openodiais a Python package which contains various tools on Odia language.- The short term goal of this package is to not make state-of-the-art methods, but to make tools which work.
- Requires Python 3.11 or higher.
- The library is tested on Python 3.11, 3.12, 3.13, 3.14, and 3.15.
pip install openodiauv is a fast Python package installer and resolver.
# Install uv if you haven't already
pip install uv
# Install openodia
uv pip install openodia
# Or add to your project
uv add openodiagit clone https://github.com/soumendrak/openodia.git
cd openodia
uv sync # or: pip install -e .For usage and further documentation please visit the Documentation page.
These are Odia-adapted variants, not classic Soundex or Metaphone.
Soundex preserves the first Odia letter, including a leading vowel, and
pads or truncates to four characters. Both encoders normalize NFC.
Vowels and matras reset duplicate suppression; halant, nukta and zero-width
joiners do not. For example, ପାପା gives Soundex ପ600 and Metaphone ପପ.
Nonempty input with no Odia letters raises ValueError, including Latin-only
input, and soundex()/metaphone() raise TypeError for a non-str such as
None. In mixed input, non-Odia characters emit nothing but reset duplicate
suppression, like vowels. Empty input produces empty keys. similarity()
does not raise for non-Odia input: it returns 0.0 when either side is None
or has no Odia letters (Latin or whitespace-only), and similarity("", "")
is 0.0. Any other type (int, bytes, list) raises TypeError. It
assigns 0.5 to Soundex equality and 0.5 to normalized Metaphone edit
similarity.
Run openodia --help or python -m openodia --help for commands, and
openodia --version for the installed version.
openodia tokenize 'ରାମ ସୀତା'
openodia tokenize --file input.txt
cat input.txt | python -m openodia tokenize -Text arguments are literal text, even when they match an existing filename.
Use --file to read a UTF-8 file explicitly (implicit file detection has been
removed). Pass - without --file to read UTF-8 text from stdin.
Every subcommand takes --json and prints one UTF-8 JSON document on stdout. Errors go to stderr with a nonzero exit code.
Version note:
--jsonon every subcommand needs a release newer than 0.1.14; PyPI 0.1.14 has it ondetect-languageonly, souvx openodia tokenize --jsonfails there with exit code 2. From a source checkout,uv run openodia ...has it. The examples below need that newer release or a source checkout; onlydetect-language --jsonworks on 0.1.14.
$ openodia tokenize --json 'ରାମ ସୀତା ଘରକୁ ଯାଏ ।'
{"tokens": ["ରାମ", "ସୀତା", "ଘରକୁ", "ଯାଏ", "।"]}
$ printf 'ରାମ ସୀତା ଘରକୁ ଯାଏ ।' | openodia detect-language - --json
{"language": "odia", "confidence_score": 0.93}See AGENTS.md for every subcommand with JSON shapes and exit codes, skills/openodia/SKILL.md for an agent skill, and llms.txt for the docs index.
If this project helps you, consider sponsoring my work on GitHub. Your support helps fund maintenance, documentation, and continued development.
Released under the MIT License.
