HBond Analysis, efficiency, usability, LAMMPS friendliness

[jaclark5]

Expected behavior

• The language in the documentation specifies the cutoffs in hbond analysis to be the min or max allowed which implies inclusive logical statements
• LAMMPS files only have types (at the moment, it is possible that molecule numbers and types could be saved as resnames and names) making this package unusable.
• A result of nan from the function is difficult to decipher
• guess_donor function takes a long time when hydrogen atoms are already known

Actual behavior

• The cutoffs in hbond analysis code use exclusive statements.
• LAMMPS files cannot be used with this function
• One must be developer savy enough to comb through the package to find out why a value of NaN was given.
• guess_donor function could use bonding information since hydrogen types are called anyway and would be faster, instead of existing distance cutoff.

[richardjgowers]

Hi @jaclark5 !

I'd fully agree that the current hydrogen bonds analysis code is very bio-leaning, so expects residues & resnames, and could be more general.