Support Polars

[vincentarelbundock]

Support Polars

[s3alfisc]

Yes, thanks for the suggestion. I had not implemented it yet because I thought it would require to make polars a hard dependency, but as you correctly suggest I can solve this with a try-except statement. I'll implement this tomorrow!

[vincentarelbundock]

Yay!

[s3alfisc]

Hi @vincentarelbundock , polars should now be supported for feols() and fepois() data input arg =)

from pyfixest.estimation import feols, fepois
from pyfixest.utils import get_data
import polars as pl

data = get_data()
data_pl = pl.from_pandas(data)
fit = feols("Y ~ X1", data=data)

fit.tidy()

#	Estimate	Std. Error	t value	Pr(>|t|)	2.5 %	97.5 %
# Coefficient						
# Intercept	2.230321	0.088980	25.065381	0.000000	2.055711	2.404931
# X1	0.333686	0.069076	4.830700	0.000002	0.198135	0.469238

I have a few other bug fixes on the way, hopefully I can publish a new version to PyPi by tomorrow.

[vincentarelbundock]

love it.

[vincentarelbundock]

sorry, one question since I can't test now: This works in fit.predict() too?

[s3alfisc]

Ah, no; I'll update this quickly.

[s3alfisc]

Support for predict() is added with #241 =)