[prakriya] let-lakara fixes phase1: Fix Bhvadi current implementation

[yajnadevam]

Fixes #236

bhū + leṬ final paradigm — exact match to Macdonell §159:

	sg	du	pl
3	BavAt / BavAti	BavAtaH	BavAn / BavAnti
2	BavAH / BavAsi	BavATaH	BavATa
1	BavA / BavAni	BavAva / BavAvaH	BavAma / BavAmaH

Code summary (vidyut-prakriya/src/tin_pratyaya.rs):

1. Lines 125-130: mip → ni (3.4.89) for leṬ 1sg — mirrors Loṭ.
2. Lines 137-150: pit/Nit fork via optional_run("3.4.94.v1") — produces both BavAti (pit) and the substrate for Nit-branch processing.
3. Lines 167-178: New 3.4.98.v1 vārttika for bare-ā 1sg paras under leṬ Uttama (luk-elides the entire pratyaya).
4. Lines 277-292: leṬ Nit-branch 3.4.100 handler, with 1sg Uttama excluded per Kashika on 3.4.103 (lakāra-derived Nit-tva doesn't propagate to the ni
   substitute).

Test edits:

• tests/integration/kashika_3_1.rs::sutra_3_1_34 — added Nit-branch outputs (patAt, cyAvayAt).
• tests/integration/kashika_3_4.rs::sutra_3_4_94, sutra_3_4_97 — same.
• tests/integration/kashika_3_4.rs::sutra_3_4_95 — added kurvEte/kurvETe from 1.2.4 ngit-cascade.
• tests/integration/kashika_3_4.rs::sutra_3_4_98 — added kurvAva/kurvAvaH/kurvAma/kurvAmaH from same cascade.
• tests/integration/kashika_3_4.rs::bhu_let_paradigm_macdonell_159 — new, locks the full bhū paradigm.

[yajnadevam]

macos-13 has hit end of life sep 15,2025 ... 8 months ago. Github actions seems to be stuck on it. Is it really necessary to support vidyut on it? @tbdasap

[yajnadevam]

@akprasad macos-13 times out ... can we merge without it

[akprasad]

@yajnadevam yeah, we can merge without it.

[yajnadevam]

@yajnadevam yeah, we can merge without it.

Lets go ahead please

[akprasad]

@tbdasap can you take a look? if you're busy over the next few days, I'll look Wednesday.

[tbdasap]

Perhaps put similar page references if applicable as appropriate in other changes ?

[tbdasap]

LGTM - The "terminology", like 1sg, 2du etc. feels different though but that also seems to be the style in the referred book.
Please edit the title to be [prakriya] let-lakara .... (as per @akprasad guidance on naming convention)

[tbdasap]

Please see the gist of differences in tinnanta output (9997).
Is it expected that it would affect only some dhatus ?

[yajnadevam]

For phase I, I have targeted Bhvadi cases where the stem is correct. This is a small fix that will address the most common issues and fixes the existing implementation.

A complete leT fix is a much larger task and I have created #239 to address it.

[akprasad]

Please edit the title to be [prakriya] let-lakara .... (as per @akprasad guidance on naming convention)

feel free to also rewrite the PR title and parts of the description on merge