Glm demean

[s3alfisc]

Add support for GLMs with fixed effects.

import numpy as np
import pandas as pd
import pyfixest as pf

data = pf.get_data(N=500, seed=42)
data["Y_binary"] = np.where(data["Y"] > 0, 1, 0)
data = data.dropna()

fit_fe = pf.feglm("Y_binary ~ X1 + X2 | f1", data=data, family="logit", vcov="hetero", ssc = pf.ssc(k_adj = False, G_adj = False))
fit_C = pf.feglm("Y_binary ~ X1 + X2 + C(f1)", data=data, family="logit", vcov="hetero", ssc = pf.ssc(k_adj = False, G_adj = False))

pf.etable([fit_fe, fit_C], keep = ["X1", "X2"], digits = 6)

To be tackled in separate PRs:

• accelerations as described in the ppmlhdfe paper
• unification of the Fepois and GLM.Poisson classes
• refactoring of the weighing logic - separate weights and irls weights, never pre-multiply Y, X as this is not explicit / confusing

@Dpananos @leostimpfle FYI (have a great start into the new year!)

[codecov]

Codecov Report

❌ Patch coverage is 6.75676% with 69 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
pyfixest/estimation/feglm_.py	3.07%	63 Missing ⚠️
pyfixest/estimation/feols_.py	0.00%	3 Missing ⚠️
pyfixest/did/lpdid.py	0.00%	2 Missing ⚠️
pyfixest/estimation/FixestMulti_.py	0.00%	1 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (d83e15b) and HEAD (926ff4a). Click for more details.

HEAD has 3 uploads less than BASE

Flag	BASE (d83e15b)	HEAD (926ff4a)
core-tests	2	0
tests-extended	1	0

Flag	Coverage Δ
core-tests	?
tests-extended	?
tests-vs-r	16.90% <6.75%> (-0.06%)	⬇️

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

Files with missing lines	Coverage Δ
pyfixest/estimation/estimation.py	9.30% <ø> (-81.40%)	⬇️
pyfixest/estimation/fegaussian_.py	61.29% <100.00%> (-25.38%)	⬇️
pyfixest/estimation/felogit_.py	54.28% <100.00%> (-33.95%)	⬇️
pyfixest/estimation/feprobit_.py	50.00% <100.00%> (-40.25%)	⬇️
pyfixest/estimation/FixestMulti_.py	18.08% <0.00%> (-60.11%)	⬇️
pyfixest/did/lpdid.py	18.27% <0.00%> (-76.35%)	⬇️
pyfixest/estimation/feols_.py	7.85% <0.00%> (-83.54%)	⬇️
pyfixest/estimation/feglm_.py	22.22% <3.07%> (-56.95%)	⬇️

... and 35 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.