rpick 0.3.0

Helps you pick items from a list by various algorithms. Example uses: pick a restaurant you haven't been to in a while, or an album to listen to.
Documentation
# 0.3.0

* [#5]https://gitlab.com/bowlofeggs/rpick/merge_requests/5: Introduced a unit test suite.
* [#8]https://gitlab.com/bowlofeggs/rpick/merge_requests/8: Defined a library for rpick so
  integrators can write their own front end to it. This also aided in testing.
* [#11]https://gitlab.com/bowlofeggs/rpick/merge_requests/11: Added documentation for the library.
* [#14]https://gitlab.com/bowlofeggs/rpick/merge_requests/14: Users will no longer be re-prompted
  for a choice they've declined in the same process, unless they decline all possible choices in a
  category.
* [d20e491b]https://gitlab.com/bowlofeggs/rpick/commit/d20e491b5971b73dd27d46bae3938f9321272517:
  Documented installation.


# 0.2.0

* [#3]https://gitlab.com/bowlofeggs/rpick/merge_requests/3: Added a new ```even``` distribution
  model, which does a nice flat random pick.
* [#4]https://gitlab.com/bowlofeggs/rpick/merge_requests/4: Added a new ```weighted```
  distribution model, which does a weighted random pick.
* [95b32b1e]https://gitlab.com/bowlofeggs/rpick/commit/95b32b1e4c103843cf3af900d94f5fef3ca286df:
  Added a new ```lottery``` distribution model, which gives lottery tickets to unpicked items and
  resets the picked item's lottery tickets to 0.


# 0.1.0

* [#1]https://gitlab.com/bowlofeggs/rpick/merge_requests/1: Added a new
  ```stddev_scaling_factor``` setting, which is optional and defaults to ```3.0```.
* [#2]https://gitlab.com/bowlofeggs/rpick/merge_requests/2: The model now defaults to "gaussian",
  so users don't have to define it by hand.