Make compatible with numpy 2.0

basf · Aug 9, 2024 · c539825 · c539825
1 parent cc98b76
commit c539825
Show file tree

Hide file tree

Showing 15 changed files with 93 additions and 92 deletions.
diff --git a/.github/workflows/linting.yml b/.github/workflows/linting.yml
@@ -34,7 +34,6 @@ jobs:
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip
-        pip install "numpy<2.0.0"
         pip install mypy
         mypy . || exit_code=$?
         mypy --install-types --non-interactive

diff --git a/molpipeline/abstract_pipeline_elements/mol2any/mol2floatvector.py b/molpipeline/abstract_pipeline_elements/mol2any/mol2floatvector.py
@@ -68,18 +68,18 @@ def n_features(self) -> int:
 
     def assemble_output(
         self,
-        value_list: Iterable[npt.NDArray[np.float_]],
-    ) -> npt.NDArray[np.float_]:
+        value_list: Iterable[npt.NDArray[np.float64]],
+    ) -> npt.NDArray[np.float64]:
         """Transform output of all transform_single operations to matrix.
 
         Parameters
         ----------
-        value_list: Iterable[npt.NDArray[np.float_]]
+        value_list: Iterable[npt.NDArray[np.float64]]
             List of numpy arrays with calculated descriptor values of each molecule.
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             Matrix with descriptor values of each molecule.
         """
         return np.vstack(list(value_list))
@@ -127,7 +127,7 @@ def set_params(self, **parameters: dict[str, Any]) -> Self:
         super().set_params(**parameter_copy)
         return self
 
-    def fit_to_result(self, values: list[npt.NDArray[np.float_]]) -> Self:
+    def fit_to_result(self, values: list[npt.NDArray[np.float64]]) -> Self:
         """Fit object to data.
 
         Parameters
@@ -146,25 +146,25 @@ def fit_to_result(self, values: list[npt.NDArray[np.float_]]) -> Self:
         return self
 
     def _normalize_matrix(
-        self, value_matrix: npt.NDArray[np.float_]
-    ) -> npt.NDArray[np.float_]:
+        self, value_matrix: npt.NDArray[np.float64]
+    ) -> npt.NDArray[np.float64]:
         """Normalize matrix with descriptor values.
 
         Parameters
         ----------
-        value_matrix: npt.NDArray[np.float_]
+        value_matrix: npt.NDArray[np.float64]
             Matrix with descriptor values of molecules.
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             Normalized matrix with descriptor values of molecules.
         """
         if self._standardizer is not None:
             return self._standardizer.transform(value_matrix)
         return value_matrix
 
-    def transform(self, values: list[RDKitMol]) -> npt.NDArray[np.float_]:
+    def transform(self, values: list[RDKitMol]) -> npt.NDArray[np.float64]:
         """Transform the list of molecules to sparse matrix.
 
         Parameters
@@ -174,13 +174,15 @@ def transform(self, values: list[RDKitMol]) -> npt.NDArray[np.float_]:
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             Matrix with descriptor values of molecules.
         """
-        descriptor_matrix: npt.NDArray[np.float_] = super().transform(values)
+        descriptor_matrix: npt.NDArray[np.float64] = super().transform(values)
         return descriptor_matrix
 
-    def finalize_single(self, value: npt.NDArray[np.float_]) -> npt.NDArray[np.float_]:
+    def finalize_single(
+        self, value: npt.NDArray[np.float64]
+    ) -> npt.NDArray[np.float64]:
         """Finalize single value. Here: standardize vector.
 
         Parameters
@@ -201,7 +203,7 @@ def finalize_single(self, value: npt.NDArray[np.float_]) -> npt.NDArray[np.float
     @abc.abstractmethod
     def pretransform_single(
         self, value: RDKitMol
-    ) -> Union[npt.NDArray[np.float_], InvalidInstance]:
+    ) -> Union[npt.NDArray[np.float64], InvalidInstance]:
         """Transform mol to dict, where items encode columns indices and values, respectively.
 
         Parameters
@@ -211,6 +213,6 @@ def pretransform_single(
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             Vector with descriptor values of molecule.
         """
diff --git a/molpipeline/estimators/chemprop/abstract.py b/molpipeline/estimators/chemprop/abstract.py
@@ -114,15 +114,15 @@ def _update_trainer(
     def fit(
         self,
         X: MoleculeDataset,  # pylint: disable=invalid-name
-        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float_],
+        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float64],
     ) -> Self:
         """Fit the model to the data.
 
         Parameters
         ----------
         X : MoleculeDataset
             The input data.
-        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float_]
+        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float64]
             The target data.
 
         Returns

diff --git a/molpipeline/estimators/chemprop/models.py b/molpipeline/estimators/chemprop/models.py
@@ -128,7 +128,7 @@ def _is_classifier(self) -> bool:
 
     def _predict(
         self, X: MoleculeDataset  # pylint: disable=invalid-name
-    ) -> npt.NDArray[np.float_]:
+    ) -> npt.NDArray[np.float64]:
         """Predict the labels.
 
         Parameters
@@ -138,7 +138,7 @@ def _predict(
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             The predictions for the input data.
         """
         self.model.eval()
@@ -164,15 +164,15 @@ def _predict(
     def fit(
         self,
         X: MoleculeDataset,
-        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float_],
+        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float64],
     ) -> Self:
         """Fit the model to the data.
 
         Parameters
         ----------
         X : MoleculeDataset
             The input data.
-        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float_]
+        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float64]
             The target data.
 
         Returns
@@ -186,7 +186,7 @@ def fit(
 
     def predict(
         self, X: MoleculeDataset  # pylint: disable=invalid-name
-    ) -> npt.NDArray[np.float_]:
+    ) -> npt.NDArray[np.float64]:
         """Predict the output.
 
         Parameters
@@ -196,7 +196,7 @@ def predict(
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             The predictions for the input data.
         """
         predictions = self._predict(X)
@@ -213,7 +213,7 @@ def predict(
     @available_if(_is_classifier)
     def predict_proba(
         self, X: MoleculeDataset  # pylint: disable=invalid-name
-    ) -> npt.NDArray[np.float_]:
+    ) -> npt.NDArray[np.float64]:
         """Predict the probabilities.
 
         Parameters
@@ -223,7 +223,7 @@ def predict_proba(
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             The probabilities of the input data.
         """
         if self._is_binary_classifier():

diff --git a/molpipeline/estimators/chemprop/neural_fingerprint.py b/molpipeline/estimators/chemprop/neural_fingerprint.py
@@ -53,15 +53,15 @@ def __init__(
     def fit(
         self,
         X: MoleculeDataset,  # pylint: disable=invalid-name
-        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float_],
+        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float64],
     ) -> Self:
         """Fit the model.
 
         Parameters
         ----------
         X : MoleculeDataset
             The input data.
-        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float_]
+        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float64]
             The target data.
 
         Returns
@@ -75,7 +75,7 @@ def fit(
 
     def transform(
         self, X: MoleculeDataset  # pylint: disable=invalid-name
-    ) -> npt.NDArray[np.float_]:
+    ) -> npt.NDArray[np.float64]:
         """Transform the input.
 
         Parameters
@@ -85,7 +85,7 @@ def transform(
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             The neural fingerprint of the input data.
         """
         self.model.eval()
@@ -95,20 +95,20 @@ def transform(
     def fit_transform(
         self,
         X: MoleculeDataset,  # pylint: disable=invalid-name
-        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float_],
-    ) -> npt.NDArray[np.float_]:
+        y: Sequence[int | float] | npt.NDArray[np.int_ | np.float64],
+    ) -> npt.NDArray[np.float64]:
         """Fit the model and transform the input.
 
         Parameters
         ----------
         X : MoleculeDataset
             The input data.
-        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float_]
+        y : Sequence[int | float] | npt.NDArray[np.int_ | np.float64]
             The target data.
 
         Returns
         -------
-        npt.NDArray[np.float_]
+        npt.NDArray[np.float64]
             The neural fingerprint of the input data.
         """
         self.fit(X, y)

diff --git a/molpipeline/estimators/nearest_neighbor.py b/molpipeline/estimators/nearest_neighbor.py
@@ -37,7 +37,7 @@
 
 AllMetrics = Union[
     SklearnNativeMetrics,
-    Callable[[Any, Any], float | npt.NDArray[np.float_] | Sequence[float]],
+    Callable[[Any, Any], float | npt.NDArray[np.float64] | Sequence[float]],
 ]
 
 
@@ -153,7 +153,7 @@ def predict(
 
         Returns
         -------
-        tuple[npt.NDArray[Any], npt.NDArray[np.float_]] | npt.NDArray[Any]
+        tuple[npt.NDArray[Any], npt.NDArray[np.float64]] | npt.NDArray[Any]
             The indices of the nearest points in the population matrix and the distances to the points.
         """
         if self.learned_names_ is None:
@@ -179,7 +179,7 @@ def fit_predict(
         y: Sequence[Any],
         return_distance: bool = False,
         n_neighbors: int | None = None,
-    ) -> tuple[npt.NDArray[Any], npt.NDArray[np.float_]] | npt.NDArray[Any]:
+    ) -> tuple[npt.NDArray[Any], npt.NDArray[np.float64]] | npt.NDArray[Any]:
         """Find the k-neighbors of a point.
 
         Parameters