Sanskrit Transliteration Tool

Supported Transliteration Schemes

Canonical ID	Scheme Name	Aliases
iast	IAST	IAST
iso	ISO 15919	ISO 15919, iso15919
kh	Harvard-Kyoto	Harvard-Kyoto, harvardkyoto
tf	Tokunaga-Fujii	Tokunaga-Fujii, tokunagafujii
slp1	SLP1	SLP1
ipa	IPA	IPA
csx	CSX (code point)	CSX

All scheme names are case-insensitive. You can use either the canonical ID or any alias listed above.

Usage

Install the package using pip:

pip install git+https://github.com/Yuzki/SanskritTransliteration.git

Import the library and use the transliterate function:

from SanskritTransliteration import transliterate

text = "agn;im ii;le pur;ohitam"
s = transliterate(text, "tf", "iso")
print(s)

This will output:

agním īḷe puróhitam

Aliases work the same way:

s = transliterate(text, "Tokunaga-Fujii", "ISO 15919")
print(s)  # agním īḷe puróhitam

Ambiguous Reverse Mappings

When multiple CSV rows map the same source token to different targets, the first row in the table takes precedence (first-win rule).

SLP1 Anunāsika Extension

ISO 15919 distinguishes anunāsika m̐ from anusvāra ṃ, but standard SLP1 uses a single token M for both. To keep ISO -> SLP1 -> ISO reversible, this library emits a project-specific extended SLP1 token for anunāsika:

ISO 15919	SLP1 output
ṃ (anusvāra)	M
m̐ (anunāsika)	M~

On the reverse path, M maps back to ṃ and M~ maps back to m̐.

M~ is a reversible extension specific to this library, not part of standard SLP1. Consumers that require strict external SLP1 compatibility should account for this extension (for example, by stripping a trailing ~ when the anunāsika distinction is not needed).