Improve some typing #2000

Yoshanuikabundi · 2025-01-23T07:03:07Z

This PR includes many changes needed for OpenFF Pablo to type check, and 1 for it to run correctly. Pablo development so far has and will continue to test and type check against this branch (or main once it is merged) until it gets into a release.

improves typing as needed by OpenFF Pablo by adding and correcting many type annotations
Handle an exception previously raised when Molecule.to_rdkit() is run on a molecule with residue numbers that cannot be converted to int (such as hexadecimal residue numbers used by OpenMM in large PDB files)
Makes Molecule.is_isomorphic_with() so that its keyword arguments are actually keyword arguments. This is not a breaking change as the **kwargs argument is still present, it's just ignored - we should deprecate and remove this ASAP
Update changelog

Sorry for the behemoth PR!

codecov · 2025-01-23T07:37:59Z

Codecov Report

Attention: Patch coverage is 97.87234% with 1 line in your changes missing coverage. Please review.

Project coverage is 82.62%. Comparing base (d5d4cf8) to head (e717570).

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

mattwthompson · 2025-02-14T14:32:42Z

Okey - sorry to screw up your PR briefly, I think I've brought it up to date with the non-trivial changes that have been happening upstream. I have some ideas about how we can revamp typing in general, as I'm sure you do, but this is neither the time or the place for me to push that

…al ints

mattwthompson

Process-wise, this is an "approve" in terms of what looks good. But I found that I broke CI at some point and Mypy isn't actually running. I would like to see it running in these CI checks (ideally on the most recent version, 1.17.0 I think) before merging this. #2087 or something similar should do it.

I have a few other comments which you are welcome to take or leave. This was fun to review and I like seeing these improvements trickle into our codebase!

mattwthompson · 2025-07-10T23:35:51Z

openff/toolkit/topology/molecule.py

-            The stereochemistry around this atom, allowed values are "CW", "CCW", or None,
+            The stereochemistry around this atom, allowed values are "R", "S", or None,


Related #2084

mattwthompson · 2025-07-14T23:26:51Z

docs/releasehistory.md

@@ -14,6 +14,8 @@ Releases follow the `major.minor.micro` scheme recommended by [PEP440](https://w

 ### Bugfixes
 - [PR #2052](https://github.com/openforcefield/openff-toolkit/pull/2052): Fixes bug where `Topology.from_pdb` couldn't load NH4+ ([Issue #2051](https://github.com/openforcefield/openff-toolkit/issues/2051))
+- [PR #2000](https://github.com/openforcefield/openff-toolkit/pull/2000): Assorted type annotation fixes
+- [PR #2000](https://github.com/openforcefield/openff-toolkit/pull/2000): `Molecule.to_rdkit()` no longer raises an exception when converting a molecule with non-decimal residue numbers to RDKit.


Could use a line here about what happens in this case since the behavior is new and not obviously well-defined

mattwthompson · 2025-07-14T23:29:09Z

openff/toolkit/topology/topology.py

+            off_atom.metadata["residue_name"] = atom.residue.name # type:ignore[attr-defined]
+            off_atom.metadata["residue_number"] = atom.residue.id # type:ignore[attr-defined]
+            off_atom.metadata["insertion_code"] = atom.residue.insertionCode # type:ignore[attr-defined]
+            off_atom.metadata["chain_id"] = atom.residue.chain.id # type:ignore[attr-defined]


Thought a little bit about how to make this cleaner, could only come up with using getattr which isn't actually any cleaner. Remind me that all of these attributes being optional is (intentionally) part of the design?

mattwthompson · 2025-07-14T23:29:43Z

openff/toolkit/topology/topology.py

@@ -2343,7 +2343,8 @@ def atom(self, atom_topology_index: int) -> "Atom":
            if next_molecule_start_index > atom_topology_index:
                atom_molecule_index = atom_topology_index - this_molecule_start_index
                # NOTE: the index here should still be in the topology index order, NOT the reference molecule's
-                return molecule.atom(atom_molecule_index)
+                # can Molecule.atom be a _SimpleAtom?


No I ... uhh ... hope not.

mattwthompson · 2025-07-14T23:33:18Z

openff/toolkit/utils/builtin_wrapper.py

@@ -131,7 +131,7 @@ def assign_partial_charges(
            partial_charges = [0.0] * molecule.n_atoms

        elif partial_charge_method == "formal_charge":
-            partial_charges = [float(atom.formal_charge.m) for atom in molecule.atoms]
+            partial_charges = [float(atom.formal_charge.m) for atom in molecule.atoms]  # type: ignore


What's the error code here?

I won't be picky enough to ask here, but in a perfect world I'd love to see us always have the specific error code in these ignores. Mypy can enforce this

mattwthompson · 2025-07-14T23:35:55Z

openff/toolkit/topology/molecule.py

-            aromatic_matching=kwargs.get("aromatic_matching", True),
-            formal_charge_matching=kwargs.get("formal_charge_matching", True),
-            bond_order_matching=kwargs.get("bond_order_matching", True),
-            atom_stereochemistry_matching=kwargs.get(
-                "atom_stereochemistry_matching", True
-            ),
-            bond_stereochemistry_matching=kwargs.get(
-                "bond_stereochemistry_matching", True
-            ),
-            strip_pyrimidal_n_atom_stereo=kwargs.get(
-                "strip_pyrimidal_n_atom_stereo", True
-            ),
-            toolkit_registry=kwargs.get("toolkit_registry", GLOBAL_TOOLKIT_REGISTRY),
+            aromatic_matching=aromatic_matching,
+            formal_charge_matching=formal_charge_matching,
+            bond_order_matching=bond_order_matching,
+            atom_stereochemistry_matching=atom_stereochemistry_matching,
+            bond_stereochemistry_matching=bond_stereochemistry_matching,
+            strip_pyrimidal_n_atom_stereo=strip_pyrimidal_n_atom_stereo,
+            toolkit_registry=toolkit_registry,


mattwthompson · 2025-07-14T23:37:14Z

openff/toolkit/topology/molecule.py

@@ -2879,7 +2885,7 @@ def to_networkx(self) -> "nx.Graph":
        """
        import networkx as nx

-        G: nx.classes.graph.Graph = nx.Graph()
+        G: nx.classes.graph.Graph[int] = nx.Graph()


Woah! Does this mean they've implemented it as a generic?

mattwthompson · 2025-07-15T16:44:57Z

I'm working on upstream fixes in #2087, will update here when that's done or if I decide to pull the plug

mattwthompson · 2025-07-15T21:58:28Z

#2087 is a mess. In the spirit of keeping the ball rolling, I'm okay merging this as-is

j-wags · 2025-07-16T21:39:34Z

openff/toolkit/topology/molecule.py

        """Set the atoms stereochemistry
        Parameters
        ----------
        value
-            The stereochemistry around this atom, allowed values are "CW", "CCW", or None,
+            The stereochemistry around this atom, allowed values are "R", "S", or None,


Suggested change

The stereochemistry around this atom, allowed values are "R", "S", or None,

The stereochemistry around this atom, suggested values are "R", "S", or None,

jameseastwood assigned Yoshanuikabundi Feb 10, 2025

Yoshanuikabundi marked this pull request as draft February 10, 2025 22:29

Yoshanuikabundi added 9 commits July 4, 2025 10:40

Improve some typing

b1f4f66

Fix type annotation for OMMQuantity

04b5550

Revert formatting changes

f68267e

Restore support for py3.10

7b4b009

Fix mypy errors

f7cb9e4

Type annotate return of _SimpleMolecule bond and atom methods

dba7493

Type annotate arg of _SimpleMolecule bond and atom methods

1b49b89

Make HierarchyScheme.identifier variadic

ce045a4

Fix fstring in _detect_undefined_stereo

89d0d29

Yoshanuikabundi force-pushed the better_typing branch 2 times, most recently from 7d7f1c8 to 3dc07df Compare July 4, 2025 00:51

Yoshanuikabundi marked this pull request as ready for review July 4, 2025 03:33

Yoshanuikabundi requested a review from j-wags as a code owner July 4, 2025 03:33

mattwthompson and others added 4 commits July 4, 2025 13:44

Fixes to get mypy passing

33b93a8

Hacks for Pint compatibility

d080402

Molecule.to_rdkit(): support residue number strings that aren't decim…

fb54cca

…al ints

Update changelog

e906ff2

Yoshanuikabundi force-pushed the better_typing branch from 2b59dd4 to e906ff2 Compare July 4, 2025 03:45

Yoshanuikabundi added 2 commits July 4, 2025 14:44

More type hints and more formatting reversions

2cc1fd4

fix missing annotation referent

e717570

j-wags assigned j-wags and unassigned Yoshanuikabundi Jul 10, 2025

j-wags requested review from mattwthompson and removed request for j-wags July 10, 2025 22:03

j-wags assigned mattwthompson and unassigned j-wags Jul 11, 2025

mattwthompson requested changes Jul 14, 2025

View reviewed changes

mattwthompson approved these changes Jul 15, 2025

View reviewed changes

j-wags assigned Yoshanuikabundi Jul 15, 2025

mattwthompson removed their assignment Jul 16, 2025

j-wags reviewed Jul 16, 2025

View reviewed changes

j-wags mentioned this pull request Jul 16, 2025

Update Atom.stereochemistry docstring #2086

Closed

5 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Improve some typing #2000

Improve some typing #2000

Uh oh!

Yoshanuikabundi commented Jan 23, 2025 •

edited

Loading

Uh oh!

codecov bot commented Jan 23, 2025 •

edited

Loading

Uh oh!

mattwthompson commented Feb 14, 2025

Uh oh!

mattwthompson left a comment

Uh oh!

mattwthompson Jul 10, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson Jul 14, 2025

Uh oh!

mattwthompson commented Jul 15, 2025

Uh oh!

mattwthompson commented Jul 15, 2025

Uh oh!

j-wags Jul 16, 2025

Uh oh!

Uh oh!

		The stereochemistry around this atom, allowed values are "CW", "CCW", or None,
		The stereochemistry around this atom, allowed values are "R", "S", or None,

	The stereochemistry around this atom, allowed values are "R", "S", or None,
	The stereochemistry around this atom, suggested values are "R", "S", or None,

Improve some typing #2000

Are you sure you want to change the base?

Improve some typing #2000

Uh oh!

Conversation

Yoshanuikabundi commented Jan 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov bot commented Jan 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

mattwthompson commented Feb 14, 2025

Uh oh!

mattwthompson left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mattwthompson commented Jul 15, 2025

Uh oh!

mattwthompson commented Jul 15, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Yoshanuikabundi commented Jan 23, 2025 •

edited

Loading

codecov bot commented Jan 23, 2025 •

edited

Loading