tackling duplication in storing the information about the runs

[segasai]

Following the issue #244 and required changes, I've decided to finally get rid of error-prone way of storing the information about the saved/based/new runs. Now there is a common object that stores various parameters. This makes a lot of logic simpler, getting rid of hundreds of lines of code.

Since the test suite is now in better shape, I think most of the bugs were caught by the tests, so I'm pretty confident about the patch. The patch is quite big, but most of the changes are pretty straightforward.

(since when I edit the code it automatically goes through yapf, there are some white-space changes as well).

There are some stylistic improvements that can be further done, like reusing RunRecord from sampler in dynamicsampler. It's also probably worth putting additional checks on the sizes of various arrays. i.e. assert that lengths of various arrays in run records are the same, but it's good enough IMO

[segasai]

All the regression tests pass now, so it's ready for review

[coveralls]

Pull Request Test Coverage Report for Build 817711672

• 219 of 262 (83.59%) changed or added relevant lines in 3 files are covered.
• 4 unchanged lines in 2 files lost coverage.
• Overall coverage decreased (-1.1%) to 70.348%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
py/dynesty/sampler.py	68	78	87.18%
py/dynesty/dynamicsampler.py	139	154	90.26%
py/dynesty/utils.py	12	30	40.0%
<!--	Total:	219	262	83.59%	-->

Files with Coverage Reduction	New Missed Lines	%
py/dynesty/sampler.py	1	81.24%
py/dynesty/utils.py	3	41.47%
<!--	Total:	4		-->

Totals
Change from base Build 814234237:	-1.1%
Covered Lines:	3293
Relevant Lines:	4681

---

💛 - Coveralls

[joshspeagle]

This is great! I've just finished going through and the internal refactoring is so much cleaner now, both for tracking things around and for possibly adding new quantities down the line. Happy to merge this in.

[avivajpeyi]

Hi @joshspeagle,

As this causes the way results are stored to change from the most up-to-date release on PyPI (v1.1), would it be possible to release a new version of dynesty on PyPI?

[joshspeagle]

There are currently a few more active issues being resolved, but the plan is to push a new version once those wrap up in the near-term future.