[NDTensors] Unwrap module for dispatching on unwrapped types

[kmp5VT]

Description

In this PR I attempt to address issues we are having with generic programming. This issue is that wrapped arrays can become too complicated for Julia to parse correctly for dispatch and as a result, incorrect functions are called which in turn slows down the code. Here I am creating a simple Expose type which will determine the base unwrapped type and launch the appropriate kernels based on this type.

This will fix #1193 (we should confirm that before merging).

Checklist:

• Create Expose type and function to create type.
• Replace NDTensors code where we implicitly unwrap code (through parent_leaftype) and instead dispatch on the expose struct
• Ensure that all tests still pass in NDTensors and ITensors
• Fix scalar indexing issues in GPU SVD code
• address merge conflicts

[codecov-commenter]

Codecov Report

Attention: 2 lines in your changes are missing coverage. Please review.

Comparison is base (70967cd) 82.87% compared to head (356c451) 54.54%.

❗ Current head 356c451 differs from pull request most recent head 130d4f7. Consider uploading reports for the commit 130d4f7 to get more accurate results

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1220       +/-   ##
===========================================
- Coverage   82.87%   54.54%   -28.34%     
===========================================
  Files          89       88        -1     
  Lines        8440     8392       -48     
===========================================
- Hits         6995     4577     -2418     
- Misses       1445     3815     +2370     

Files	Coverage Δ
src/mps/dmrg.jl	0.00% <0.00%> (-79.11%)	⬇️
src/set_types.jl	0.00% <0.00%> (ø)

... and 36 files with indirect coverage changes

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

[mtfishman]

@kmp5VT I fixed the merge conflicts.

This looks good to go from my end, is this ready from your end? Any more scalar indexing issues you wanted to address?

[mtfishman]

Ah looks like some tests are failing, I introduced code into main that uses leaf_parenttype and parenttype which needs to get updated.

[kmp5VT]

Ah looks like some tests are failing, I introduced code into main that uses leaf_parenttype and parenttype which needs to get updated.

Yeah I found those areas and fixed them

[kmp5VT]

@mtfishman Okay this PR should be good now. Thanks!