Add examples from BUGS volume 2

[sunxd3]

Fix https://github.com/TuringLang/JuliaBUGS.jl/issues/133 (also #188, which seems to be a numerical thing, fixed by using approximately equality test)

[coveralls]

Pull Request Test Coverage Report for Build 9747173277

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 15 of 15 (100.0%) changed or added relevant lines in 1 file are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage increased (+0.05%) to 82.624%

---

Files with Coverage Reduction	New Missed Lines	%
src/parser/bugs_parser.jl	1	89.06%
<!--	Total:	1		-->

Totals
Change from base Build 9237944383:	0.05%
Covered Lines:	1436
Relevant Lines:	1738

---

💛 - Coveralls

[yebai]

@sunxd3 looks good; do we cover all BUGS models in Vol 1 and Vol 2?

[sunxd3]

Other than the last four examples from Vol 2, all examples from Vol 1 and 2 are here. I didn't include those examples because either they are not examples on doing inferences (https://www.multibugs.org/examples/latest/Stagnant.html) or they are not very illustrative.

I tried doing HMC (or NUTS) on pretty much all the examples, the results are here. Most examples actually just work. The biggest failure mode is discrete random variable, for instance

a ~ dcat() # Categorical
b ~ dnorm(c[a], 1)

the issue here is a is casted into a float (TrackedReal), but the conversion back to Int is not defined for TrackedReal). MH with individual proposal should work, although with LogDensityProblems interface, it is not straightforward to do.