAdi, 2, , Karmani, Let, Prathama, Bahu	एध्यातै , एध्याते	एध्याते
eDa~, BvAdi, 2, , Karmani, Let, Madhyama, Eka	एध्यासै , एध्यासे	एध्यासे
eDa~, BvAdi, 2, , Karmani, Let, Madhyama, Bahu	एध्याध्वै , एध्याध्वे	एध्याध्वे
eDa~, BvAdi, 2, , Karmani, Let, Uttama, Dvi	एध्यावहै , एध्यावहे	एध्यावहे
eDa~, BvAdi, 2, , Karmani, Let, Uttama, Bahu	एध्यामहै , एध्यामहे	एध्यामहे

[yajnadevam]

Under leṬ Whitney §720: the strict dhātupāṭha pada distinction is relaxed; many dhātus take either pada in the RV. However, since we are not supporting non-Panini let forms, I will fix it

[yajnadevam]

Any chance this will be merged soon?

[tbdasap]

Any chance this will be merged soon?

Sorry about the delay. I've gist'd the diffs 110K+. Github displays this partially only. So please use a local markdown viewer (for eg. a chrome extension) to view the raw files (2 parts).

I will look at the code diffs over the coming week.

[yajnadevam]

Looks correct.

Sutras 3.4.97.v1 (paras i-lopa) and 3.4.93 (atm e→ai) — together roughly double the paradigm. The biggest driver overall is the LetStem Lun/Lit split backed by 3.4.113 / 3.4.117, which adds two entirely new stem paradigms.

Summary table these are the sutra that create the new forms:

Sūtra	What it adds	Scope / multiplier
3.4.97.v1 leṭo'pi parasmaipadeṣv ito lopo bahulam	short forms bhavāt/bhavān/bhavās alongside bhavāti/bhavānti/bhavāsi	dominant — fires on every parasmaipada cell (except 1sg uttama). Nearly doubles all paras forms.
3.4.93 (generalized) e → ai	yajātai alongside yajāte, kṛṇavai, bravāvahai	every ātmanepada cell except non-uttama dual. Doubles atm forms.
3.4.98 + vārttika (s-lopa)	karavāva ∥ karavāvaḥ	uttama dual/plural with final -s
3.4.98.v1 leṭuttamapuruṣe … pratyayalopo vā	bare-ā 1sg bhavā alongside bhavāni	1sg uttama paras
7.3.70 ghor lopo leṭi vā	dadhat ∥ dadhāti (and dā)	ghu roots dā/dhā
3.1.34 (now optional for yaj/vah/śru)	yakṣi/vakṣi/śroṣi ∥ yajāti/vahāti/śṛṇavat	sip-vikaraṇa branch
3.4.94.v2 siḍvikaraṇayoge … aṭ anityatvam	no-agama 2sg yakṣi ∥ yakṣasi	sip + 2sg paras
3.4.94 aṭ/āṭ (liṭ-stem) leṭo'ḍāṭau	long-ā papṛcāsi/jujozāt ∥ short-a	consonant-final reduplicated stem

[tbdasap]

The rest of the Let changes look good to me.

[tbdasap]

I am assuming these 3 "vartikkas" are in some specific texts.
Could you please put the links to the specific pages ?

[tbdasap]

This may not be the right place for a "7.3.*" karya. Currently, I am unable to suggest
the right "prakarana" file where this would be fit in a way that the tests pass.
I will mark a TODO for myself to move this snippet to the right file/location.