Generalize case where only 1 particle of each type is present

[shivupa]

Proposed changes

Closes #3137
This implements the fixes @markdewing recommended. It generalizes the special case of the two body jastrow where there are only 1 particle of each type present.

What type(s) of changes does this code introduce?

• Bugfix

Does this introduce a breaking change?

• No

What systems has this change been tested on?

Ryzen workstation (5900x, gcc version 10.2.0)

Checklist

Update the following with a yes where the items apply. If you're unsure about any of them, don't hesitate to ask. This is
simply a reminder of what we are going to look for before merging your code.

• Yes. This PR is up to date with current the current state of 'develop'
• Yes. Code added or changed in the PR has been clang-formatted
• No. This PR adds tests to cover any new code, or to catch a bug that is being fixed
• No. Documentation has been added (if appropriate)

[qmc-robot]

Can one of the admins verify this patch?

[prckent]

OK to test

[ye-luo]

Please add a unit test, The J2 in opt/qmcpack/src/QMCWaveFunctions/tests/test_bspline_jastrow.cpp can be a starting point.

[prckent]

Many thanks for this Shiv. FYI I ran the full (but limited) set of optimzation tests and they seemed to be OK.

I agree that adding a unit test by expanding the example Ye pointed to would be good.

If you don't have time to do that let us know.

[shivupa]

Yup sounds good. Thanks for the example Ye. I'll give it a try and report back

[markdewing]

The tests in test_DiffTwoBodyJastrowOrbital.cpp would be a good starting point, since they target the previous round of problems in this code.

[ye-luo]

@shivupa The added test doesn't trigger failure on develop. This then becomes a less effective test.
The test is named as "DiffTwoBodyJastrowOrbital" but there is no call to routines calculating parameter derivatives. If calling derivatives is the only way to trigger the failure, then adding derivative call is necessary. Otherwise, rename the title.
Please update the test.

[shivupa]

Thanks working on this now. I think I misunderstood what was going on and Mark's previous test. Sorry about that

[shivupa]

Sorry is it possible to stop checks on this? I though converting to draft would, but I this these will pass regardless as the previous work did. I mainly pushed to ask some questions.

[prckent]

We can't straightforwardly stop tests on a per-PR basis. However, we ignore the test results until the PR is marked as ready.

[shivupa]

Makes sense. I didn't want to waste the computational time.

This is ready for review now! Thanks @ye-luo and @markdewing for the discussion. The test now checks the F vector for any nullptrs. This fails on develop as expected (the F vector is private on develop so once the function allowing access for F is added on develop, this will still fail as expected).

[markdewing]

Looks good. Thanks for doing this!

[ye-luo]

Thank you!!!!