Add workflow, fix mypy

Author: atulcthakur

.github/workflows/test.yml

@@ -81,6 +81,9 @@ jobs:
           - os: ubuntu-latest
             python: "3.13"
             optionals: "ops"
+          - os: ubuntu-latest
+            python: "3.13"
+            optionals: "alchmtk"
           - os: ubuntu-latest
             python: "3.13"
           - os: macos-latest

src/matgl/ext/_tensornet_alchmtk.py

@@ -148,12 +148,11 @@ def __init__(
         self.model = model
 
         # ZBL nuclear repulsion (fixed analytical potential)
+        self.repuls: NuclearRepulsionPyG | None = None  # type: ignore[name-defined]
         if calc_repuls:
             from matgl.layers._zbl_pyg import NuclearRepulsionPyG
 
             self.repuls = NuclearRepulsionPyG(float(model.cutoff))
-        else:
-            self.repuls = None
 
         self.model_config = ModelConfig(
             outputs=frozenset({"energy", "forces", "stress"}),
@@ -222,8 +221,8 @@ def from_potential(cls, potential: Potential) -> TensorNetWrapper:
 
         return cls(
             model=potential.model,
-            data_mean=potential.data_mean.clone(),
-            data_std=potential.data_std.clone(),
+            data_mean=potential.data_mean.clone(),  # type: ignore[operator]
+            data_std=potential.data_std.clone(),  # type: ignore[operator]
             element_refs=element_refs,
             calc_repuls=getattr(potential, "calc_repuls", False),
         )
@@ -277,7 +276,7 @@ def adapt_input(self, data: AtomicData | Batch, **kwargs: Any) -> dict[str, Any]
         device = data.positions.device
         B: int = data.num_graphs
 
-        node_type = self._z_to_type[data.atomic_numbers]
+        node_type = self._z_to_type[data.atomic_numbers]  # type: ignore[index]
 
         # nvalchemi (E, 2) -> TensorNet/PyG (2, E)
         edge_index = data.neighbor_list.T  # [2, E]