(Closes #2360) Fixes for LFRic dependency analysis

[arporter]

A couple of small fixes identified when doing OpenACC for LFRic.

[codecov]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (d90d03d) 99.84% compared to head (4f7e2f8) 99.84%.
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2361      +/-   ##
==========================================
- Coverage   99.84%   99.84%   -0.01%     
==========================================
  Files         348      348              
  Lines       47062    47073      +11     
==========================================
+ Hits        46990    47000      +10     
- Misses         72       73       +1     

Files	Coverage Δ
src/psyclone/dynamo0p3.py	99.78% <100.00%> (-0.03%)	⬇️
...clone/psyir/transformations/parallel_loop_trans.py	100.00% <100.00%> (ø)
src/psyclone/transformations.py	99.40% <100.00%> (+<0.01%)	⬆️

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

[arporter]

Small fixes that change the way DA is done for LFRic loops. These were needed to do ACC PARALLEL/LOOP for LFRic.

[sergisiso]

@arporter it looks good, I just have some general questions about how these fits with other LFRic transformations/kernels. But feel free to disregard if it leads to complications.

[arporter]

CI permitting, this is ready for another look.

[sergisiso]

@arporter Thanks for expanding the comments, the rest of the PR looks good. There are no TODOs remaining, linters don't show new issues, test and coverage pass. I am running integration but unless there is some error there, this is ready to merge.