Atomtype parameterization methods for parameterizing a topology in GMSO. #644

CalCraven · 2022-04-14T20:38:14Z

This PR will add the atomtyping module to GMSO for passing a GMSO.Forcefield
and a GMSO.Topology and match atomtype using foyer as the backend. Then,
the corresponding connection types will be found in the Forcefield and
applied to the connections in the topology.

Create parameterize.py, which has the apply function which can take a
topology, and a gmso forcefield to apply to it. This can use subgraph
isomorphism to identify molecular structures in the topology through
the bondgraph and bin those into unique molecules that match a specified
forcefield. This apply function can also do the standard atomtyping of
the entire topology in one step.
Create isomorph.py which uses networkx graphs to identify disconnected
components and isomorphism to identify repeated structures.

~~Checklist:~~
~~- [x] Use site residues to type repeated structures~~
~~- [ ] Write a function that will add residue information to specified sites~~
~~- [ ] Evaluate docstring clarity~~
~~- [ ] Write unit tests~~
~~- [x] Match pre-commit style~~

Updated Checklist:

Feature Parity with Parmed
Multi--Forcefield application
Isomorphic Test

This PR will add the atomtyping module to GMSO for passing a GMSO.Forcefield and a GMSO.Topology and match atomtype using foyer as the backend. Then, the corresponding connection types will be found in the Forcefield and applied to the connections in the topology. * Create parameterize.py, which has the apply function which can take a topology, and a gmso forcefield to apply to it. This can use subgraph isomorphism to identify molecular structures in the topology through the bondgraph and bin those into unique molecules that match a specified forcefield. This apply function can also do the standard atomtyping of the entire topology in one step. * Create isomorph.py which uses networkx graphs to identify disconnected components and isomorphism to identify repeated structures.

lgtm-com · 2022-04-14T20:43:44Z

This pull request introduces 2 alerts when merging 8e69d6e into efeaad2 - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

…orts

lgtm-com · 2022-04-14T21:57:55Z

This pull request introduces 2 alerts when merging 9074909 into efeaad2 - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

umesh-timalsina

This is shaping up nicely @CalCraven. I have added some comments by skimming through it.

gmso/atomtyping/isomorph.py

umesh-timalsina · 2022-04-19T19:44:31Z

gmso/atomtyping/isomorph.py

+ matcher = nx.algorithms.isomorphism.GraphMatcher(
+ graph, graph_of_interest


I think we need to add a node_match function (based on the element of the node) IIRC.

What exactly would a node_match function do? I'm not entirely familiar, is it something passed to the GraphMatcher to do the node matching?

just tested this offline, I think the current use of GraphMatcher only concern about the shape of the graph but not about the nodes involved. So this could cause issue with molecules with same shape but different element would still be recognized as isomorphic (so lumped into the same group).

ethanol = mb.load("CCO", smiles=True) ethanol.name = "Ethanol" eth1 = mb.clone(ethanol) eth2 = mb.clone(ethanol) eth1_top = mb_convert(eth1) eth1_top.sites[0].name = "K" eth1_top.sites[0].element = element_by_symbol("K") eth2_top = mb_convert(eth2)

Note below I am using the get_modified_topology_graph from the example notebook.

eth1_grp = get_modified_topology_graph(eth1_top) eth2_grp = get_modified_topology_graph(eth2_top) matcher = nx.isomorphism.GraphMatcher(eth1_grp, eth2_grp)

Then
matcher.is_isomorphic() will still return True

on a different note, I think we can use this GraphMatcher for the issue with the assumed site order in a molecule (or subtopology) (also this mosdef-hub/foyer#504). Will have to sacrifice some performance for checking but the number of checks performed will still be fewer that going the completely splitting the topology, so should still be a bit faster.

Okay this makes more sense to me now, thanks @daico007. I was a little hazy on the exact implementation of GraphMatcher. I think having a few different options for matching makes sense, which trade of increasing performance for increasing matching precision. This reminds me a bit of the filters used in PR #649, which can be used to specify how you want to filter unique parameter types. I think it makes sense to at least try a basic one that checks at least for elements in the correct positions, and allow that to be a built in option.

gmso/atomtyping/isomorph.py

gmso/atomtyping/parameterize.py

umesh-timalsina · 2022-04-19T19:56:23Z

gmso/atomtyping/parameterize.py

+ for connect in getattr(top, "_" + parameter + "s"):
+ sub_top_label = connect.connection_members[0].label_
+ ff_of_interest = forcefields[sub_top_label]
+ type_getter = attrgetter("atom_type.atomclass")


Is a None-Type check prudent here? I think foyer will raise an error if it cannot find an atomtype right?

Would we want to use the assert_bond_params flag to see if we should or should not look for all instances of that connection to be typed? This will definitely throw an error if it's missing as of now. Line 97 essentially handles this check as of now.

Added an error message in e1f3d4e.

gmso/atomtyping/parameterize.py

umesh-timalsina · 2022-04-19T19:57:45Z

gmso/atomtyping/parameterize.py

+ index=gmso_topology.get_index(atom),
+ atomic_number=None,
+ element=atom.name,
+ subtopology_label=atom.label,


Not sure where we want to address this but we need a better identifier than atom.label.

I agree. I will think about this, and see if there is something better we can be using.

Should not the atom be the atom number in the topology?

The index is the atom_number in the topology. The subtopology_label should be the string that we use to apply a given forcefield to the sites with this label.

According to PR #638, this info will be stored in site.molecule. I think whatever convention we decide to use there can be the reference label we use to identify the molecule name, and therefore the identifier we use for the forcefield

gmso/atomtyping/parameterize.py

for more information, see https://pre-commit.ci

lgtm-com · 2022-04-25T04:32:51Z

This pull request introduces 2 alerts when merging 6376da7 into f9e7332 - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

…lop-gmsoff-parameterize

codecov · 2022-04-26T15:28:01Z

Codecov Report

Merging #644 (318c5ad) into develop (4436e35) will increase coverage by 0.59%.
The diff coverage is 99.53%.

@@             Coverage Diff             @@
##           develop     #644      +/-   ##
===========================================
+ Coverage    91.10%   91.70%   +0.59%     
===========================================
  Files           57       64       +7     
  Lines         4758     4964     +206     
===========================================
+ Hits          4335     4552     +217     
+ Misses         423      412      -11

Impacted Files	Coverage Δ
gmso/parameterization/topology_parameterizer.py	`99.15% <99.15%> (ø)`
gmso/core/topology.py	`96.90% <100.00%> (ø)`
gmso/external/convert_parmed.py	`95.59% <100.00%> (+0.33%)`	⬆️
gmso/parameterization/__init__.py	`100.00% <100.00%> (ø)`
gmso/parameterization/foyer_utils.py	`100.00% <100.00%> (ø)`
gmso/parameterization/isomorph.py	`100.00% <100.00%> (ø)`
gmso/parameterization/parameterize.py	`100.00% <100.00%> (ø)`
gmso/parameterization/subtopology_utils.py	`100.00% <100.00%> (ø)`
gmso/parameterization/utils.py	`100.00% <100.00%> (ø)`
gmso/core/angle.py	`96.00% <0.00%> (-0.30%)`	⬇️
... and 8 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4436e35...318c5ad. Read the comment docs.

lgtm-com · 2022-04-28T16:33:44Z

This pull request introduces 2 alerts when merging 790aacc into cd68411 - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

…arameterize

lgtm-com · 2022-04-28T19:57:37Z

This pull request introduces 2 alerts when merging 2d7a770 into 1818e46 - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

…arameterize

lgtm-com · 2022-04-29T18:12:40Z

This pull request introduces 2 alerts when merging 40f6639 into f9c40ea - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

…neralize where the molecule definition is pulled from

for more information, see https://pre-commit.ci

lgtm-com · 2022-05-02T15:39:29Z

This pull request introduces 2 alerts when merging 1503578 into f9c40ea - view on LGTM.com

new alerts:

1 for Unused local variable
1 for Illegal raise

daico007 · 2022-05-03T09:23:54Z

gmso/atomtyping/parameterize.py

+ } # TODO validate improper order
+ molecule_label = "label_" # place to grab info referencing molecule name
+ for connect in getattr(top, "_" + parameter + "s"):
+ sub_top_label = connect.connection_members[0].get(molecule_label)


Suggested change

sub_top_label = connect.connection_members[0].get(molecule_label)

sub_top_label = connect.connection_members[0].__dict__.get(molecule_label)

or

Suggested change

sub_top_label = connect.connection_members[0].get(molecule_label)

sub_top_label = connect.connection_members[0].molecule_label

daico007 · 2022-05-03T09:24:54Z

gmso/atomtyping/parameterize.py

+ molecule_label = "label_" # place to grab info referencing molecule name
+ for connect in getattr(top, "_" + parameter + "s"):
+ sub_top_label = connect.connection_members[0].get(molecule_label)
+ if sub_top_label is None:


Suggested change

if sub_top_label is None:

if not sub_top_label:

I think this would also catch the case where the sub_top_label is empty string.

CalCraven · 2022-05-04T16:55:33Z

Rename module to to parameterization, not atomtyping

…arameterize

lgtm-com · 2022-05-06T19:17:29Z

This pull request introduces 1 alert when merging 7be5f57 into 2605c65 - view on LGTM.com

new alerts:

1 for Illegal raise

lgtm-com · 2022-05-06T22:00:58Z

This pull request introduces 1 alert when merging 7366452 into 2605c65 - view on LGTM.com

new alerts:

1 for Unused import

* Include ImproperType reading in "from_parmed" This PR is branched from PR #644. The parmed conversion to GMSO topology was not able to read in various Improper connections to GMSO. It would aggregate them with the Dihedral connections, from when GMSO did not have explicit treatment of impropers, similar to ParmEd. This PR aims to check a ParmEd object for structure.impropers or structure.dihedrals with the dihedral.impropers flag set to True and convert those to `gmso.Impropers`. Likewise, it looks to read in periodic or harmonic impropers and generate the correct `gmso.ImproperType` parametric potential for that type, if the flag `refer_types` is set to True in the `from_parmed` call. * Add pmd_improper_types_map function to convert_parmed * Generate a parametric potential with harmonic type expression from pmd.impropers * Generate a parametric potential with periodic type expression from pmd.dihedrals and dihedral.improper=True * Validate testing with foyer opls_validation when using functions in PR #644 Harmonic type expression for impropers is taken from https://manual.gromacs.org/current/reference-manual/functions/bonded-interactions.html#improper-dihedrals-harmonic-type This uses Xi as the independent variable. If someone testing this PR can validate that this is the formalism we want to use when reading in ImproperTypes from ParmEd improper_types, that would be a good sanity check. The periodic type expression is taken from https://manual.gromacs.org/current/reference-manual/functions/bonded-interactions.html#equation-eqnperiodicpropdihedral with phi as the independent variable. * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Add test for number of impropers from NNdimethylformamide molecule * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * WIP- include gmso in impropertype testing * WIP- Add support for NoneTypes in site charge/mass * Address Review comments and create impropertype expression from HarmonicImproperPotential in library * WIP- Add testing for impropers * Modification for improper potential forms from the PotentialTemplateLibrary * WIP- Use id instead of object for storing mapping; impropertypes test * WIP- combined test for dihedral/impropertypes * Fix test case and id mapping in conversion * Add tests for harmonic impropers * Add proper properties for improper_types * WIP-Additional test cases Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Umesh Timalsina <umesh.timalsina@vanderbilt.edu>

…arameterize

daico007 · 2022-06-15T14:59:14Z

One edge case popped up when @bc118 was testing on this PR: when a topology with no subtopology (so only two layers topology and sites) is being typed, the apply can only take one single forcefield XML file, and not in the form of a dictionary ({molecule_name: forcefield_file}). Handling this issue include erroring out and asking user for the appropriate input or consider the topology as a subtop (so matching top.name to the key in the XML dictionary provided).

During that discussion, another question raised, do all forcefield provided in the forcefield XML dictionary has to be used or do we allow users to supercharge and provide excess forcefields. One idea is to not allow it by default but only when expert_mode=True.

…raven/gmso into develop-gmsoff-parameterize

* Modify scaling factors to a numpy array * Misc Changes 1. Split property into two parts. global_scaling_factors and per_molecule_scaling_factors 2. Handle Nan/NoneType cases, non float cases 3. Add additional functions for handling scaling factors 4. Change formats/json.py to incorporate new properties 5. Change function name to prefix `get_`, formats/mcf.py change * WIP- Refactor global_scaling_factors to scaling_factors * Additional tests; Simplificaiton of get_scaling_factors method * Atomtype parameterization methods for parameterizing a topology in GMSO. (#644) * Atomtype parameterization methods for parameterizing a topology in GMSO. This PR will add the atomtyping module to GMSO for passing a GMSO.Forcefield and a GMSO.Topology and match atomtype using foyer as the backend. Then, the corresponding connection types will be found in the Forcefield and applied to the connections in the topology. * Create parameterize.py, which has the apply function which can take a topology, and a gmso forcefield to apply to it. This can use subgraph isomorphism to identify molecular structures in the topology through the bondgraph and bin those into unique molecules that match a specified forcefield. This apply function can also do the standard atomtyping of the entire topology in one step. * Create isomorph.py which uses networkx graphs to identify disconnected components and isomorphism to identify repeated structures. * Move module imports for apply into atomtyping to prevent circular imports * Add a quick fix which will do atomtyping if no residue flag is passed * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * correctly update sites when adding subtop * Changes to doc strings for clarity. Add a subtop_label variable to generalize where the molecule definition is pulled from * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * More modular architecture * WIP- Add testing, minor refactoring of the APIs. * WIP- Better error handling * Misc Changes 1. Update Topology after parametrization 2. Add dependency for forcefield_utilities in env files 3. Add tests for trappe forcefield 4. Patch parmed (should be moved to a cal's PR #658 * WIP- Skip opls for now, full TrappE tests * Avoid accidental overwriting of typemap when using isomorphism * WIP- Remove unused import * Use enumerate for atom index while converting to TopologyGraph * Fix argument order * WIP- Add test for subtopology parameterization * Make opls/trappe global fixtures, Add tests for isomorphism * Further testing isomorphism * REVERT - skip OPLS tests * Copy scaling factors and combining rules after parametrization * Proper OPLS tests * WIP- Refactor the test module * WIP- Remove unused import * WIP- Add test for parameterization with impropers * WIP- Additional impropers test; Separate module for testing impropers * Minor refacotors; additional edge cases coverage/tests * Docstring minor fix * Remove rel_to_module as is obsolete in forcefield_utilities * Change trappe_ua to trappe-ua for correct loading * fix typo, add note about specific use case * pip install forcefield-utilites until new release Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Co Quach <daico007@gmail.com> Co-authored-by: Co Quach <43968221+daico007@users.noreply.github.com> Co-authored-by: Umesh Timalsina <umesh.timalsina@vanderbilt.edu> * Properly apply different scaling factors while parameterizing Co-authored-by: CalCraven <54594941+CalCraven@users.noreply.github.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Co Quach <daico007@gmail.com> Co-authored-by: Co Quach <43968221+daico007@users.noreply.github.com>

Move module imports for apply into atomtyping to prevent circular imp…

9074909

…orts

umesh-timalsina reviewed Apr 19, 2022

View reviewed changes

CalCraven and others added 2 commits April 24, 2022 23:25

Add a quick fix which will do atomtyping if no residue flag is passed

80c1850

[pre-commit.ci] auto fixes from pre-commit.com hooks

6376da7

for more information, see https://pre-commit.ci

Merge branch 'master' of https://github.com/mosdef-hub/gmso into deve…

7d0a790

…lop-gmsoff-parameterize

correctly update sites when adding subtop

99c0793

umesh-timalsina mentioned this pull request Apr 26, 2022

Feature: Add non_element_types support for atomtyping #648

Merged

Merge branch 'develop' into develop-gmsoff-parameterize

790aacc

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

2d7a770

…arameterize

umesh-timalsina added 2 commits April 29, 2022 10:30

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

e6da3f7

…arameterize

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

40f6639

…arameterize

CalCraven and others added 2 commits May 2, 2022 10:33

Changes to doc strings for clarity. Add a subtop_label variable to ge…

e1f3d4e

…neralize where the molecule definition is pulled from

[pre-commit.ci] auto fixes from pre-commit.com hooks

1503578

for more information, see https://pre-commit.ci

daico007 reviewed May 3, 2022

View reviewed changes

umesh-timalsina added 2 commits May 6, 2022 10:48

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

b13bf94

…arameterize

More modular architecture

7be5f57

WIP- Add testing, minor refactoring of the APIs.

7366452

umesh-timalsina added 6 commits June 3, 2022 09:55

WIP- Remove unused import

c7ea346

Use enumerate for atom index while converting to TopologyGraph

d267d2f

Fix argument order

9aab528

WIP- Add test for subtopology parameterization

51b1081

Make opls/trappe global fixtures, Add tests for isomorphism

06df2ba

Further testing isomorphism

d7b08d6

umesh-timalsina added 9 commits June 7, 2022 10:18

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

031e703

…arameterize

REVERT - skip OPLS tests

4a523e6

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

403a36c

…arameterize

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

593c15a

…arameterize

Copy scaling factors and combining rules after parametrization

341ecc5

Proper OPLS tests

fe352fd

WIP- Refactor the test module

b3e045f

WIP- Remove unused import

4834b5f

WIP- Add test for parameterization with impropers

0a4c780

umesh-timalsina mentioned this pull request Jun 9, 2022

Expanded search for improper_types #665

Merged

Merge remote-tracking branch 'upstream/develop' into develop-gmsoff-p…

5633c52

…arameterize

umesh-timalsina and others added 8 commits June 15, 2022 13:19

WIP- Additional impropers test; Separate module for testing impropers

3711172

Minor refacotors; additional edge cases coverage/tests

956977c

Docstring minor fix

f0fd1e5

Remove rel_to_module as is obsolete in forcefield_utilities

1e866d7

Change trappe_ua to trappe-ua for correct loading

da3a23e

fix typo, add note about specific use case

b018a6c

Merge branch 'develop-gmsoff-parameterize' of https://github.com/CalC…

4c62c5c

…raven/gmso into develop-gmsoff-parameterize

pip install forcefield-utilites until new release

318c5ad

daico007 merged commit 7fe2dc5 into mosdef-hub:develop Jun 15, 2022

CalCraven mentioned this pull request Aug 10, 2022

Specifying forcefields to specific residues mosdef-hub/mbuild#751

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Atomtype parameterization methods for parameterizing a topology in GMSO. #644

Atomtype parameterization methods for parameterizing a topology in GMSO. #644

CalCraven commented Apr 14, 2022 •

edited by umesh-timalsina

Loading

lgtm-com bot commented Apr 14, 2022

lgtm-com bot commented Apr 14, 2022

umesh-timalsina left a comment

umesh-timalsina Apr 19, 2022

CalCraven Apr 28, 2022

daico007 May 2, 2022

daico007 May 2, 2022

daico007 May 2, 2022

CalCraven May 2, 2022

umesh-timalsina Apr 19, 2022

CalCraven Apr 28, 2022

CalCraven May 2, 2022

umesh-timalsina Apr 19, 2022

CalCraven Apr 27, 2022

bc118 Apr 28, 2022

CalCraven May 2, 2022

CalCraven May 2, 2022

lgtm-com bot commented Apr 25, 2022

codecov bot commented Apr 26, 2022 •

edited

Loading

lgtm-com bot commented Apr 28, 2022

lgtm-com bot commented Apr 28, 2022

lgtm-com bot commented Apr 29, 2022

lgtm-com bot commented May 2, 2022

daico007 May 3, 2022

daico007 May 3, 2022

CalCraven commented May 4, 2022

lgtm-com bot commented May 6, 2022

lgtm-com bot commented May 6, 2022

daico007 commented Jun 15, 2022

		matcher = nx.algorithms.isomorphism.GraphMatcher(
		graph, graph_of_interest

	sub_top_label = connect.connection_members[0].get(molecule_label)
	sub_top_label = connect.connection_members[0].__dict__.get(molecule_label)

Atomtype parameterization methods for parameterizing a topology in GMSO. #644

Atomtype parameterization methods for parameterizing a topology in GMSO. #644

Conversation

CalCraven commented Apr 14, 2022 • edited by umesh-timalsina Loading

Updated Checklist:

lgtm-com bot commented Apr 14, 2022

lgtm-com bot commented Apr 14, 2022

umesh-timalsina left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lgtm-com bot commented Apr 25, 2022

codecov bot commented Apr 26, 2022 • edited Loading

Codecov Report

lgtm-com bot commented Apr 28, 2022

lgtm-com bot commented Apr 28, 2022

lgtm-com bot commented Apr 29, 2022

lgtm-com bot commented May 2, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

CalCraven commented May 4, 2022

lgtm-com bot commented May 6, 2022

lgtm-com bot commented May 6, 2022

daico007 commented Jun 15, 2022

CalCraven commented Apr 14, 2022 •

edited by umesh-timalsina

Loading

codecov bot commented Apr 26, 2022 •

edited

Loading