gemseo-multi-fidelity 0.0.1__py3-none-any.whl

gemseo_multi_fidelity/mf_rosenbrock/rosen_mf_funcs.py

@@ -0,0 +1,350 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+# Copyright (c) 2018 AIRBUS OPERATIONS
+
+#
+# Contributors:
+#    INITIAL AUTHORS - API and implementation and/or documentation
+#      :author: Romain Olivanti
+#    OTHER AUTHORS   - MACROSCOPIC CHANGES
+"""Customized multi-fidelity Rosenbrock functions."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from numpy import asarray
+from numpy import fill_diagonal
+from numpy import roll
+from numpy import sqrt
+from numpy import sum as np_sum
+from numpy import zeros
+
+if TYPE_CHECKING:
+    from numpy.typing import NDArray
+
+
+def rosen(x_in: NDArray, coef: float = 1e2) -> NDArray:
+    """Rosenbrock function.
+
+    Scalable Rosenbrock function:
+    coef * sum(100*(x[1:] - x[:-1]**2)**2 + (1 - x[:-1])**2.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+
+    Returns:
+        The function value.
+    """
+    x_in = asarray(x_in)
+    return coef * np_sum((x_in[1:] - x_in[:-1] ** 2) ** 2) + np_sum(
+        (x_in[:-1] - 1.0) ** 2
+    )
+
+
+def rosen_grad(x_in: NDArray, coef: float = 1e2) -> NDArray:
+    """Rosenbrock gradient.
+
+    Gradient of the scalable Rosenbrock function:
+    coef * sum(100*(x[1:] - x[:-1]**2)**2 + (1 - x[:-1])**2.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+
+    Returns:
+        The gradient value.
+    """
+    x_in = asarray(x_in)
+    ret = (
+        2.0 * coef * (x_in - roll(x_in, 1) ** 2)
+        - 4.0 * coef * x_in * (roll(x_in, -1) - x_in**2)
+        + 2.0 * (x_in - 1.0)
+    )
+    # Border corrections
+    ret[0] = 2.0 * (x_in[0] - 1.0) - 4.0 * coef * x_in[0] * (x_in[1] - x_in[0] ** 2)
+    ret[-1] = 2.0 * coef * (x_in[-1] - x_in[-2] ** 2)
+    return ret
+
+
+def rosen_hess(x_in: NDArray, coef: float = 1e2) -> NDArray:
+    """Rosenbrock Hessian.
+
+    Hessian of the scalable Rosenbrock function:
+    coef * sum(100*(x[1:] - x[:-1]**2)**2 + (1 - x[:-1])**2.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+
+    Returns:
+        The Hessian.
+    """
+    x_in = asarray(x_in)
+    dim = x_in.shape[0]
+    ret = zeros((dim,) * 2)
+    for i in range(1, dim - 1):
+        ret[i, i - 1] = -4.0 * coef * x_in[i - 1]
+        ret[i, i] = 2.0 * (coef + 1.0) - 4 * coef * (x_in[i + 1] - 3.0 * x_in[i] ** 2)
+        ret[i, i + 1] = -4 * coef * x_in[i]
+    # Border corrections
+    ret[0, 0] = 2.0 - 4 * coef * (x_in[1] - 3.0 * x_in[0] ** 2)
+    ret[0, 1] = -4 * coef * x_in[0]
+    ret[-1, -1] = 2.0 * coef
+    ret[-1, -2] = -4.0 * coef * x_in[-2]
+    return ret
+
+
+def rosen_sqr(x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None) -> NDArray:
+    """Rosenbrock function (with 'squared' info).
+
+    Low-fidelity Rosenbrock function with 'squared' information:
+    coef * sum(x[1:]**2) + sum(x[:-1]**2).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The function value.
+    """
+    x_in = asarray(x_in)
+    x_vect = x_in - x_sol if x_sol is not None else x_in
+    return coef * np_sum(x_vect[1:] ** 2) + np_sum(x_vect[:-1] ** 2)
+
+
+def rosen_sqr_grad(x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None) -> NDArray:
+    """Rosenbrock gradient (with 'squared' info).
+
+    Gradient of the low-fidelity Rosenbrock function with 'squared' information:
+    coef * sum(x[1:]**2) + sum(x[:-1]**2).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The gradient value.
+    """
+    x_in = asarray(x_in)
+    x_vect = x_in - x_sol if x_sol is not None else x_in
+    ret = 2.0 * (coef + 1.0) * x_vect
+    # Border corrections
+    ret[0] = 2.0 * x_vect[0]
+    ret[-1] = 2.0 * coef * x_vect[-1]
+    return ret
+
+
+def rosen_sqr_hess(x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None) -> NDArray:
+    """Rosenbrock Hessian (with 'squared' info).
+
+    Hessian of the low-fidelity Rosenbrock function with 'squared' information:
+    coef * sum(x[1:]**2) + sum(x[:-1]**2).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The Hessian.
+    """
+    x_in = asarray(x_in)
+    dim = x_in.shape[0]
+    ret = zeros((dim,) * 2)
+    fill_diagonal(ret, 2.0 * (coef + 1))
+    # Border corrections
+    ret[0, 0] = 2.0
+    ret[-1, -1] = 2.0 * coef
+    return ret
+
+
+def rosen_quartic(x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None) -> NDArray:
+    """Rosenbrock function (with 'quartic' info).
+
+    Low-fidelity Rosenbrock function with 'quartic' information:
+    coef * (sum(x[1:]**2) + sum(x[:-1]**4)).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The function value.
+    """
+    x_in = asarray(x_in)
+    x_vect = x_in - x_sol if x_sol is not None else x_in
+    return coef * (np_sum(x_vect[1:] ** 2) + np_sum(x_vect[:-1] ** 4))
+
+
+def rosen_quartic_grad(
+    x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None
+) -> NDArray:
+    """Rosenbrock gradient (with 'quartic' info).
+
+    Gradient of the low-fidelity Rosenbrock function with 'quartic' information:
+    coef * (sum(x[1:]**2) + sum(x[:-1]**4)).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The gradient value.
+    """
+    x_in = asarray(x_in)
+    x_vect = x_in - x_sol if x_sol is not None else x_in
+    ret = 2.0 * coef * x_vect * (1.0 + 2.0 * x_vect**2)
+    # Border corrections
+    ret[0] = 4.0 * coef * x_vect[0] ** 3
+    ret[-1] = 2.0 * coef * x_vect[-1]
+    return ret
+
+
+def rosen_quartic_hess(
+    x_in: NDArray, coef: float = 1e2, x_sol: NDArray = None
+) -> NDArray:
+    """Rosenbrock Hessian (with 'quartic' info).
+
+    Hessian of the low-fidelity Rosenbrock function with 'quartic' information:
+    coef * (sum(x[1:]**2) + sum(x[:-1]**4)).
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The tunable optimum.
+
+    Returns:
+        The Hessian.
+    """
+    x_in = asarray(x_in)
+    x_vect = x_in - x_sol if x_sol is not None else x_in
+    dim = x_vect.shape[0]
+    ret = zeros((dim,) * 2)
+    ret[list(range(1, dim - 1)), list(range(1, dim - 1))] = (
+        2.0 * coef * (1.0 + 6.0 * x_vect[1:-1] ** 2)
+    )
+    # Border corrections
+    ret[0, 0] = 12.0 * coef * x_vect[0] ** 2
+    ret[-1, -1] = 2.0 * coef
+    return ret
+
+
+def rosen_custom(
+    x_in: NDArray, coef: float = 1e2, x_sol: float | None = None
+) -> NDArray:
+    """Custom Rosenbrock function.
+
+    Rosenbrock custom low-fidelity function:
+    coef * sum((x_sol * x[1:] - x[:-1]**2)**2) + 2. * sum((x - x_sol)**2)
+    - (x[0] - x_sol)**2 - (x[-1] - x_sol)**2
+    Min 0. located at x_sol.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The generalized position of the solution [x_sol] * dim.
+
+    Returns:
+        The function value.
+    """
+    if x_sol is None:
+        x_sol = sqrt(3)
+    x_in = asarray(x_in)
+    return (
+        coef * np_sum((x_sol * x_in[1:] - x_in[:-1] ** 2) ** 2)
+        + 2.0 * np_sum((x_in - x_sol) ** 2)
+        - (x_in[0] - x_sol) ** 2
+        - (x_in[-1] - x_sol) ** 2
+    )
+
+
+def rosen_custom_grad(
+    x_in: NDArray, coef: float = 1e2, x_sol: float | None = None
+) -> NDArray:
+    """Custom Rosenbrock gradient.
+
+    Gradient of the Rosenbrock custom low-fidelity function:
+    coef * sum((x_sol * x[1:] - x[:-1]**2)**2) + 2. * sum((x - x_sol)**2)
+    - (x[0] - x_sol)**2 - (x[-1] - x_sol)**2
+    Min 0. located at x_sol.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The generalized position of the solution [x_sol] * dim.
+
+    Returns:
+        The gradient value.
+    """
+    if x_sol is None:
+        x_sol = sqrt(3)
+    x_in = asarray(x_in)
+    ret = (
+        2.0 * coef * x_sol * (x_sol * x_in - roll(x_in, 1) ** 2)
+        - 4.0 * coef * x_in * (x_sol * roll(x_in, -1) - x_in**2)
+        + 4.0 * (x_in - x_sol)
+    )
+    # Border corrections
+    ret[0] = 2.0 * (x_in[0] - x_sol) - 4.0 * coef * x_in[0] * (
+        x_sol * x_in[1] - x_in[0] ** 2
+    )
+    ret[-1] = 2.0 * (x_in[-1] - x_sol) + 2.0 * coef * x_sol * (
+        x_sol * x_in[-1] - x_in[-2] ** 2
+    )
+    return ret
+
+
+def rosen_custom_hess(
+    x_in: NDArray, coef: float = 1e2, x_sol: float | None = None
+) -> NDArray:
+    """Custom Rosenbrock Hessian.
+
+    Hessian of the Rosenbrock custom low-fidelity function:
+    coef * sum((x_sol * x[1:] - x[:-1]**2)**2) + 2. * sum((x - x_sol)**2)
+    - (x[0] - x_sol)**2 - (x[-1] - x_sol)**2
+    Min 0. located at x_sol.
+
+    Args:
+        x_in: The input.
+        coef: The scaling coefficient.
+        x_sol: The generalized position of the solution [x_sol] * dim.
+
+    Returns:
+        The Hessian.
+    """
+    if x_sol is None:
+        x_sol = sqrt(3)
+    x_in = asarray(x_in)
+    dim = x_in.shape[0]
+    ret = zeros((dim,) * 2)
+    for i in range(1, dim - 1):
+        ret[i, i - 1] = -4.0 * coef * x_sol * x_in[i - 1]
+        ret[i, i] = 2.0 * (coef * x_sol**2 + 2.0) - 4 * coef * (
+            x_sol * x_in[i + 1] - 3.0 * x_in[i] ** 2
+        )
+        ret[i, i + 1] = -4 * coef * x_sol * x_in[i]
+    # Border corrections
+    ret[0, 0] = 2.0 - 4 * coef * (x_sol * x_in[1] - 3.0 * x_in[0] ** 2)
+    ret[0, 1] = -4 * coef * x_sol * x_in[0]
+    ret[-1, -1] = 2.0 * (1.0 + coef * x_sol**2)
+    ret[-1, -2] = -4.0 * coef * x_sol * x_in[-2]
+    return ret

gemseo_multi_fidelity/models/__init__.py

@@ -0,0 +1,16 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+"""Models."""

gemseo_multi_fidelity/models/fake_updater.py

@@ -0,0 +1,112 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+# Copyright (c) 2018 AIRBUS OPERATIONS
+
+#
+# Contributors:
+#    INITIAL AUTHORS - API and implementation and/or documentation
+#      :author: Romain Olivanti
+#    OTHER AUTHORS   - MACROSCOPIC CHANGES
+"""Fake updater."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from gemseo.core.mdo_functions.mdo_function import MDOFunction
+from numpy import array
+from numpy import where
+from numpy import zeros
+from numpy import zeros_like
+from numpy.linalg import norm
+
+from gemseo_multi_fidelity.models.model_updater import ModelUpdater
+
+if TYPE_CHECKING:
+    from numpy.typing import NDArray
+
+try:
+    _ = MDOFunction.hess
+    HAS_HESSIAN = True
+except AttributeError:
+    HAS_HESSIAN = False
+
+
+class FakeUpdater(ModelUpdater):
+    """Simple updater.
+
+    Storing and returning either the reference values for the reference points or
+    (0,...) for other points.
+
+    Only meant to be used for testing purposes and not meant to be differentiated.
+    """
+
+    HANDLES_MULTI_UPDATE = True
+
+    def __init__(self, **kwargs) -> None:
+        """Constructor.
+
+        Args:
+            kwargs: The keywords arguments.
+        """
+        self.x_ref = None
+        self.vals = None
+        self.grads = None
+        self.hess = None
+        super().__init__()
+
+    def _get_vect_index(self, x_vect: NDArray) -> int | None:
+        """Return the index of the stored data.
+
+        Only if x_vect exactly matches one row of self.x_ref.
+
+
+        """
+        if self.x_ref is not None:
+            ind_val = where(norm(self.x_ref - x_vect, axis=1) == 0.0)[0]
+            return ind_val[0] if len(ind_val) != 0 else None
+        return None
+
+    def _func(self, x_vect: NDArray) -> float:
+        index = self._get_vect_index(x_vect)
+        return 0.0 if index is None else self.vals[index]
+
+    def _jac(self, x_vect: NDArray) -> NDArray:
+        index = self._get_vect_index(x_vect)
+        return zeros_like(x_vect) if index is None else self.grads[index]
+
+    def _hess(self, x_vect: NDArray) -> NDArray:
+        index = self._get_vect_index(x_vect)
+        return zeros((len(x_vect),) * 2) if index is None else self.hess[index]
+
+    def _set_function(self) -> None:
+        """Set the MDOFunction."""
+        self.function = MDOFunction(self._func, "Fake", jac=self._jac)
+        if HAS_HESSIAN:
+            self.function.hess = self._hess
+
+    def _update(
+        self, vects: NDArray, vals: NDArray, grads: NDArray = None, hess: NDArray = None
+    ) -> None:
+        """Use all the points provided."""
+        self.x_ref = array(vects)
+        self.vals = array(vals)
+        self.grads = None
+        self.hess = None
+        if grads is not None:
+            self.grads = array(grads)
+        if hess is not None:
+            self.hess = array(hess)

gemseo_multi_fidelity/models/model_updater.py

@@ -0,0 +1,91 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+# Copyright (c) 2018 AIRBUS OPERATIONS
+
+#
+# Contributors:
+#    INITIAL AUTHORS - API and implementation and/or documentation
+#      :author: Romain Olivanti
+#    OTHER AUTHORS   - MACROSCOPIC CHANGES
+"""Model Updater."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from gemseo.core.mdo_functions.mdo_function import MDOFunction
+
+if TYPE_CHECKING:
+    from numpy.typing import NDArray
+
+
+class ModelUpdater:
+    """Base class for model updaters."""
+
+    HANDLES_MULTI_UPDATE = False
+
+    def __init__(self, **kwargs) -> None:
+        """Constructor.
+
+        Args:
+            kwargs: The keywords arguments.
+        """
+        # MDOFunction wrapping the model
+        self.function = None
+        self._is_set = False
+        self._set_function()
+        if not isinstance(self.function, MDOFunction):
+            msg = "Updater not setting a MDOFunction"
+            raise TypeError(msg)
+
+    @property
+    def is_set(self) -> bool:
+        """Return is_set attribute."""
+        return self._is_set
+
+    def _set_function(self):
+        """Set the MDOFunction. To be overloaded by subclasses."""
+        raise NotImplementedError
+
+    def update(
+        self, vects: NDArray, vals: NDArray, grads: NDArray = None, hess: NDArray = None
+    ) -> None:
+        """Update the model using the provided data.
+
+        The last point is guaranteed to be provided first. Gradient and hessian are
+        optional and may not be provided for each vect.
+
+        Args:
+            vects: The input space (3D array).
+            vals: The output values (1D array).
+            grads: The output gradients (2D array).
+            hess: The output hessians (3D array).
+        """
+        self._update(vects, vals, grads=grads, hess=hess)
+        self._is_set = True
+
+    def _update(
+        self, vects: NDArray, vals: NDArray, grads: NDArray = None, hess: NDArray = None
+    ):
+        """Real update routine. To be overloaded by subclasses.
+
+        Args:
+            vects: The input space (3D array).
+            vals: The output values (1D array).
+            grads: The output gradients (2D array).
+            hess: The output hessians (3D array).
+        """
+        raise NotImplementedError

gemseo_multi_fidelity/models/rbf/__init__.py

@@ -0,0 +1,16 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+"""RBF model."""

gemseo_multi_fidelity/models/rbf/kernel_factory.py

@@ -0,0 +1,66 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+# Copyright (c) 2018 AIRBUS OPERATIONS
+
+#
+# Contributors:
+#    INITIAL AUTHORS - API and implementation and/or documentation
+#      :author: Romain Olivanti
+#    OTHER AUTHORS   - MACROSCOPIC CHANGES
+"""Kernel factory."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+from typing import ClassVar
+
+from gemseo_multi_fidelity.models.rbf.kernels.gaussian import GaussianKernel
+from gemseo_multi_fidelity.models.rbf.kernels.matern_3_2 import Matern32Kernel
+from gemseo_multi_fidelity.models.rbf.kernels.matern_5_2 import Matern52Kernel
+
+if TYPE_CHECKING:
+    from gemseo_multi_fidelity.models.rbf.kernels.rbf_kernel import RBFKernel
+
+
+class KernelFactory:
+    """Kernel factory to generate a kernel from its name."""
+
+    GAUSSIAN = "gaussian"
+    MATERN_3_2 = "matern_3_2"
+    MATERN_5_2 = "matern_5_2"
+    AVAILABLE_KERNELS: ClassVar[list] = [GAUSSIAN, MATERN_3_2, MATERN_5_2]
+    """The list of available kernels."""
+
+    def create(self, name, scaling=1.0) -> RBFKernel:
+        """Factory method to create an RBF kernel.
+
+        Args:
+            name: The name of the kernel.
+            scaling: The scaling of the kernel.
+
+        Returns:
+            The RBFKernel object.
+        """
+        if name == self.GAUSSIAN:
+            kernel = GaussianKernel(scaling=scaling)
+        elif name == self.MATERN_3_2:
+            kernel = Matern32Kernel(scaling=scaling)
+        elif name == self.MATERN_5_2:
+            kernel = Matern52Kernel(scaling=scaling)
+        else:
+            msg = f"{name} not in {self.AVAILABLE_KERNELS}"
+            raise ValueError(msg)
+        return kernel

gemseo_multi_fidelity/models/rbf/kernels/__init__.py

@@ -0,0 +1,16 @@
+# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License version 3 as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+"""RBF model kernels."""