Improve LoopCombinator

[LanderOtto]

This commit improves LoopCombinator performance by skipping the possibility to combine tokens with different tag levels (a state that should never occur in loops). Most importantly, this change also resolves potential errors in recovery workflows where input tokens may arrive out of order. Before this commit, during recovery, an output token could have the correct tag while its associated input tokens were incorrect due to tag mismatches (e.g., a token with tag 0.0 was generated using input tokens with tags [0, 0.0, 0]).

Moreover, the restore method of the LoopCombinator and LoopCombinatorStep classes was refactored.

[codecov]

Codecov Report

❌ Patch coverage is 92.30769% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 73.69%. Comparing base (3653a82) to head (465b58e).
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
streamflow/workflow/step.py	88.23%	0 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #952      +/-   ##
==========================================
+ Coverage   73.66%   73.69%   +0.02%     
==========================================
  Files          90       90              
  Lines       12223    12224       +1     
  Branches     2161     2160       -1     
==========================================
+ Hits         9004     9008       +4     
+ Misses       2704     2702       -2     
+ Partials      515      514       -1     

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