Percentage Point Functions for parameters

[vhaasteren]

This PR adds the Percentage Point Function (PPF) to Enterprise parameters. The PPF is the inverse of the CDF, so it can be used to sample from the distribution. Some Nested Sampling implementations require the PPF as input: for instance MultiNest requires the "hypercube transformation". That is done with the PPF.

Now we can do something like:

from enterprise.signals.parameter import LinearExp

le = LinearExp(1, 3)("LEXP")
le.get_ppf(0.5)

The Truncated Normal PPF is not yet implemented, but LinearExp, Normal, and Uniform are there.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.48%. Comparing base (5c1e455) to head (fbff5be).
Report is 16 commits behind head on dev.

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #353      +/-   ##
==========================================
- Coverage   88.44%   87.48%   -0.96%     
==========================================
  Files          13       13              
  Lines        3037     3133      +96     
==========================================
+ Hits         2686     2741      +55     
- Misses        351      392      +41     

Files	Coverage Δ
enterprise/signals/parameter.py	91.66% <100.00%> (+7.60%)	⬆️

... and 2 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8b1f24c...fbff5be. Read the comment docs.

[vhaasteren]

It seems that this requires some extra unit tests. Actually, the unit tests for the entire parameter class do not check for exceptions. I'll just update it to check for everything...

[AaronDJohnson]

@vhaasteren Is there anything that needs to be done on this one?

[vhaasteren]

No, is ready to go