fix(exchange and gnc package cellids)

[spaulins-usgs]

Fixed exchange and gnc package cellids to resolve model number properly, for simulations with multiple models (#1841)

[codecov]

Codecov Report

Merging #1866 (59ff92f) into develop (a644b52) will decrease coverage by 0.1%.
The diff coverage is 87.5%.

@@            Coverage Diff            @@
##           develop   #1866     +/-   ##
=========================================
- Coverage     72.3%   72.2%   -0.1%     
=========================================
  Files          256     256             
  Lines        56112   56133     +21     
=========================================
+ Hits         40575   40584      +9     
- Misses       15537   15549     +12     

Impacted Files	Coverage Δ
flopy/mf6/coordinates/modeldimensions.py	79.0% <60.0%> (+0.4%)	⬆️
flopy/mf6/utils/testutils.py	9.9% <66.6%> (-0.1%)	⬇️
flopy/utils/datautil.py	63.9% <92.3%> (+0.7%)	⬆️
flopy/mf6/data/mfdatalist.py	72.3% <100.0%> (+<0.1%)	⬆️
flopy/mf6/data/mfdatastorage.py	71.2% <100.0%> (ø)
flopy/mf6/data/mfdatautil.py	54.4% <100.0%> (+0.1%)	⬆️
flopy/mf6/data/mffileaccess.py	71.7% <100.0%> (+<0.1%)	⬆️
flopy/mf6/modflow/mfsimulation.py	65.6% <100.0%> (+<0.1%)	⬆️

... and 2 files with indirect coverage changes

[langevin-usgs]

I couldn't follow all the low-level changes, but if this allows the example problem to work correctly, then it seems okay. I don't see any comparison of the simulated output to a known solution. Do we need that somehow? If not, then go ahead and merge.

[spaulins-usgs]

I think my tests that flopy reads and writes correctly are sufficient. Having some problems merging this PR with develop, so I am going to close it and open a new PR for this.