fix coalescence

[sajjadazimi]

Purpose

Make coalescence computations gpu compatible.

I have:

• followed the codebase contribution guide: https://clima.github.io/ClimateMachine.jl/latest/Contributing/
• followed the style guide: https://clima.github.io/ClimateMachine.jl/latest/DevDocs/CodeStyle/
• followed the documentation policy: https://github.com/CliMA/policies/wiki/Documentation-Policy
• checked that this PR does not duplicate an open PR.

In the Content, I have included

• relevant unit tests, and integration tests,
• appropriate docstrings on all functions, structs, and modules, and included relevant documentation.
• I have read and checked the items on the review checklist.

[edejong-caltech]

• Make moments a Tuple of Tuples (as in condensation/sedimentation tendencies) instead of SMatrix
• Use M from the new style of CoalescenceTensor to get the order of the moments needed
• In get_finite_2d_integrals, consider doing two maps: first, just compute the lower right triangle; second, copy those to the upper right triangle
• Maybe try to make the Q, R, S computations a bit more reasonable? Consider adding helper functions

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 96.49123% with 6 lines in your changes are missing coverage. Please review.

Project coverage is 97.67%. Comparing base (29b59bb) to head (2233170). Report is 12 commits behind head on main.

Files	Patch %	Lines
src/Sources/Coalescence.jl	97.38%	4 Missing :warning:
src/ParticleDistributions/ParticleDistributions.jl	0.00%	1 Missing :warning:
src/helper_functions.jl	91.66%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #161 +/- ## ========================================== - Coverage 98.17% 97.67% -0.51% ========================================== Files 8 9 +1 Lines 493 473 -20 ========================================== - Hits 484 462 -22 - Misses 9 11 +2 ```

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

[edejong-caltech]

Let's also check on the 3_modes and 4_modes test cases to make sure they are working as expected

[sajjadazimi]

Let's also check on the 3_modes and 4_modes test cases to make sure they are working as expected

There was a minor bug in the constructor of CoalescenceData: it used to normalize kernels twice. Now the results look correct again.