Linearmodels

This release fixes a single bug that affected the computation of the first-stage F-statistic.

This is a minor release that fixes a wheel-building issue. There are no other changes from 4.18.

This is a performance and bug-fix release.

* Improved performance of `AbsorbingLS.fit` by deferring some operations.
* Added support for the method available in PyHDFE in `AbsorbingLS`. These methods can only be used when the variables absorbed are categorical (i.e., fixed-effects only) and when the model is unweighted.
* Added alternative measures of R-squared using the squared correlation. See `PanelResults.corr_squared_overall`, `PanelResults.corr_squared_within`, and `PanelResults.corr_squared_between`.
* Added a clustered covariance estimator (`ClusteredCovariance`) for system regressions.
* Fixed a bug in `kernel_optimal_bandwidth` which used incorrect values for a tuning parameter in the bandwidth estimation for the Parzen and Quadratic Spectral kernels.

This is a minor release that corrects multiple incorrect type definitions.

This release being two features and some bug fixes:

* Added FamaMacBethResults which exposes the set of estimated regression coefficients used to construct the model estimator.
* Added typing to all functions
* Refactored code to statically type with mypy and fixed some subtle bugs

This release fixes a small number of bugs, adds new measure of model fit, and provides some essential maintenance:

* Blackened the code.
* Added McElroy's and Berndt's measures of system fit.
* Removed support for Python 3.5 inline with NEP-29.
* Fixed a bad test that only appeared when packaging from 4.15.3.