Fixed: Resample requesting outside available ROI for interpolation padding

[rhoadesScholar]

Pre-Merge Checklist:

• [ ] if this PR adds a feature: request merge into the latest development branch (vX.Y-dev)
• [x] if this PR fixes a bug: request merge into the latest patch branch (patch-X.Y.Z)
• [x] PR branch name is short and describes the feature/bug addressed in this PR
• [x] commit messages are consistent
• [ ] changes reviewed by another contributor
• [x] tests cover changes
• [x] all tests pass

[pattonw]

Good catch. To avoid interpolation border artifacts it might be better to shrink the provided roi by 1 voxel in setup rather than doing an roi intersection which will still lead to border artifacts if the request happens to be at the boundary of what is provided. Also I'm not completely clear on the purpose of ndim, can you please update the docstring?

[rhoadesScholar]

To avoid interpolation border artifacts it might be better to shrink the provided roi by 1 voxel in setup rather than doing an roi intersection which will still lead to border artifacts if the request happens to be at the boundary of what is provided.

The downside of this is that the Resample node can no longer be placed downstream of a RandomLocation node, because RandomLocation provides an (inf, inf, inf) shaped ROI, which breaks the setup.

[rhoadesScholar]

To avoid interpolation border artifacts it might be better to shrink the provided roi by 1 voxel in setup rather than doing an roi intersection which will still lead to border artifacts if the request happens to be at the boundary of what is provided.

The downside of this is that the Resample node can no longer be placed downstream of a RandomLocation node, because RandomLocation provides an (inf, inf, inf) shaped ROI, which breaks the setup.

Fixed (d092313)

[pattonw]

Looks pretty good to me. If no one else has any comments I'll merge it tomorrow