Post-processing needed for various Brahmic schemes

[akprasad]

Such schemes include Tibetan, Khmer, etc. For rationale and examples, see reshape.rs in vidyut-lipi:

https://github.com/ambuda-org/vidyut/blob/main/vidyut-lipi/src/reshape.rs

Most of these changes can be modeled as regular expressions (see ConvertFix.py from Aksharamukha, e.g.).

Options I see, from easiest to hardest:

• Do nothing. This will leave various "supported" schemes in a broken state.
• Remove support for schemes that need post-processing.
• Add the corresponding regexes in common_maps.toml. I think this should work for most cases.
• Write this logic in a single language (e.g. Rust) and bind it to each implementation language.
• Write this logic in each implementation language: Python, JavaScript, etc.

[vvasuki]

यतो भवता यथाकथञ्चिद् रस्टि परिहरिष्यते, तस्यैवात्र प्रयोगं विदधातु नाम …