Adopt lazy imports from Terra

qiskit_machine_learning/algorithms/distribution_learners/qgan/_pytorch_discriminator_net.py

@@ -1,6 +1,6 @@
 # This code is part of Qiskit.
 #
-# (C) Copyright IBM 2019, 2021.
+# (C) Copyright IBM 2019, 2022.
 #
 # This code is licensed under the Apache License, Version 2.0. You may
 # obtain a copy of this license in the LICENSE.txt file in the root directory
@@ -14,22 +14,10 @@
 Discriminator
 """
 
-import logging
-from qiskit.exceptions import MissingOptionalLibraryError
+import qiskit_machine_learning.optionals as _optionals
 
-logger = logging.getLogger(__name__)
-
-try:
+if _optionals.HAS_TORCH:
     import torch
-    from torch import nn
-except ImportError:
-    if logger.isEnabledFor(logging.INFO):
-        EXC = MissingOptionalLibraryError(
-            libname="Pytorch",
-            name="DiscriminatorNet",
-            pip_install="pip install 'qiskit-machine-learning[torch]'",
-        )
-        logger.info(str(EXC))
 
 # torch 1.6.0 fixed a mypy error about not applying contravariance rules
 # to inputs by defining forward as a value, rather than a function.  See also
@@ -39,6 +27,7 @@
 # The pylint disable=abstract-method fixes it.
 
 
+@_optionals.HAS_TORCH.require_in_instance
 class DiscriminatorNet(torch.nn.Module):  # pylint: disable=abstract-method
     """
     Discriminator
@@ -54,6 +43,9 @@ def __init__(self, n_features: int = 1, n_out: int = 1) -> None:
         """
 
         super().__init__()
+        # pylint: disable=import-error
+        from torch import nn
+
         self.n_features = n_features
 
         self.hidden0 = nn.Sequential(

qiskit_machine_learning/algorithms/distribution_learners/qgan/numpy_discriminator.py

@@ -1,6 +1,6 @@
 # This code is part of Qiskit.
 #
-# (C) Copyright IBM 2019, 2021.
+# (C) Copyright IBM 2019, 2022.
 #
 # This code is licensed under the Apache License, Version 2.0. You may
 # obtain a copy of this license in the LICENSE.txt file in the root directory
@@ -274,7 +274,7 @@ def load_model(self, load_dir):
         Load discriminator model
 
         Args:
-            load_dir (str): file with stored pytorch discriminator model to be loaded
+            load_dir (str): file with stored PyTorch discriminator model to be loaded
         """
         self._discriminator.architecture = np.load(
             os.path.join(load_dir, "np_discriminator_architecture.csv")

qiskit_machine_learning/algorithms/distribution_learners/qgan/pytorch_discriminator.py

@@ -1,6 +1,6 @@
 # This code is part of Qiskit.
 #
-# (C) Copyright IBM 2019, 2021.
+# (C) Copyright IBM 2019, 2022.
 #
 # This code is licensed under the Apache License, Version 2.0. You may
 # obtain a copy of this license in the LICENSE.txt file in the root directory
@@ -17,20 +17,12 @@
 from typing import Dict, Any, Iterable, Optional, Sequence, cast
 import os
 import numpy as np
-from qiskit.exceptions import MissingOptionalLibraryError
 from qiskit.utils import QuantumInstance
+import qiskit_machine_learning.optionals as _optionals
 from .discriminative_network import DiscriminativeNetwork
 
-try:
-    import torch
-    from torch import nn, optim
-    from torch.autograd.variable import Variable
-
-    _HAS_TORCH = True
-except ImportError:
-    _HAS_TORCH = False
-
 
+@_optionals.HAS_TORCH.require_in_instance
 class PyTorchDiscriminator(DiscriminativeNetwork):
     """
     Discriminator based on PyTorch
@@ -41,17 +33,10 @@ def __init__(self, n_features: int = 1, n_out: int = 1) -> None:
         Args:
             n_features: Dimension of input data vector.
             n_out: Dimension of the discriminator's output vector.
-
-        Raises:
-            MissingOptionalLibraryError: Pytorch not installed
         """
         super().__init__()
-        if not _HAS_TORCH:
-            raise MissingOptionalLibraryError(
-                libname="Pytorch",
-                name="PyTorchDiscriminator",
-                pip_install="pip install 'qiskit-meachine-learning[torch]'",
-            )
+        # pylint: disable=import-outside-toplevel
+        from torch import optim
 
         self._n_features = n_features
         self._n_out = n_out
@@ -73,6 +58,8 @@ def set_seed(self, seed: int):
         Args:
             seed: seed
         """
+        import torch
+
         torch.manual_seed(seed)
 
     def save_model(self, snapshot_dir: str):
@@ -82,15 +69,19 @@ def save_model(self, snapshot_dir: str):
         Args:
             snapshot_dir:  directory path for saving the model
         """
+        import torch
+
         torch.save(self._discriminator, os.path.join(snapshot_dir, "discriminator.pt"))
 
     def load_model(self, load_dir: str):
         """
         Load discriminator model
 
         Args:
-            load_dir: file with stored pytorch discriminator model to be loaded
+            load_dir: file with stored PyTorch discriminator model to be loaded
         """
+        import torch
+
         self._discriminator = torch.load(load_dir)
 
     @property
@@ -118,11 +109,14 @@ def get_label(self, x, detach=False):  # pylint: disable=arguments-differ
         Returns:
             torch.Tensor: Discriminator output, i.e. data label
         """
-
         # pylint: disable=not-callable, no-member
+        import torch
+
         if isinstance(x, torch.Tensor):
             pass
         else:
+            from torch.autograd.variable import Variable
+
             x = torch.tensor(x, dtype=torch.float32)
             x = Variable(x)
 
@@ -143,6 +137,8 @@ def loss(self, x, y, weights=None):
         Returns:
             torch.Tensor: Loss w.r.t to the generated data points.
         """
+        from torch import nn
+
         if weights is not None:
             loss_funct = nn.BCELoss(weight=weights, reduction="sum")
         else:
@@ -164,9 +160,13 @@ def gradient_penalty(self, x, lambda_=5.0, k=0.01, c=1.0):
             torch.Tensor: Gradient penalty.
         """
         # pylint: disable=not-callable, no-member
+        import torch
+
         if isinstance(x, torch.Tensor):
             pass
         else:
+            from torch.autograd.variable import Variable
+
             x = torch.tensor(x, dtype=torch.float32)
             x = Variable(x)
         # pylint: disable=no-member
@@ -207,6 +207,9 @@ def train(
         """
         # pylint: disable=E1101
         # pylint: disable=E1102
+        import torch
+        from torch.autograd.variable import Variable
+
         # Reset gradients
         self._optimizer.zero_grad()
         real_batch = cast(Sequence, data)[0]

qiskit_machine_learning/algorithms/objective_functions.py

@@ -1,6 +1,6 @@
 # This code is part of Qiskit.
 #
-# (C) Copyright IBM 2021.
+# (C) Copyright IBM 2021, 2022.
 #
 # This code is licensed under the Apache License, Version 2.0. You may
 # obtain a copy of this license in the LICENSE.txt file in the root directory
@@ -17,9 +17,14 @@
 
 import numpy as np
 
-try:
+import qiskit_machine_learning.optionals as _optionals
+from qiskit_machine_learning.neural_networks import NeuralNetwork
+from qiskit_machine_learning.utils.loss_functions import Loss
+
+if _optionals.HAS_SPARSE:
+    # pylint: disable=import-error
     from sparse import SparseArray
-except ImportError:
+else:
 
     class SparseArray:  # type: ignore
         """Empty SparseArray class
@@ -29,10 +34,6 @@ class SparseArray:  # type: ignore
         pass
 
 
-from qiskit_machine_learning.neural_networks import NeuralNetwork
-from qiskit_machine_learning.utils.loss_functions import Loss
-
-
 class ObjectiveFunction:
     """An abstract objective function. Provides methods for computing objective value and
     gradients for forward and backward passes."""

qiskit_machine_learning/connectors/torch_connector.py

@@ -14,24 +14,17 @@
 
 from typing import Tuple, Any, Optional, cast, Union
 import numpy as np
-from qiskit.exceptions import MissingOptionalLibraryError
-
-try:
-    from sparse import SparseArray, COO
-
-    _HAS_SPARSE = True
-except ImportError:
-    _HAS_SPARSE = False
 
+import qiskit_machine_learning.optionals as _optionals
 from ..neural_networks import NeuralNetwork
 from ..exceptions import QiskitMachineLearningError
 from ..deprecation import deprecate_property
 
-try:
+if _optionals.HAS_TORCH:
     from torch import Tensor, sparse_coo_tensor, einsum
     from torch.autograd import Function
     from torch.nn import Module, Parameter as TorchParam
-except ImportError:
+else:
 
     class Function:  # type: ignore
         """Empty Function class
@@ -50,17 +43,12 @@ class Tensor:  # type: ignore
     class Module:  # type: ignore
         """Empty Module class
         Replacement if torch.nn.Module is not present.
-        Always fails to initialize
         """
 
-        def __init__(self) -> None:
-            raise MissingOptionalLibraryError(
-                libname="Pytorch",
-                name="TorchConnector",
-                pip_install="pip install 'qiskit-machine-learning[torch]'",
-            )
+        pass
 
 
+@_optionals.HAS_TORCH.require_in_instance
 class TorchConnector(Module):
     """Connects a Qiskit (Quantum) Neural Network to PyTorch."""
 
@@ -88,7 +76,6 @@ def forward(  # type: ignore
 
             Raises:
                 QiskitMachineLearningError: Invalid input data.
-                MissingOptionalLibraryError: sparse not installed.
             """
 
             # validate input shape
@@ -109,12 +96,10 @@ def forward(  # type: ignore
                 input_data.detach().cpu().numpy(), weights.detach().cpu().numpy()
             )
             if neural_network.sparse and sparse:
-                if not _HAS_SPARSE:
-                    raise MissingOptionalLibraryError(
-                        libname="sparse",
-                        name="COO",
-                        pip_install="pip install 'qiskit-machine-learning[sparse]'",
-                    )
+                _optionals.HAS_SPARSE.require_now("COO")
+                # pylint: disable=import-error
+                from sparse import SparseArray, COO
+
                 result = cast(COO, cast(SparseArray, result).asformat("coo"))
                 result_tensor = sparse_coo_tensor(result.coords, result.data)
             else:

qiskit_machine_learning/datasets/ad_hoc.py

@@ -16,8 +16,7 @@
 
 import numpy as np
 import scipy
-from qiskit.utils import algorithm_globals
-from qiskit.exceptions import MissingOptionalLibraryError
+from qiskit.utils import algorithm_globals, optionals
 
 from qiskit_machine_learning.datasets.dataset_helper import (
     features_and_labels_transform,
@@ -167,14 +166,9 @@ def ad_hoc_data(
         }
 
         if plot_data:
-            try:
-                import matplotlib.pyplot as plt
-            except ImportError as ex:
-                raise MissingOptionalLibraryError(
-                    libname="Matplotlib",
-                    name="ad_hoc_data",
-                    pip_install="pip install matplotlib",
-                ) from ex
+            optionals.HAS_MATPLOTLIB.require_now("ad_hoc_data")
+            # pylint: disable=import-error
+            import matplotlib.pyplot as plt
 
             plt.show()
             fig2 = plt.figure()
@@ -260,14 +254,10 @@ def ad_hoc_data(
         }
 
         if plot_data:
-            try:
-                import matplotlib.pyplot as plt
-            except ImportError as ex:
-                raise MissingOptionalLibraryError(
-                    libname="Matplotlib",
-                    name="ad_hoc_data",
-                    pip_install="pip install matplotlib",
-                ) from ex
+            optionals.HAS_MATPLOTLIB.require_now("ad_hoc_data")
+            # pylint: disable=import-error
+            import matplotlib.pyplot as plt
+
             sample_total_a = np.asarray(sample_total_a)
             sample_total_b = np.asarray(sample_total_b)
             x_1 = sample_total_a[:, 0]

qiskit_machine_learning/datasets/breast_cancer.py

@@ -19,7 +19,7 @@
 from sklearn.model_selection import train_test_split
 from sklearn.preprocessing import StandardScaler, MinMaxScaler
 from sklearn.decomposition import PCA
-from qiskit.exceptions import MissingOptionalLibraryError
+from qiskit.utils import optionals
 from .dataset_helper import features_and_labels_transform
 from ..deprecation import deprecate_function
 
@@ -68,14 +68,10 @@ def breast_cancer(training_size, test_size, n, plot_data=False, one_hot=True):
     )
 
     if plot_data:
-        try:
-            import matplotlib.pyplot as plt
-        except ImportError as ex:
-            raise MissingOptionalLibraryError(
-                libname="Matplotlib",
-                name="breast_cancer",
-                pip_install="pip install matplotlib",
-            ) from ex
+        optionals.HAS_MATPLOTLIB.require_now("breast_cancer")
+        # pylint: disable=import-error
+        import matplotlib.pyplot as plt
+
         for k in range(0, 2):
             plt.scatter(
                 sample_train[label_train == k, 0][:training_size],