udtools
Python tools for Universal Dependencies
What it is and what it does
udtools provides the official validator and scoring tools for Universal Dependencies, a cross-linguistic annotation framework. It reads CoNLL-U formatted files and checks compliance with UD specifications across multiple validity levels, from basic file structure to detailed linguistic annotation rules. The package also includes the official parsing scorer used in CoNLL and IWPT shared tasks.
The validator can be invoked from the command line or imported as a Python library. When used programmatically, it returns a State object containing validation results that can be inspected for errors and warnings, optionally suppressing console output. The validator supports language-specific checks and can load custom data files from a local UD tools repository, though the pip-installed version may lag behind the authoritative specification.
Use it for:
- Validate CoNLL-U treebank files before submission to Universal Dependencies releases or shared task competitions.
- Integrate validation into a linguistic annotation pipeline to catch format and specification violations early.
- Score dependency parsing system outputs against gold-standard UD treebanks in research or evaluation workflows.
- Check compliance of custom language-specific UD annotations with the framework's rules and feature inventories.
- Batch-validate multiple treebank files and collect structured error reports for quality assurance.
Worth the install?
AI-flagged interpretation of the facts on this page — verify before relying
Validates CoNLL-U format files against Universal Dependencies specifications and scores dependency parsing results from shared tasks.
Yes, if you work with Universal Dependencies data or need to validate CoNLL-U files. The package is actively maintained, has low install friction, and provides the official validator and scorer. The copyleft license is a consideration for proprietary work. Note that pip-installed versions may not have the most current language-specific data; clone the authoritative repository if you need cutting-edge specifications.
Install
udtools on PyPI
pip
pip install udtoolsuv
uv add udtoolspoetry
poetry add udtoolsInstalling udtools
Before you install
Low install friction with only two runtime dependencies (udapi and regex). Active maintenance with a recent release 57 days ago and ongoing repository activity.
License in practice
GPL-2.0-or-later copyleft license means any derivative work or distribution must also be open-source under a compatible license; suitable for research and open-source projects but requires careful consideration in proprietary contexts.
Quickstart
pip install udtools
from udtools import Validator
validator = Validator(lang='la', output=None)
state = validator.validate_files(['la_proiel-ud-train.conllu'])
if state:
print('Validation passed')
else:
print('Validation failed')
Requires Python 3.10 or later.
Verify before relying
- Whether language-specific data files bundled with pip releases stay current with the authoritative UD specification
- Performance characteristics when validating very large treebanks or many files in sequence
Package facts
| License | GPL-2.0-or-later (copyleft) |
| Python support | supports the current Python release (>=3.10) |
| Install friction | low — pure-Python wheel |
| Runtime dependencies | 2 — udapi, regex |
| Maintenance | actively maintained — 57 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 251,021/month — #8,602 on PyPI (30-day window, as of 2026-08-14) |
| Known vulnerabilities | none known (OSV.dev, checked 2026-08-14) |
Evidence: udtools-0.2.8-py3-none-any.whl
Tags
More Linguistic packages
Detects and normalizes text encoding from…
permissive · top 100 on PyPI
tiktokentiktoken is a fast BPE tokenizer that converts…
permissive · top 1,000 on PyPI
chardetDetects character encoding and language in byte…
permissive · top 1,000 on PyPI
text-unidecodeConverts Unicode text to ASCII by…
copyleft · top 1,000 on PyPI
larkLark is a parsing library that builds abstract…
permissive · top 1,000 on PyPI
tree-sitterPython bindings to the tree-sitter parsing…
permissive · top 1,000 on PyPI
udapiUdapi is a Python framework for reading,…
copyleft · top 15,000 on PyPI
conlluParses CoNLL-U formatted text (a standard NLP…
unclear · top 15,000 on PyPI
confuConfu validates and generates configuration…
permissive · top 15,000 on PyPI
ginzaGiNZA is a Japanese NLP library that performs…
permissive · top 15,000 on PyPI
lemminflectLemmatizes and inflects English words using…
permissive · top 15,000 on PyPI
swagger-spec-validatorValidates Swagger/OpenAPI specifications…
permissive · top 15,000 on PyPI
stix2-validatorValidates STIX 2.x JSON documents against the…
permissive · top 15,000 on PyPI
iregexp-checkValidates regular expressions against RFC 9485…
permissive · top 15,000 on PyPI
simplemmaSimplemma converts inflected word forms to…
permissive · top 15,000 on PyPI
stanzaStanza is a Python NLP library that runs…
permissive · top 5,000 on PyPI