Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Christoph Klein
    @ctk3b
    based on info in forcedata.py
    Jeff Thompson
    @jpthompson17
    ok I think missed it but I'll check it out
    Christoph Klein
    @ctk3b
    It’s also a bit messy but it means we never have to make the forces by hand
    but if we want to add this canonical thing we’ll have to add a line to that so it can pick the write conversion
    Jeff Thompson
    @jpthompson17
    so if I wanted to add, say, a functional form for a lammps harmonic dihedral (http://lammps.sandia.gov/doc/dihedral_harmonic.html), which would also cover the identical cvff improper functional form, I could use that?
    the question I have then is...if I were to go in in the future to add an entirely new plugin--for NAMD or Amber or Tinker or whatever--and I again needed to define a new functional form, should it still go into the common 'forces' subdirectory?
    I suppose this model makes sense since there is a lot of overlap in functional forms and it encourages the most reuse
    Christoph Klein
    @ctk3b
    Yes that’s the idea. Is the lammps harmonic not covered by intermol’s proper periodic?
    But I really do like this canonical function idea
    less code that each parser has to handle
    again, factoring out something common
    Jeff Thompson
    @jpthompson17
    not quite...I mean the information it contains is identical, but the coefficients are different so it fails in the automated parsing of coefficients
    Christoph Klein
    @ctk3b
    ah
    wait so
    is the issue just that intermol is doing something silly right now?
    Jeff Thompson
    @jpthompson17
    so even though it only takes a line to go from K (1 + d cos(nphi) ) to K (1 + cos(n*phi - phase) ), it seems it would be best to create a new class
    Christoph Klein
    @ctk3b
    If I’m understanding, correctly, you’d only have to get the phase set to zero
    oooo
    Jeff Thompson
    @jpthompson17
    if d == 1: phase = 0; else phase = 180
    Christoph Klein
    @ctk3b
    missed the d coefficient
    Jeff Thompson
    @jpthompson17
    yeah
    ...well really elif d == -1
    please let me know if I'm missing something
    Christoph Klein
    @ctk3b
    So this sounds almost like we should just be adding another parameter to our proper dihedral
    If it defaults to zero, the gromacs code doesnt even need to be changed
    Jeff Thompson
    @jpthompson17
    yeah I thought about that--have ProperPeriodicDihedral store a consistent d and phi
    or...no?
    Christoph Klein
    @ctk3b
    Consistent? I was just suggesting adding another parameter - is that what you’re saying? And then we’d have to modify the convert_XXX function
    Ok so let’s see. Organizationally, how would we start with this canonical method idea?
    Add the function to each dihedral force that we care about. It would go in the XxxType super class probably since that’s where the parameters are
    and this function calls the appropriate convert_dihedral_from_xxx_to_yyy
    But this purely converts the functional form so the parsers will still have to do some modification for that specific package
    Jeff Thompson
    @jpthompson17
    yeah except ultimately the convert_dihedral_from_x_to_y would be replaced by conversion to/from canonical form
    Christoph Klein
    @ctk3b
    Right. It would always be convert x to Trig or Periodic Improper
    er
    ImproperHarmonic
    Jeff Thompson
    @jpthompson17
    but also the subtypes should be able to do the opposite conversion
    Christoph Klein
    @ctk3b
    subtypes?
    the forces?
    Jeff Thompson
    @jpthompson17
    and then the parser would just key into the subtype (specific functional form) to print coeffs
    yes the forces
    I have to run and eat
    may be back later
    Christoph Klein
    @ctk3b
    alrighty
    I’ll try to write this up as a more defined issue. Probably not before tomorrow morning
    Stephan Grein
    @stephanmg
    hey.
    is it possible to convert from GROMACS to LAMMPS (and generating then for LAMMPS a CHARMM force field) with Intermol?
    Christoph Klein
    @ctk3b
    I believe that should work. GROMACS to LAMMPS conversions support most functional forms.
    Make sure you compare the energies with the -e flag and let me know if you run into any problems!
    Ghost
    @ghost~5be1bc53d73408ce4fae0cc8
    Hi guys,
    Nice to meet you.
    I need your help with moleculer.
    Please let me know what tool can be used instead of istanbul for code coverage in moleculer.
    Or what is the best tool for code coverage in moleculer?