Fix for expectations of very small numbers

Missed one line when updating min(A)-1
to min(A)-logfactor.  Affected the value of min(A)
coming out (usually not used).

Noticed one bug in setlogzero commit

* adding conda env and github workflow

* switch to pytest

* remove nose from setup.cfg

* switch to pytest in recipe

* everything is now switched to pytest, need to fix some issues with type errors now

* pin statsmodels since version 0.13 has breaking changes

* don't worry about linting or testing examples

* update badge

* change test to check if: absolute(fe_sigma - dfe_t[0,1]) <= (0.01 * absolute(dfe_t[0,1]))

* run on pymbar3 branch

* Starting bootstrap.

* Bootstrap sampling for free energies.

* Added test for bootstraps, fixed bug revealed by test.

* Boostrap pause.

* update with bootstrapping at each K.

* fixed float->int typing problem.

* fix test

* Bootstrap starting from the initial point.

* allow solver tolerance specification, skip sc iterations for bootstrapping, fix docs

* clean up solver_tolerance

* add test

* fix call to harmonic oscillators test case in test

* Make sure can return uncertainies without dictionaries.

Including in case with bootstrap.

Co-authored-by: Ivy Zhang <[email protected]>
Co-authored-by: Mike Henry <[email protected]>

Co-authored-by: Mike Henry <[email protected]>

* sphinx.ext.pngmath is now sphinx.ext.imgmath

* configure bibtex_bibfiles setting

* quick pass on the docs, plenty more to fix but will take care of it later

* add warning to the docs

* add warning to readme