feat: improve amino acid features

[DaniBodor]

Summary of changes:

• Main changes to aminoacidlist.py
  • Corrections of amino acid properties in aminoacidlist.py
    • The charge was incorrect for every single amino acid (there was a blanket calculation error)
    • Few other mistakes were found in some amino acids
  • New amino acid properties
    • residue mass
    • isolectric point
  • Sources added for all information
    • Websites indicated in comments
    • Summary in an excel file in the package
• Other changes:
  • Read new properties in deeprankcore.features.components
  • Reordered code that deals with amino acid properties to have the same order in all modules
  • Rename APOLAR --> NONPOLAR to align with most literature sources
  • Minor updates to notebook for generating test data
  • Explicitly test all amino acid features in test_components
  • Allow for exceptions in test_aminoacidlist
• Recreate hdf5 files with new/corrected features
  • Using 25 data points for residue.hdf5, instead of previous 50(?) to reduce size

[gcroci2]

Great fixes and additions :)

[DaniBodor]

Coverage appears to be decreasing because I noticed a missing init in neuralnets and added it (despite not really being part of this PR). 2 of the modules in neuralnets are not currently tested, but due to the missing init, this was not picked up coveralls.

In reality, the coverage of molstruct/amino_acid.py and features/components.py increased slightly and nothing else was affected.