How to Test PyNE

This guide will teach you the basics of how to test PyNE code.

Unit Testing

First, install:

To perform all unit tests:

$ cd tests/
$ ./travis-run-tests.sh

This will recursively look through the currently directory, open up every file named test_* and run every function (or method) named test_*.

Nosetests can take file(s) as an argument. For example, to just run the mcnp and material module tests:

nosetests test_mcnp.py test_material.py

A clean build/nucdatamake should yield a version of PyNE with all tests passing.

Example Testing

The examples directory should also be kept up-to-date as much as possible. PyNE examples are either in Python files or IPython notebooks. This means that to test the examples requires a recent version of IPython. Furthermore, the examples themseleves may have many other optional dependencies. Don’t be alarmed if testing the examples fails do to a lack of having a dependency installed. For this reason, testing the examples is not as important as unit tests, but still should be done occassionally.

To run the examples automatically, go to the examples directory and run the execer.py file from the root pyne dir.

$ cd examples
$ ../execer.py

Tutorial Testing

Tutorial testing is very similar to example testing except that all of the tutorials are IPython notebooks.

To run the tutorials automatically, go to the tutorial directory and run the execer.py file from the root pyne dir.

$ cd tutorial
$ ../execer.py

Putting It All Together

If you’d like to run all of the tests automatically from the root pyne dir, you can chain the following BASH commands together:

$ cd tests && ./travis-run-tests.sh && cd ../examples && ../execer.py || cd ../tutorial && \
  ../execer.py || cd ..

If you get any test errors try running nuc_data_make –clean=1 Happy testing!