add {min,max}val2code transformations (towards #2105)

[rupertford]

Adds two new transformations which replace minval and maxval intrinsics with equivalent code. The requirements for these are similar to the existing sum2code transformation.

[codecov]

Codecov Report

Patch coverage: 100.00% and no project coverage change.

Comparison is base (f408b3b) 99.85% compared to head (8505d92) 99.85%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2148   +/-   ##
=======================================
  Coverage   99.85%   99.85%           
=======================================
  Files         339      342    +3     
  Lines       46069    46144   +75     
=======================================
+ Hits        46000    46075   +75     
  Misses         69       69           

Files Changed	Coverage Δ
src/psyclone/psyir/transformations/__init__.py	100.00% <100.00%> (ø)
...ir/transformations/intrinsics/maxval2code_trans.py	100.00% <100.00%> (ø)
...ir/transformations/intrinsics/minval2code_trans.py	100.00% <100.00%> (ø)
...psyir/transformations/intrinsics/mms_base_trans.py	100.00% <100.00%> (ø)
...psyir/transformations/intrinsics/sum2code_trans.py	100.00% <100.00%> (ø)

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

[rupertford]

This PR needs the HUGE and TINY intrinsics (or equivalent) to set initial values. PR #2130 has now added these so we can continue with this PR.

[rupertford]

Ready for first review from @arporter or @sergisiso

[arporter]

These are complicated transformations! I've only looked at maxval2code and the mms base class so far and have a lot of questions so I'm stopping here for now.
All the tests and examples are OK with compilation.
The updated (thanks) UG builds fine.

[rupertford]

Ready for next review from @arporter

[arporter]

Good job. It's a nice implementation now. Despite the number of comments this is nearly there.
I was about to say I'll fire off the integration tests but there's no point as they don't actually use these transformations. That should probably be a separate PR.
I'll do it next time though to check the compilation of the tests.
I've checked everything with pylint.
Ref. guide builds without new warnings.

[rupertford]

Ready for next review from @arporter

[arporter]

Very nearly there.
Integration tests are still underway.
I just have one outstanding query about the pair of tests that you say are doing different things...

[rupertford]

Ready for next review from @arporter

[arporter]

Requested change has been made.
Integration tests were all fine.
Ref. guide has no more warnings that on master.
Will proceed to merge.