Resolve some issues when using custom sampler

[ColmTalbot]

I ran into some issues trying to use a customised Sampler subclass and with a different proposal distribution.

The changes in this PR allow me to fairly simple define the new class and function to modify the analysis.

[ColmTalbot]

Ping @joshspeagle @segasai

[segasai]

Thanks for adding a new test. that was previously lacking, because I never used this functionality
One quick comment is that I think it would be better to make the test not part of test_egg because here IMO we don't particularly care about the complex function to sample, we just care that the code can run (i.e. a single gaussian like in test_gau would do).. Also I'm not sure the test needs to be part of the itertools.product over bounds too (i.e. custom+rslice and multi+sample_rwalk combinations would do).
The reason for these comments is that the test execution time is a concern, so one should try to make tests as short as possible.

[coveralls]

Pull Request Test Coverage Report for Build 3191988185

• 6 of 6 (100.0%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.1%) to 91.472%

---

Totals
Change from base Build 3185506400:	0.1%
Covered Lines:	3990
Relevant Lines:	4362

---

💛 - Coveralls

[segasai]

thank for the PR.
I've merged the edited version of it in #402