Fix Issue 162 - SVD with empty rows and cols

[lkdvos]

Fixes #162

This more properly handles some edge cases that had been reported, by also migrating fully empty row/col pairs to the end, and correctly adapting the indexing.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 75.84%. Comparing base (83ed436) to head (9426dc8).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #164      +/-   ##
==========================================
- Coverage   75.95%   75.84%   -0.12%     
==========================================
  Files          36       36              
  Lines        2059     2049      -10     
==========================================
- Hits         1564     1554      -10     
  Misses        495      495              

Flag	Coverage Δ
docs	7.28% <0.00%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[lkdvos]

@mtfishman I'm not sure if I can actually fix this in such a way that it stays compatible with gradedarrays (also, I don't think it is great that GradedArrays depend so heavily on the internal implementation here).
The issue is, from what I can tell, that for blocksparsearrays, when a block is missing there's not much more for me to do than simply migrate it to the end, otherwise I get into trouble with rectangular shapes (for example the bug here). This does break some of the things that GradedArrays was implicitly depending on. In some sense, I would argue that this is not a breaking change, and it is mostly the GradedArrays implementation that was too dependent on internals here.