Repository navigation
Use pytest-mpl - #168
Use pytest-mpl#168
Conversation
There was a problem hiding this comment.
As mentioned in https://github.com/astrofrog/pytest-mpl/issues/7#issuecomment-115407826, I think having to repeat baseline_dir='baseline_images' is not nice and it would be better to have this as a global option (just in case I didn't make myself clear there).
There was a problem hiding this comment.
There was a problem hiding this comment.
Ok, this should now be fixed in this PR using pytest-mpl v0.2
|
@astrofrog – Looks like all the image diffs are now above tolerance and all tests fail? This is just moving code to a different place, the results should be the same, no? |
|
The failures are only with the developer version of astropy. Looking into it. |
|
The failures are on master too, it's just that the tests don't get run regularly enough with astropy changes. |
|
I'm going to guess it's related to WCSLib 5 |
|
I guess the way to debug image difference failures is to blink the images and maybe compute and look at difference images? Could be useful to say a few sentences how to do this in |
|
This is a nice example. Expected Actual I can include these in the pytest-mpl docs. The bad commit in astropy according to bisect is: |
|
Finally managed to make a minimal example to reproduce the issue with Astropy: astropy/astropy#3891 |
|
@cdeil - added an example to the pytest-mpl README based on the above: https://github.com/astrofrog/pytest-mpl/blob/master/README.md#test-failure-example |
a42fb5e to
669037b
Compare


This will need the 0.2 release of pytest-mpl to work. Will update .travis.yml once 0.2 is available.
@cdeil, just pinging you so you can see it in use in a real package.