python: fix handling of N/A output in image.statistics()

[jdtournier]

Discussed in #2248

I consider this a bug fix, but I suggest merging to dev given there is scope for this to change behaviour of existing apps. Up for discussion.

[Lestropie]

Would prefer to do the copyright message update in its own PR so that it doesn't corrupt the diff of unrelated changes. It looks like other stuff unrelated to the topic at hand has also made its way into here, likely due to a difference in base between the local branch and the PR, which would also make more sense in its own PR.

I consider this a bug fix, but I suggest merging to dev given there is scope for this to change behaviour of existing apps.

If the only "change of behaviour" is provision of a more interpretable error message where continuation of processing is not possible in any case, I don't think that merging to master is a problem.

[jdtournier]

Fair enough to split it from the changes in master. I'll clean this up once #2255 is merged.

I consider this a bug fix, but I suggest merging to dev given there is scope for this to change behaviour of existing apps.

If the only "change of behaviour" is provision of a more interpretable error message where continuation of processing is not possible in any case, I don't think that merging to master is a problem.

Well, in the case of dwi2response, I assume processing would stop in any case. But there may be other cases where a single-voxel mask would have caused a failure prior to these changes, but would complete successfully now. But if that's the case, I would agree that these would still nonetheless class as bug fixes. So I'll rebase it onto master as you suggest.