Tool to find the distribution of a function of a R.V.

[Mv77]

Imagine x is a DiscreteDistribution object and you want to find the distribution of f(x) for some f.

This pr introduces a tool called distr_of_function. distr_of_function(x, f) will return a DiscreteDistribution object that represents the distribution of f(x). I wrote tests and made it work for f: 1->1, f: n->1, f: 1->n, and f: n->m. You only need f to return a np.array.

Please ensure your pull request adheres to the following guidelines:

• Tests for new functionality/models or Tests to reproduce the bug-fix in code.
• Updated documentation of features that add new functionality.
• Update CHANGELOG.md with major/minor changes.

[Mv77]

I see this being useful as we move toward the representation of models using transition equations.

Imagine I'm in period t and have:

• A function for v[t+1] and its derivatives
• A distribution for t+1 shocks ShockDstn.
• A transition equation such that state[t+1] = transition(shocks, state[t]).

One could do something like FutureStateDstn = distr_of_function(ShockDstn, transition, state[t]). And then with calc_expectations compute the things we usually need, say, calc_expectations(FutureStateDstn, dvdm[t+1]).

[Mv77]

@sbenthall could you review? This is heavily inspired on your calc_expectations tool, which I am a huge fan of.

[codecov-commenter]

Codecov Report

Merging #1144 (21c6e21) into master (0bec58f) will increase coverage by 0.07%.
The diff coverage is 100.00%.

❗ Current head 21c6e21 differs from pull request most recent head b9a4377. Consider uploading reports for the commit b9a4377 to get more accurate results

@@            Coverage Diff             @@
##           master    #1144      +/-   ##
==========================================
+ Coverage   73.99%   74.07%   +0.07%     
==========================================
  Files          70       70              
  Lines       10776    10809      +33     
==========================================
+ Hits         7974     8007      +33     
  Misses       2802     2802              

Impacted Files	Coverage Δ
HARK/distribution.py	85.39% <100.00%> (+0.30%)	⬆️
HARK/tests/test_distribution.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0bec58f...b9a4377. Read the comment docs.

[sbenthall]

Yes, I can review this.

[sbenthall]

Looks great! Merging now...