Hierarch

What's Changed
* Bump pytest from 7.3.1 to 7.3.2 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/155
* Bump actions/checkout from 3.5.2 to 3.5.3 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/154
* Bump sphinx-rtd-theme from 1.2.1 to 1.2.2 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/153
* Bump pandas from 2.0.1 to 2.0.2 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/152
* Bump numba from 0.57.0 to 0.57.1 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/156
* Bump pytest from 7.3.2 to 7.4.0 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/157
* [FEAT] Adds `hierarch.stats.hierarchical_randomization`, a generator that yields resampled datasets by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/164
* Bump sphinx-rtd-theme from 1.2.2 to 1.3.0 by dependabot in https://github.com/rishi-kulkarni/hierarch/pull/163


**Full Changelog**: https://github.com/rishi-kulkarni/hierarch/compare/v1.1.4...v1.1.6

This one mostly fixes Numba 0.57 compatibility.

What's Changed
* Development by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/34
* New project configuration and minor updates by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/69
* Add CI and update docs by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/70
* fix action by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/71
* [FIX] Numba 0.57 compatibility by rishi-kulkarni in https://github.com/rishi-kulkarni/hierarch/pull/151


**Full Changelog**: https://github.com/rishi-kulkarni/hierarch/compare/v1.1.3...v1.1.4

Changelog
------------
* Improved initial compile time
* If the input data is a pandas DataFrame, the treatment column can be specified using the column name as a string.

Changelog
* Improved convergence rate of confidence_interval function
* Added tests for a number of workhorse functions

Changelog
* added docstrings and minor bugfixes
* prevented hypothesis_test from giving a p-value of 0 when there are no permutations that can be performed

Changelog

* Combined two_sample_test and linear_regression test to a single hypothesis_test function - setting compare to "means" calls two_sample_test, while setting compare to "corr" calls linear_regression_test
* confidence_interval now searches for the upper and lower bounds separately, which improves coverage significantly for skewed data
* General speed improvements