Add feature to automatically rename categorical variables in etable

[dsliwka]

This adds a feature to etable that automatically renames categorical variables in regression tables. The user can specify a formatting template (default is "{variable}={value}"). It should work either for pandas categoricals (i.e "y ~ x + c" when c is categorical) and with categorical variables generated in the regression formula with the C() or i() operator.

To do this, the existing rename_categoricals function is slightly extended to allow using formatting templates.

[codecov]

Codecov Report

Attention: Patch coverage is 96.42857% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
pyfixest/report/utils.py	95.00%	1 Missing ⚠️

Flag	Coverage Δ
core-tests	80.89% <96.42%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
pyfixest/report/summarize.py	87.42% <100.00%> (-0.17%)	⬇️
pyfixest/report/utils.py	97.29% <95.00%> (+6.97%)	⬆️

... and 1 file with indirect coverage changes

[s3alfisc]

pre-commit.ci autofix

[s3alfisc]

pre-commit.ci autofix

[dsliwka]

Hi @s3alfisc,

thanks for reviewing this PR!

I just realized that I forgot to mention that when using "i(x, ref=v)" then having blanks in category (i.e. level) names leads to an error. That is, i(gender, ref="Male") works well but i(ethnicity, ref="Native American") leads to "ValueError: Value 'NativeAmerican' for TreatmentContrasts.base is not among the provided levels." even though "Native American" is in the levels of the categorical column. As it is often convenient to have category names including blanks, I had added set_first_cat as a workaround. But of course I understand that is independent of pyfixest functionality so it makes sense to drop it.
But would it be possible to keep blanks in level names in parsing?

[s3alfisc]

"i(x, ref=v)" then having blanks in category (i.e. level) names leads to an error.

Ah ok, I see! I think we should treat this as a bug of the i() parser and I might open a separate issue for it?

Btw, I have also changed the default to not apply the relabeling by default - the reason is that I think there should be consistency in internal variable naming and users should explicitly ask for renaming. Else we might run into users who see variables names in etable() but struggle to understand how they might relate to what is included in .coef(), .tidy() or .summary()?

[dsliwka]

Yes, that all makes sense, thanks! From: Alexander Fischer ***@***.***> Sent: Friday, February 7, 2025 4:28 PM To: py-econometrics/pyfixest ***@***.***> Cc: KT Sliwka Dirk ***@***.***>; Author ***@***.***> Subject: Re: [py-econometrics/pyfixest] Add feature to automatically rename categorical variables in etable (PR #798) "i(x, ref=v)" then having blanks in category (i.e. level) names leads to an error. Ah ok, I see! I think we should treat this as a bug of the i() parser and I might open a separate issue for it? Btw, I have also changed the default to not apply the relabeling by default - the reason is that I think there should be consistency in internal variable naming and users should explicitly ask for renaming. Else we might run into users who see variables names in etable() but struggle to understand how they might relate to what is included in .coef(), .tidy() or .summary()? — Reply to this email directly, view it on GitHub<#798 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALY442WHAZPDX6OV6XLPGST2OTGJRAVCNFSM6AAAAABV3P4TVGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNBTGI2DONRSGA>. You are receiving this because you authored the thread.Message ID: ***@***.******@***.***>>

[s3alfisc]

Merged this one for now, thanks Dirk! On the i-ref and white space issue - let's see if I will manage to fix it, if not happy to discuss bringing back the set_first_cat function =)

[dsliwka]

Thanks!! From: Alexander Fischer ***@***.***> Sent: Friday, February 7, 2025 4:36 PM To: py-econometrics/pyfixest ***@***.***> Cc: KT Sliwka Dirk ***@***.***>; Author ***@***.***> Subject: Re: [py-econometrics/pyfixest] Add feature to automatically rename categorical variables in etable (PR #798) Merged this one for now, thanks Dirk! On the i-ref and white space issue - let's see if I will manage to fix it, if not happy to discuss bringing back the set_first_cat function =) — Reply to this email directly, view it on GitHub<#798 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALY442R4DPY2TZY2YFWDNC32OTHFBAVCNFSM6AAAAABV3P4TVGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNBTGI3DIMRYGE>. You are receiving this because you authored the thread.Message ID: ***@***.******@***.***>>