Polars RefKit

Parse, inspect, render, and format BibTeX columns inside Polars eager and lazy queries. Each source row owns an independent bibliography and citation state.

Documentation · Expression reference · Python objects

Render a citation column

Install in Python 3.10 through 3.14 with Polars 1.29 or newer:

python -m pip install polars-refkit

import polars as pl
import polars_refkit

frame = pl.DataFrame(
    {
        "bibtex": ["@article{doe2024, author={Doe, Jane}, title={Fast Citations}, year={2024}}"],
        "key": ["doe2024"],
    }
)
result = frame.select(
    citation=pl.col("bibtex").refkit.cite("key"),
    entries=pl.col("bibtex").refkit.entry_count(),
)
print(result.to_dicts())

[{'citation': '(Doe, 2024)', 'entries': 1}]

Importing polars_refkit registers pl.Expr.refkit. A string argument names a column. Use pl.lit(...) for literal bibliography text or citation keys.

Choose an operation

Choose output="text", "html", or "rendered" on rendering expressions. Pass formatting settings as an options= dictionary. Value expressions return null on row-local failures. Reports retain structured diagnostics, errors, or rename records beside the source row.

Use with agents

The package includes version-matched task guidance:

import polars_refkit.agent

help(polars_refkit.agent)

Agent integration covers discovery. Data shapes defines reports.

polars-refkit supports native CPython on Linux, macOS, and Windows. For Python in Pyodide, use refkit directly. The Polars adapter does not support Pyodide.

RefKit is alpha software. Public APIs may change before 1.0.

Apache-2.0 license