Skip to content

Transliterate Odia

Convert Odia to another Indic script, or to Latin letters. Both conversions run offline, and the same input always gives the same output.

indic and roman need a release newer than 0.1.14. From a source checkout, uv run python can import them. Every example on this page is run by the test suite.

Convert between Indic scripts

indic.transliterate changes only characters that belong to from_script. Every other character stays. The names you can pass are indic.SCRIPTS.

>>> from openodia import indic
>>> indic.SCRIPTS
('devanagari', 'bengali', 'assamese', 'gurmukhi', 'gujarati', 'odia', 'tamil', 'telugu', 'kannada', 'malayalam')
>>> indic.transliterate("ଓଡ଼ିଆ", from_script="odia", to_script="devanagari")
'ओड़िआ'
>>> indic.transliterate("नमस्ते", from_script="devanagari", to_script="odia")
'ନମସ୍ତେ'
>>> indic.transliterate("file କ", from_script="odia", to_script="devanagari")
'file क'

The same script on both sides returns the text unchanged. An unknown name raises ValueError.

>>> indic.transliterate("ଓଡ଼ିଆ", from_script="odia", to_script="odia")
'ଓଡ଼ିଆ'
>>> indic.transliterate("କ", from_script="nope", to_script="odia")
Traceback (most recent call last):
  ...
ValueError: unknown from_script 'nope'; supported: ('devanagari', 'bengali', 'assamese', 'gurmukhi', 'gujarati', 'odia', 'tamil', 'telugu', 'kannada', 'malayalam')

A letter with no counterpart stays as it is, and some pairs do not round-trip.

>>> indic.transliterate("ଖ", from_script="odia", to_script="tamil")
'ଖ'
>>> indic.transliterate("ୱ", from_script="odia", to_script="devanagari")
'ୱ'
>>> indic.transliterate("व", from_script="devanagari", to_script="bengali")
'ব'
>>> indic.transliterate("ব", from_script="bengali", to_script="devanagari")
'ब'

ଖ stays because Tamil has no letter for that Odia consonant. Odia ୱ (WA) stays because the other scripts have no WA. Devanagari व (VA) is written as Bengali BA, and that Bengali letter comes back as Devanagari BA.

Write Latin letters

roman.to_roman writes Latin letters. roman.from_roman reads them back. The schemes are ISO15919, Hunterian, and ITRANS.

>>> from openodia import roman
>>> roman.to_roman("ଓଡ଼ିଆ", "ISO15919")
'ōṛiā'
>>> roman.to_roman("ଓଡ଼ିଆ", "Hunterian")
'oria'
>>> roman.to_roman("ଓଡ଼ିଆ", "ITRANS")
'o.DiA'
>>> roman.from_roman(roman.to_roman("ଓଡ଼ିଆ", "ISO15919"), "ISO15919")
'ଓଡ଼ିଆ'

Hunterian drops vowel length, so reading a Hunterian spelling does not put that length back. Hunterian and ITRANS write Odia digits as ASCII digits. Hunterian still does that when reversible=True.

>>> roman.from_roman("namaskara", "Hunterian")
'ନମସ୍କର'
>>> roman.to_roman("୨୦୨୬", "ITRANS")
'2026'
>>> roman.to_roman("୨୦୨୬", "Hunterian", reversible=True)
'2026'

Characters outside the Odia block are copied through, unless you pass reversible=True. to_roman("କi") returns kai, which is not a reversible spelling of that string. reversible=True quotes the non-Odia part so from_roman can return it. Hunterian ignores reversible. An unknown scheme raises ValueError.

>>> roman.to_roman("କi")
'kai'
>>> roman.to_roman("କi", reversible=True)
'ka:{i:}'
>>> roman.from_roman(roman.to_roman("କi", reversible=True), "ISO15919")
'କi'
>>> roman.to_roman("କ", "nope")
Traceback (most recent call last):
  ...
ValueError: unknown scheme 'nope'; supported: ('ISO15919', 'Hunterian', 'ITRANS')

Next