Add section on how to test models

[alexpghayes]

Things that should be included:

• How to test a formula / recipe interface
• How to test code that compares results to computation from other packages, or software written in other languages (i.e. suppose you want to compare your R implementation to someone's reference implementation in MATLAB)
• How to test models that are expensive to compute / will time out on Travis (i.e. saving models)

Some of this should overlap with the R package book (see related issues https://github.com/hadley/r-pkgs/issues/477, https://github.com/hadley/r-pkgs/issues/481).

[alexpghayes]

In particular we may want to provide advice on how to test against the Python ecosystem since it's such a nightmare to get that all set up

[DavisVaughan]

For yardstick, I took the approach of saving small output files for later use in tests.

One file generates them, and they just sit in the tests/ folder: https://github.com/tidymodels/yardstick/tree/master/tests/pycompare

I needed a helper for reading in the files so the same path could work interactively and during a CTRL+SHIFT+T test run (testthat changes the path). https://github.com/tidymodels/yardstick/blob/master/tests/testthat/helper-read_pydata.R

And then the actual tests are fairly simple: https://github.com/tidymodels/yardstick/blob/master/tests/testthat/test-class-precision.R#L24