vidyut-lipi has reasonably good support for a variety of schemes, which are stored in the Scheme enum. But I am not an expert on all schemes (I am barely an expert in one), and I am sure that most of our schemes have at least one error.
Proposed solution
Pick a scheme in Scheme and read its documentation on our doc comments (or, find some documentation if the docstring doesn't have any). For roman schemes like ITRANS, every possible value in that scheme should be mapped to at least one Devanagari sound. For other schemes, every possible value to the extent it's possible should be mapped to at least one Devanagari sound.
Support for a scheme should include:
support for alternates, e.g. ITRANS supports both I and ee.
unit tests of the following kinds:
converting to Devanagari
converting from Devanagari
(if the scheme has alternates) converting to and from Devanagari, and confirming that the result uses the preferred form of the alternate (e.g. ee to Devanagari and back should be I.)
Problem
vidyut-lipi has reasonably good support for a variety of schemes, which are stored in the
Scheme
enum. But I am not an expert on all schemes (I am barely an expert in one), and I am sure that most of our schemes have at least one error.Proposed solution
Pick a scheme in
Scheme
and read its documentation on our doc comments (or, find some documentation if the docstring doesn't have any). For roman schemes like ITRANS, every possible value in that scheme should be mapped to at least one Devanagari sound. For other schemes, every possible value to the extent it's possible should be mapped to at least one Devanagari sound.Support for a scheme should include:
I
andee
.ee
to Devanagari and back should beI
.)