Data
====

We have designed `vidyut-prakriya` so that it can run without any side data.
But in practice, it is useful to have a dhatupatha available to generate
various words and stems. For this reason, the `vidyut-prakriya` crate includes
`dhatupatha.tsv`, a comprehensive Dhatupatha that we also use in our exhaustive
test suite.


Creating the data file
----------------------

`dhatupatha.tsv` was sourced from [ashtadhyayi.com][a-com], and the author
of ashtadhyayi.com has graciously agreed to share their dhatupatha with us
under an MIT license.

This dhatupatha is a superset of traditional dhatupathas from five different
sources:

- the *Siddhāntakaumudī*
- the *Bṛhaddhātukusumākaraḥ*
- the *Mādhavīyadhātuvṛttiḥ*
- the *Kṣīrataraṅgiṇī*
- the *Dhātupradīpaḥ*

The source data is [available on Github][dhatupatha-json] as a JSON file. To
create `dhatupatha.tsv`, we downloaded this file and made the following changes:

1. First, we converted the original JSON file to a TSV file with three columns.
   For our current needs, these fields are sufficient.

2. Next, we added svaras to the dhatus in this list. Many of the dhatus here
   overlap with the dhatus used in [SanskritVerb][sanskrit-verb], a PHP-based
   Paninian generator that heavily inspired `vidyut-prakriya`. Where we could
   find an overlap, we used the svaras provided by SanskritVerb. Otherwise, we
   added svaras through a simple mechanical procedure: *aniṭ* roots have
   *anudātta* on the root vowel, *ātmanepadī* roots without *ṅit* are
   *anudāttet*, and *ubhayapadī* roots without *ñit* are *svaritet*.

3. Finally, we made a few corrections to the dhatus in this list. The original
   JSON file will likely be updated with these corrections, so we have omitted
   the details for now.

[dhatupatha-json]: https://github.com/ashtadhyayi-com/data/blob/master/dhatu/data.txt
[sanskrit-verb]: https://github.com/drdhaval2785/SanskritVerb


[a-com]: https://ashtadhyayi.com