-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix tests and update examples #36
Conversation
When EigenData2._get_temperature_index had an input that equals the boundary, it had been issuing a warning. I changed this so that the warning is not issued when the temperature equals the minimum or maximum temperature in the grid. When I start a simulation at 10000 K, I no longer get a warning.
Hello @namurphy! Thanks for updating the PR.
Comment last updated on June 08, 2018 at 00:33 Hours UTC |
Non-Python files need to be included in MANIFEST.in so that they are kept track of while running `python setup.py install` and `python setup.py test`. This should recursively include everything in the data directory.
This snuck back in when I pulled in the latest changes from the astropy affiliated package template.
For many of the packages in requirements/requirements.txt, the minimum required version is probably more recent that is necessary for most of these. The exceptions are NumPy (which had some io changes as of 1.14) and Cython (which had some bugs for ~0.27 that showed up in PlasmaPy).
There have been problems with some NumPy and SciPy functions dropping units from Quantity instances. More recent versions of Astropy with NumPy 1.13 or 1.14ish can handle these special functions better. This will probably be important since users may use NumPy and SciPy functions in the functions that define density and temperature evolution.
For some reason I keep getting an error that the coveragerc file has not been found. Here I'm handling the coveragerc file (for code test coverage checking) in the same way as the HDF5 files.
Tests for test_eigenvaluetable.py were timing out in Travis CI because there was no screen output for 10 minutes. This now allows 30 minutes for no screen output. We will probably want to shorten the tests though, if possible.
I temporarily commented out a test that depended on ChiantiPy because it might be causes Travis CI to get stalled and time out. I also parametrized another test so that there would be more frequent screen output, and the tests would be performed independently of each other.
There was a bug in PlasmaPy where there was a TypeError raised when an int was expected and a numpy.dtype of something like int64 was used.
Codecov Report
@@ Coverage Diff @@
## master #36 +/- ##
=========================================
Coverage ? 67.95%
=========================================
Files ? 6
Lines ? 1267
Branches ? 0
=========================================
Hits ? 861
Misses ? 406
Partials ? 0
Continue to review full report at Codecov.
|
I also parametrized this to check a few different elements. For some reason, there have been some problems with Travis CI in creating EigenData2 instances for heavier elements (e.g., Fe).
I finally got the tests to pass again on Travis CI, though I'm still having some problems with |
...so that Travis CI doesn't time out.
I am going through and fixing tests, mostly in
nei.py
andeigenvaluetable.py
. I am also about to make a few changes to how things are set up for Travis CI, so I'll need to keep this pull request open to make sure the test settings are correct.