MDAnalysis / GridDataFormats

GridDataFormats is a pure Python library to handle data on a regular grid using commonly used file formats in molecular simulations.
https://mdanalysis.org/GridDataFormats
GNU Lesser General Public License v3.0
29 stars 18 forks source link

unit tests #6

Closed orbeckst closed 6 years ago

orbeckst commented 9 years ago

We should have some minimal unit tests for the package. They could be part of the package and wouldn't have to be separate. Look to use cStringIO to use tiny inline fixtures.

kain88-de commented 9 years ago

I started adding some unit tests. There is one small one to check dx files 8467c7bc9068.

It would be nice to know if this files covers all the information we potentially want to extract from a dx file.

TODO:

orbeckst commented 9 years ago

Thanks @kain88-de for all the work you've been putting into updating the package. You should be able to write to the repo now. For bigger changes, do a feature branch/PR and assign to someone else like @richardjgowers or myself for review. By all means, ping me if nothing seems to be moving forward.

For the tests we need small example files. I should be able to get a small plt file with gridcount but I have no ccp4 one, maybe @holocronweaver has one?

holocronweaver commented 9 years ago

Sure, I can make one for testing. Do you have a particular 3D data set you want to fill it with?

orbeckst commented 9 years ago

On 11 Sep, 2015, at 14:21, Jesse Johnson wrote:

Sure, I can make one for testing. Do you have a particular 3D data set you want to fill it with?

I don't have one at hand. The smaller the better so that we can include it as test data.

kain88-de commented 9 years ago

@holocronweaver have a look at the dx test file. Something like this is enough. It should be small and contain all meta-data you find important.

For the data I choose to have 8 points (2 in every direction) each with a value of 1. That is enough to see if the reading works.

orbeckst commented 9 years ago

On 11 Sep, 2015, at 14:46, kain88-de wrote:

@holocronweaver have a look at the dx test file. Something like this is enough. It should be as small and contain all meta-data you find important.

For the data I choose to have 8 points (2 in every direction) each with a value of 1. That is enough to see if the reading works.

Actually, less symmetry is better, something like 2 x 3 x 5 because this can help with catching errors of swapping axes.

holocronweaver commented 9 years ago

Here is a test CCP4 file: https://app.box.com/s/d3l9boh11x7q0bjtskskj003v09uw92z It is 8 x 4 x 6 and the grid points have values equal to their column-major index in a 1D array. Hope it is sufficient for unit testing. (BTW, creating this file uncovered a major bug in my software! Good thing it hasn't been released to the public yet. Sometimes contributing to OSS has unexpected benefits.)

orbeckst commented 6 years ago

Not sure why I didn't close this one. We have had CI with tests for a long time...