Details
-
Type:
Story
-
Status: To Do
-
Resolution: Unresolved
-
Fix Version/s: None
-
Component/s: utils
-
Labels:
Description
lsst.utils.assertFloatsAlmostEqual cannot handle lists of floats. Thus it cannot handle values returned from C++ as std::vector. Similarly for tuples of floats, lists of lists of floats, etc., though those may be rarer.
The usual workaround is to use self.assertTrue(numpy.allclose(arr1, arr2)) but it would be nice not to have to do that.
Also please consider adding a equal_nan flag like numpy.allclose. In both cases it is worth considering using numpy allclose as the code that performs the test.
Attachments
Issue Links
Activity
See afw/tests/test_endpoint.py and test_transform.py for examples where this would be useful.
Also, I observe that the first argument cannot be a list, but the second argument can be.
I'm not sure how important this is. The numpy.testing asserts seem to work fine. I would be interested to know why we provide our own assert.
I don't remember the details, but I added assertClose specifically because assert(numpy.allclose(...)) didn't do what I needed. I believe it was some combination of better diagnostics on failure and maybe some subtlety in defining relative tolerances.
See my comment in DM-7276 about the differences in behavior.
Poking this as I was just burned by the not quite complete behaviour on lists of floats. Might be a good one to deal with as a pair coding exercise.
Adding a ticket I filed earlier about using np.allclose. Also removed myself as the assignee: I'm not taking responsibility for this code, I just documented its availability!