feat(modflow): support dataframe for pkg data

[wpbonelli]

• close feature: support dataframe for package data #1917
• support pandas dataframe for node_data, stress_period_data etc for flopy.modflow.* pkgs
• just convert to recarray — not reimplementing data storage with pandas
• test with mnw and wel

[codecov]

Codecov Report

Merging #2010 (d128440) into develop (7cc5e6f) will increase coverage by 0.0%.
Report is 2 commits behind head on develop.
The diff coverage is 55.1%.

Additional details and impacted files

@@           Coverage Diff           @@
##           develop   #2010   +/-   ##
=======================================
  Coverage     72.7%   72.8%           
=======================================
  Files          259     259           
  Lines        58956   59025   +69     
=======================================
+ Hits         42916   42975   +59     
- Misses       16040   16050   +10     

Files	Coverage Δ
flopy/modflow/mfag.py	91.1% <100.0%> (+<0.1%)	⬆️
flopy/modflow/mfchd.py	90.3% <ø> (ø)
flopy/modflow/mfdrn.py	84.1% <ø> (ø)
flopy/modflow/mfdrt.py	67.5% <ø> (ø)
flopy/modflow/mfghb.py	90.2% <ø> (ø)
flopy/modflow/mfmnw2.py	83.0% <100.0%> (+0.1%)	⬆️
flopy/modflow/mfriv.py	93.2% <ø> (ø)
flopy/modflow/mfwel.py	88.6% <ø> (ø)
flopy/utils/mflistfile.py	71.2% <ø> (-0.1%)	⬇️
flopy/modflow/mfgage.py	89.0% <66.6%> (-0.5%)	⬇️
... and 5 more

... and 2 files with indirect coverage changes

[langevin-usgs]

Are there any requirements on the column names for the pandas implementation? Looks like the requirement is that the number of columns must be correct. Seems good to me.

[wpbonelli]

Colnames for dataframe should be the same as for recarray. Internally MfList still tries to coerce everything to recarray, checking dtypes as it does. I added another case in test_modflow.py to try both and make sure it handles them.