Conversation
|
rebased against develop (in particular with #552 fixed). |
|
Rebased against develop and force-pushed. Added new requirement for GridDataFormats >= 0.3.1. |
|
No idea why, on travis, python appears to be unable to find GridDataFormats, even though it is clearly being installed (in version 0.3.1): from gridData import Gridfails in |
|
@orbeckst did you check locally of the gridData package can be loaded in a new virtualenv? |
|
rebased against develop |
|
Installing GridDataFormats into fresh virtualenv worked and all tests passed. |
|
rebased against develop |
|
Now part of the tests pass in build #1038:
I don't quite understand the failure, though, because the travis config always does a |
|
Yeah I also don't quite get the error. You could try to start python and import |
|
@kain88-de thanks for the idea: now I know why the minimal fails: GridDataFormat wants scipy. So I will decorate the tests. |
|
... or see if I can make scipy optional in GridDataFormats, i.e. only loaded if one wants to do the interpolations. |
|
Currently this is not possible. The ndimage function is already called at the initialization of the |
|
ah, crap... this fails at the top level as soon as Well, wait for MDAnalysis/GridDataFormats#25 ... |
|
Just released GridDataFormat 0.3.2: will update and rebase this branch and then it should work... |
- The new release 0.3.1 of GridDataFormats contains the fix for issue MDAnalysis/GridDataFormats#21 which is needed to fix issue #544. - The release 0.3.2 allows importing of gridData without scipy.ndimage present (MDAnalysis/GridDataFormats#25): this allows the minimal tests to pass
|
(I am a bit miffed that QC makes the PR look as if it does not pass the tests... most of the QC issues are things that I will simply ignore.) |
|
How I see it we have to tweak as we go. Any given test can be disabled, so go ahead and disable the ones you think won't apply to our style (or tell me which and I'll do it). As to the build "failing", I think it's harmless, and draws the attention of the PR reviewer to possible code issues. When we prune to leave only the tests we consider important a failure will be much more meaningful. |
|
I reviewed the QC issue and I will not fix any of the reported issues --- they all relate to how we write our tests and I don't think that there's anything wrong with it, @kain88-de, please consider the PR complete from my end (short of other problems you or anyone else might identify). |
test for writing DX files from density (Issue #544)
Requires a 0.3.1 release of GridDataFormats, see MDAnalysis/GridDataFormats#24.