Re: [cclib-devel] Distance, Angle, Dihedral

[Eric B. <er...@pi...>]

Hi Tyler,

I know you didn't ask this, but regarding the different implementations,
I'll implement them and try them out. In general, I think that assuming
performance doesn't vary too wildly, we're more interested in
implementations that are close to the original intent (the equations),
thought the "fast" one seems reasonable. A better timing test would compare
both small molecules and very large ones, like MOFs or proteins, using
`timeit`.

> should one calculate all bonds, angles, dihedrals automatically, or
should the code simply return specific bonds/angles/dihedrals that are
requested by the user (or both)?

What I propose is a new method (see
https://github.com/cclib/cclib/tree/master/src/cclib/method) that would
take a `ccdata` instance and calculate all possible distances, angles, and
torsions. In general, when we parse, we don't compute derived data; the
only things we compute are quantities for attributes _if_ they can be
computed from other attributes. The need for this is rare (see trickery for
`coreelectrons` in Q-Chem). Once the `ccdata` is passed to the method, it
is probably fine to precompute all values rather than on-the-fly, but this
is not as important as keeping the calculation separate from the parsing.
Eventually we could do both.

See if you can write a method that would basically look like the `Molecule`
but is initialized from a `ccdata` instance.

Eric

On Mon, Apr 2, 2018 at 10:10 AM, Tyler Josephson <jos...@um...> wrote:

> Thanks, Eric, this is helpful!
>
> I did find a useful and simple implementation of dihedrals in Python on
> this page: https://stackoverflow.com/questions/20305272/dihedral-
> torsion-angle-from-four-points-in-cartesian-coordinates-in-python
> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstackoverflow.com%2Fquestions%2F20305272%2Fdihedral-torsion-angle-from-four-points-in-cartesian-coordinates-in-python&data=01%7C01%7Cerb74%40pitt.edu%7C0cb5cf0f698b4584176e08d598a3a50c%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=LRIupf4KL2vU2HJH0SuJpkEdUie0rUOiSgEGq61VZv0%3D&reserved=0>
> I especially like Praexolitic's careful evaluation of the different
> equations in order to select the fastest one.
>
> For my current application, I added this in a rather ad-hoc manner to a
> copy of the ccget script, calling it from within an if statement to be
> executed when attr == 'atomcoords' . So I have what I need for now. I would
> be interested in helping add this to cclib in a more formal way. After
> seeing your code, I realize the community should probably have a discussion
> about what this should entail and how it should be organized - should one
> calculate all bonds, angles, dihedrals automatically, or should the code
> simply return specific bonds/angles/dihedrals that are requested by the
> user (or both)?
>
> What would you recommend? I'm rather new to collaborative code development.
>
> Regards,
> Tyler
>
>
>
> On Sun, Apr 1, 2018 at 10:09 PM, Eric Berquist <er...@pi...> wrote:
>
>> Hi Tyler,
>>
>> Nothing is present right now in the package to do that, though it may be
>> a useful method in the future.
>>
>> For reference, if you're looking for the expressions, I've coded up
>> distances, torsions, and angles in the past from this project:
>> https://github.com/CrawfordGroup/ProgrammingProject
>> s/tree/master/Project%2301
>> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FCrawfordGroup%2FProgrammingProjects%2Ftree%2Fmaster%2FProject%252301&data=01%7C01%7Cerb74%40pitt.edu%7C0cb5cf0f698b4584176e08d598a3a50c%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=RVJgFfkFGATY7cQmozjV8zYWZuCOa2KWhxYBE5l3uC8%3D&reserved=0>
>>
>> Some of it is available (https://github.com/berquist/p
>> rogramming_party/blob/master/eric/molecule.py
>> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fberquist%2Fprogramming_party%2Fblob%2Fmaster%2Feric%2Fmolecule.py&data=01%7C01%7Cerb74%40pitt.edu%7C0cb5cf0f698b4584176e08d598a3a50c%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=rHZe4h49qc4HM3Tow2VMDdfHYm9hWxlLEBr58Kk5zL4%3D&reserved=0>),
>> though the torsions aren't present for some reason. I can find them if
>> you're interested.
>>
>> Eric
>>
>> On Sat, Mar 31, 2018 at 7:14 AM, Tyler Josephson <jos...@um...>
>> wrote:
>>
>>> Hi there!
>>>
>>> I'm new to cclib, and enjoying it so far! It's a powerful tool. I wanted
>>> to ask the community if you've worked on functions that would extract
>>> particular distance, angle, or dihedral measurements (I need dihedrals for
>>> my particular problem). I could read in the atomcoords and calculate them
>>> myself, but I wanted to check if something else is available.
>>>
>>> Regards,
>>> Tyler
>>>
>>> --
>>> Tyler Josephson
>>> PhD Chemical Engineering
>>> Postdoctoral Research Associate, Siepmann Group
>>> University of Minnesota, Twin Cities
>>> 651-269-1433 |  | LinkedIn
>>> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.linkedin.com%2Fin%2Ftrjosephson%2F&data=01%7C01%7Cerb74%40pitt.edu%7C99323476a9ec40b3fa1d08d596fb6271%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=bqoJnaPbbc1W%2BSq81RrBgbARpQHFmXsemHI2Sbi9eiY%3D&reserved=0>
>>>
>>> ------------------------------------------------------------
>>> ------------------
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, Slashdot.org! https://na01.safelinks.protect
>>> ion.outlook.com/?url=http%3A%2F%2Fsdm.link%2Fslashdot&data=0
>>> 1%7C01%7Cerb74%40pitt.edu%7C99323476a9ec40b3fa1d08d596fb6271
>>> %7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=9ijts%2FUMc6h6
>>> 5RYv2EYB3lVuEHgIjuFVLInwZ3z1SKI%3D&reserved=0
>>> _______________________________________________
>>> cclib-devel mailing list
>>> ccl...@li...
>>> https://na01.safelinks.protection.outlook.com/?url=https%3A%
>>> 2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fcclib-devel&
>>> data=01%7C01%7Cerb74%40pitt.edu%7C99323476a9ec40b3fa1d08d5
>>> 96fb6271%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=
>>> vtaWqxXdaJ13NDCnSkApw%2FCihxRwbWBxJK9fDfK%2FpeY%3D&reserved=0
>>>
>>>
>>
>
>
> --
> Tyler Josephson
> PhD Chemical Engineering
> Postdoctoral Research Associate, Siepmann Group
> University of Minnesota, Twin Cities
> 651-269-1433 |  | LinkedIn
> <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.linkedin.com%2Fin%2Ftrjosephson%2F&data=01%7C01%7Cerb74%40pitt.edu%7C0cb5cf0f698b4584176e08d598a3a50c%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=qV%2BEiuoXrqlCNf9K8scXONaMiRvip5fImikj%2BYFdalc%3D&reserved=0>
>