Pandasaurus

Pandasaurus supports simple queries over ontology annotations in dataframes, powered by Ubergraph SPARQL queries. It keeps dependencies light while still offering CURIE validation, enrichment utilities, and graph exports for downstream tooling.

Features

• Validate and update seed CURIEs, catching obsoleted terms with replacement suggestions.
• Enrich seed lists via simple, minimal, full, contextual, and ancestor-based strategies.
• Build tabular outputs (pandas.DataFrame) and transitive-reduced graphs (rdflib.Graph) for visualization.
• Batched SPARQL queries and deterministic tests with built-in mocking examples.

Installation

pip install pandasaurus

or with Poetry:

poetry add pandasaurus

Requires Python 3.9–3.11.

Quick Example

from pandasaurus.curie_validator import CurieValidator
from pandasaurus.query import Query

seeds = ["CL:0000084", "CL:0000787", "CL:0000636"]

terms = CurieValidator.construct_term_list(seeds)
CurieValidator.get_validation_report(terms)  # raises if invalid or obsoleted

query = Query(seeds, force_fail=True)
df = query.simple_enrichment()
print(df.head())

See the Quick Start guide for a step-by-step workflow.

Documentation

Full documentation (quick start, recipes, developer guide, and API reference) lives under docs/ and is published from the gh-pages branch:

• Hosted documentation
• Quick Start (source)
• Guides (source)
• API reference (source)

To build docs locally:

poetry install -E docs
poetry run sphinx-build -b html docs docs/_build/html

Contributing

Pull requests are welcome! See docs/guides/contributing.rst for details on environment setup, testing, linting, and the release workflow. Pandasaurus aims to remain a small, focused library; please open an issue before introducing large new features.

Background

The first planned use case is to provide enrichment/query tooling for AnnData Cell x Gene matrices following the CZ single cell curation standard.