Add dirichlet distribution

[reubenharry]

Implements issue #1998.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[reubenharry]

I've marked this as a draft because there are currently no tests, but suggestions are welcome.

[SiegeLordEx]

Thanks! It looks mostly good (I added comments); there's many formatting issues, but I'll take care of them on our end when we import change.

[reubenharry]

Thanks for the comments - I think all addressed!

A few points:

• I didn't add a test running HMC on the distribution (although I have done that using my own code), mainly because I'm not very familiar with the tensorflow probability sampling code, and I struggled to debug it.
• I'd like to add a sample transformation for lambda x: x*2, but haven't yet found the closed form fourth moment of the Dirichlet (although I presume it exists).
• I'm happy to run autoformatting stuff for future PRs of this nature, but wasn't sure what tools you used