WIP: Add function to irregularly rebin data in N dimensions

[DanRyanIrish]

To Do:

• [x] Make function accept an NDCube instance, strip the data and mask, and return the rebinned data and mask arrays.
• [x] Make function propagate uncertainties
• [x] Interpolate all coordinate grids
• [ ] Construct new WCS using gwcs and ndcube TableCoordinates from adapted coordinate values
• [x] Think how to handle axis-aware sliceable metadata
• [ ] Make function return same NDCube type as input cube.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 16.78832% with 114 lines in your changes are missing coverage. Please review.

Project coverage is 26.65%. Comparing base (d44a1c8) to head (bd428a0).

Files	Patch %	Lines
stixpy/coordinates/transforms.py	18.29%	67 Missing :warning:
stixpy/calibration/visibility.py	6.06%	31 Missing :warning:
stixpy/map/stix.py	0.00%	7 Missing :warning:
stixpy/product/sources/science.py	37.50%	5 Missing :warning:
stixpy/calibration/energy.py	33.33%	2 Missing :warning:
stixpy/io/readers.py	33.33%	2 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## ndcubify #109 +/- ## ============================================= - Coverage 66.33% 26.65% -39.69% ============================================= Files 31 31 Lines 1821 1887 +66 ============================================= - Hits 1208 503 -705 - Misses 613 1384 +771 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[DanRyanIrish]

Regarding the top-level checkpoint on handling sliceable metadata, once this PR is merged into ndcube and released, we could use ndcube.Meta and VisMeta included in the xrayvision Visibility PR as parent classes for a STIXVisMeta class. We could then implement a STIXVisMeta.rebin method that does a certain number of general, but well-defined things to the visibilities metadata on rebinning, e.g. adding the detector string labels together so we can see which detectors are included in a rebinned bin.

[DanRyanIrish]

@samaloney Could you merge/backport #106 and #107 into the TCDSolar ndcubify branch to make the diff more readable?