Videre

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/__init__.py

@@ -0,0 +1,20 @@
+from .main import minimize
+from .utils import show_versions
+
+# PEP0440 compatible formatted version, see:
+# https://www.python.org/dev/peps/pep-0440/
+#
+# Final release markers:
+#   X.Y.0   # For first release after an increment in Y
+#   X.Y.Z   # For bugfix releases
+#
+# Admissible pre-release markers:
+#   X.YaN   # Alpha release
+#   X.YbN   # Beta release
+#   X.YrcN  # Release Candidate
+#
+# Dev branch marker is: 'X.Y.dev' or 'X.Y.devN' where N is an integer.
+# 'X.Y.dev0' is the canonical version of 'X.Y.dev'.
+__version__ = "1.1.3"
+
+__all__ = ["minimize", "show_versions"]

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/settings.py

@@ -0,0 +1,132 @@
+import sys
+from enum import Enum
+
+import numpy as np
+
+
+# Exit status.
+class ExitStatus(Enum):
+    """
+    Exit statuses.
+    """
+
+    RADIUS_SUCCESS = 0
+    TARGET_SUCCESS = 1
+    FIXED_SUCCESS = 2
+    CALLBACK_SUCCESS = 3
+    FEASIBLE_SUCCESS = 4
+    MAX_EVAL_WARNING = 5
+    MAX_ITER_WARNING = 6
+    INFEASIBLE_ERROR = -1
+    LINALG_ERROR = -2
+
+
+class Options(str, Enum):
+    """
+    Options.
+    """
+
+    DEBUG = "debug"
+    FEASIBILITY_TOL = "feasibility_tol"
+    FILTER_SIZE = "filter_size"
+    HISTORY_SIZE = "history_size"
+    MAX_EVAL = "maxfev"
+    MAX_ITER = "maxiter"
+    NPT = "nb_points"
+    RHOBEG = "radius_init"
+    RHOEND = "radius_final"
+    SCALE = "scale"
+    STORE_HISTORY = "store_history"
+    TARGET = "target"
+    VERBOSE = "disp"
+
+
+class Constants(str, Enum):
+    """
+    Constants.
+    """
+
+    DECREASE_RADIUS_FACTOR = "decrease_radius_factor"
+    INCREASE_RADIUS_FACTOR = "increase_radius_factor"
+    INCREASE_RADIUS_THRESHOLD = "increase_radius_threshold"
+    DECREASE_RADIUS_THRESHOLD = "decrease_radius_threshold"
+    DECREASE_RESOLUTION_FACTOR = "decrease_resolution_factor"
+    LARGE_RESOLUTION_THRESHOLD = "large_resolution_threshold"
+    MODERATE_RESOLUTION_THRESHOLD = "moderate_resolution_threshold"
+    LOW_RATIO = "low_ratio"
+    HIGH_RATIO = "high_ratio"
+    VERY_LOW_RATIO = "very_low_ratio"
+    PENALTY_INCREASE_THRESHOLD = "penalty_increase_threshold"
+    PENALTY_INCREASE_FACTOR = "penalty_increase_factor"
+    SHORT_STEP_THRESHOLD = "short_step_threshold"
+    LOW_RADIUS_FACTOR = "low_radius_factor"
+    BYRD_OMOJOKUN_FACTOR = "byrd_omojokun_factor"
+    THRESHOLD_RATIO_CONSTRAINTS = "threshold_ratio_constraints"
+    LARGE_SHIFT_FACTOR = "large_shift_factor"
+    LARGE_GRADIENT_FACTOR = "large_gradient_factor"
+    RESOLUTION_FACTOR = "resolution_factor"
+    IMPROVE_TCG = "improve_tcg"
+
+
+# Default options.
+DEFAULT_OPTIONS = {
+    Options.DEBUG.value: False,
+    Options.FEASIBILITY_TOL.value: np.sqrt(np.finfo(float).eps),
+    Options.FILTER_SIZE.value: sys.maxsize,
+    Options.HISTORY_SIZE.value: sys.maxsize,
+    Options.MAX_EVAL.value: lambda n: 500 * n,
+    Options.MAX_ITER.value: lambda n: 1000 * n,
+    Options.NPT.value: lambda n: 2 * n + 1,
+    Options.RHOBEG.value: 1.0,
+    Options.RHOEND.value: 1e-6,
+    Options.SCALE.value: False,
+    Options.STORE_HISTORY.value: False,
+    Options.TARGET.value: -np.inf,
+    Options.VERBOSE.value: False,
+}
+
+# Default constants.
+DEFAULT_CONSTANTS = {
+    Constants.DECREASE_RADIUS_FACTOR.value: 0.5,
+    Constants.INCREASE_RADIUS_FACTOR.value: np.sqrt(2.0),
+    Constants.INCREASE_RADIUS_THRESHOLD.value: 2.0,
+    Constants.DECREASE_RADIUS_THRESHOLD.value: 1.4,
+    Constants.DECREASE_RESOLUTION_FACTOR.value: 0.1,
+    Constants.LARGE_RESOLUTION_THRESHOLD.value: 250.0,
+    Constants.MODERATE_RESOLUTION_THRESHOLD.value: 16.0,
+    Constants.LOW_RATIO.value: 0.1,
+    Constants.HIGH_RATIO.value: 0.7,
+    Constants.VERY_LOW_RATIO.value: 0.01,
+    Constants.PENALTY_INCREASE_THRESHOLD.value: 1.5,
+    Constants.PENALTY_INCREASE_FACTOR.value: 2.0,
+    Constants.SHORT_STEP_THRESHOLD.value: 0.5,
+    Constants.LOW_RADIUS_FACTOR.value: 0.1,
+    Constants.BYRD_OMOJOKUN_FACTOR.value: 0.8,
+    Constants.THRESHOLD_RATIO_CONSTRAINTS.value: 2.0,
+    Constants.LARGE_SHIFT_FACTOR.value: 10.0,
+    Constants.LARGE_GRADIENT_FACTOR.value: 10.0,
+    Constants.RESOLUTION_FACTOR.value: 2.0,
+    Constants.IMPROVE_TCG.value: True,
+}
+
+# Printing options.
+PRINT_OPTIONS = {
+    "threshold": 6,
+    "edgeitems": 2,
+    "linewidth": sys.maxsize,
+    "formatter": {
+        "float_kind": lambda x: np.format_float_scientific(
+            x,
+            precision=3,
+            unique=False,
+            pad_left=2,
+        )
+    },
+}
+
+# Constants.
+BARRIER = 2.0 ** min(
+    100,
+    np.finfo(float).maxexp // 2,
+    -np.finfo(float).minexp // 2,
+)

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/subsolvers/__init__.py

@@ -0,0 +1,14 @@
+from .geometry import cauchy_geometry, spider_geometry
+from .optim import (
+    tangential_byrd_omojokun,
+    constrained_tangential_byrd_omojokun,
+    normal_byrd_omojokun,
+)
+
+__all__ = [
+    "cauchy_geometry",
+    "spider_geometry",
+    "tangential_byrd_omojokun",
+    "constrained_tangential_byrd_omojokun",
+    "normal_byrd_omojokun",
+]

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/subsolvers/geometry.py

@@ -0,0 +1,387 @@
+import inspect
+
+import numpy as np
+
+from ..utils import get_arrays_tol
+
+
+TINY = np.finfo(float).tiny
+
+
+def cauchy_geometry(const, grad, curv, xl, xu, delta, debug):
+    r"""
+    Maximize approximately the absolute value of a quadratic function subject
+    to bound constraints in a trust region.
+
+    This function solves approximately
+
+    .. math::
+
+        \max_{s \in \mathbb{R}^n} \quad \bigg\lvert c + g^{\mathsf{T}} s +
+        \frac{1}{2} s^{\mathsf{T}} H s \bigg\rvert \quad \text{s.t.} \quad
+        \left\{ \begin{array}{l}
+            l \le s \le u,\\
+            \lVert s \rVert \le \Delta,
+        \end{array} \right.
+
+    by maximizing the objective function along the constrained Cauchy
+    direction.
+
+    Parameters
+    ----------
+    const : float
+        Constant :math:`c` as shown above.
+    grad : `numpy.ndarray`, shape (n,)
+        Gradient :math:`g` as shown above.
+    curv : callable
+        Curvature of :math:`H` along any vector.
+
+            ``curv(s) -> float``
+
+        returns :math:`s^{\mathsf{T}} H s`.
+    xl : `numpy.ndarray`, shape (n,)
+        Lower bounds :math:`l` as shown above.
+    xu : `numpy.ndarray`, shape (n,)
+        Upper bounds :math:`u` as shown above.
+    delta : float
+        Trust-region radius :math:`\Delta` as shown above.
+    debug : bool
+        Whether to make debugging tests during the execution.
+
+    Returns
+    -------
+    `numpy.ndarray`, shape (n,)
+        Approximate solution :math:`s`.
+
+    Notes
+    -----
+    This function is described as the first alternative in Section 6.5 of [1]_.
+    It is assumed that the origin is feasible with respect to the bound
+    constraints and that `delta` is finite and positive.
+
+    References
+    ----------
+    .. [1] T. M. Ragonneau. *Model-Based Derivative-Free Optimization Methods
+       and Software*. PhD thesis, Department of Applied Mathematics, The Hong
+       Kong Polytechnic University, Hong Kong, China, 2022. URL:
+       https://theses.lib.polyu.edu.hk/handle/200/12294.
+    """
+    if debug:
+        assert isinstance(const, float)
+        assert isinstance(grad, np.ndarray) and grad.ndim == 1
+        assert inspect.signature(curv).bind(grad)
+        assert isinstance(xl, np.ndarray) and xl.shape == grad.shape
+        assert isinstance(xu, np.ndarray) and xu.shape == grad.shape
+        assert isinstance(delta, float)
+        assert isinstance(debug, bool)
+        tol = get_arrays_tol(xl, xu)
+        assert np.all(xl <= tol)
+        assert np.all(xu >= -tol)
+        assert np.isfinite(delta) and delta > 0.0
+    xl = np.minimum(xl, 0.0)
+    xu = np.maximum(xu, 0.0)
+
+    # To maximize the absolute value of a quadratic function, we maximize the
+    # function itself or its negative, and we choose the solution that provides
+    # the largest function value.
+    step1, q_val1 = _cauchy_geom(const, grad, curv, xl, xu, delta, debug)
+    step2, q_val2 = _cauchy_geom(
+        -const,
+        -grad,
+        lambda x: -curv(x),
+        xl,
+        xu,
+        delta,
+        debug,
+    )
+    step = step1 if abs(q_val1) >= abs(q_val2) else step2
+
+    if debug:
+        assert np.all(xl <= step)
+        assert np.all(step <= xu)
+        assert np.linalg.norm(step) < 1.1 * delta
+    return step
+
+
+def spider_geometry(const, grad, curv, xpt, xl, xu, delta, debug):
+    r"""
+    Maximize approximately the absolute value of a quadratic function subject
+    to bound constraints in a trust region.
+
+    This function solves approximately
+
+    .. math::
+
+        \max_{s \in \mathbb{R}^n} \quad \bigg\lvert c + g^{\mathsf{T}} s +
+        \frac{1}{2} s^{\mathsf{T}} H s \bigg\rvert \quad \text{s.t.} \quad
+        \left\{ \begin{array}{l}
+            l \le s \le u,\\
+            \lVert s \rVert \le \Delta,
+        \end{array} \right.
+
+    by maximizing the objective function along given straight lines.
+
+    Parameters
+    ----------
+    const : float
+        Constant :math:`c` as shown above.
+    grad : `numpy.ndarray`, shape (n,)
+        Gradient :math:`g` as shown above.
+    curv : callable
+        Curvature of :math:`H` along any vector.
+
+            ``curv(s) -> float``
+
+        returns :math:`s^{\mathsf{T}} H s`.
+    xpt : `numpy.ndarray`, shape (n, npt)
+        Points defining the straight lines. The straight lines considered are
+        the ones passing through the origin and the points in `xpt`.
+    xl : `numpy.ndarray`, shape (n,)
+        Lower bounds :math:`l` as shown above.
+    xu : `numpy.ndarray`, shape (n,)
+        Upper bounds :math:`u` as shown above.
+    delta : float
+        Trust-region radius :math:`\Delta` as shown above.
+    debug : bool
+        Whether to make debugging tests during the execution.
+
+    Returns
+    -------
+    `numpy.ndarray`, shape (n,)
+        Approximate solution :math:`s`.
+
+    Notes
+    -----
+    This function is described as the second alternative in Section 6.5 of
+    [1]_. It is assumed that the origin is feasible with respect to the bound
+    constraints and that `delta` is finite and positive.
+
+    References
+    ----------
+    .. [1] T. M. Ragonneau. *Model-Based Derivative-Free Optimization Methods
+       and Software*. PhD thesis, Department of Applied Mathematics, The Hong
+       Kong Polytechnic University, Hong Kong, China, 2022. URL:
+       https://theses.lib.polyu.edu.hk/handle/200/12294.
+    """
+    if debug:
+        assert isinstance(const, float)
+        assert isinstance(grad, np.ndarray) and grad.ndim == 1
+        assert inspect.signature(curv).bind(grad)
+        assert (
+            isinstance(xpt, np.ndarray)
+            and xpt.ndim == 2
+            and xpt.shape[0] == grad.size
+        )
+        assert isinstance(xl, np.ndarray) and xl.shape == grad.shape
+        assert isinstance(xu, np.ndarray) and xu.shape == grad.shape
+        assert isinstance(delta, float)
+        assert isinstance(debug, bool)
+        tol = get_arrays_tol(xl, xu)
+        assert np.all(xl <= tol)
+        assert np.all(xu >= -tol)
+        assert np.isfinite(delta) and delta > 0.0
+    xl = np.minimum(xl, 0.0)
+    xu = np.maximum(xu, 0.0)
+
+    # Iterate through the straight lines.
+    step = np.zeros_like(grad)
+    q_val = const
+    s_norm = np.linalg.norm(xpt, axis=0)
+
+    # Set alpha_xl to the step size for the lower-bound constraint and
+    # alpha_xu to the step size for the upper-bound constraint.
+
+    # xl.shape = (N,)
+    # xpt.shape = (N, M)
+    # i_xl_pos.shape = (M, N)
+    i_xl_pos = (xl > -np.inf) & (xpt.T > -TINY * xl)
+    i_xl_neg = (xl > -np.inf) & (xpt.T < TINY * xl)
+    i_xu_pos = (xu < np.inf) & (xpt.T > TINY * xu)
+    i_xu_neg = (xu < np.inf) & (xpt.T < -TINY * xu)
+
+    # (M, N)
+    alpha_xl_pos = np.atleast_2d(
+        np.broadcast_to(xl, i_xl_pos.shape)[i_xl_pos] / xpt.T[i_xl_pos]
+    )
+    # (M,)
+    alpha_xl_pos = np.max(alpha_xl_pos, axis=1, initial=-np.inf)
+    # make sure it's (M,)
+    alpha_xl_pos = np.broadcast_to(np.atleast_1d(alpha_xl_pos), xpt.shape[1])
+
+    alpha_xl_neg = np.atleast_2d(
+        np.broadcast_to(xl, i_xl_neg.shape)[i_xl_neg] / xpt.T[i_xl_neg]
+    )
+    alpha_xl_neg = np.max(alpha_xl_neg, axis=1, initial=np.inf)
+    alpha_xl_neg = np.broadcast_to(np.atleast_1d(alpha_xl_neg), xpt.shape[1])
+
+    alpha_xu_neg = np.atleast_2d(
+        np.broadcast_to(xu, i_xu_neg.shape)[i_xu_neg] / xpt.T[i_xu_neg]
+    )
+    alpha_xu_neg = np.max(alpha_xu_neg, axis=1, initial=-np.inf)
+    alpha_xu_neg = np.broadcast_to(np.atleast_1d(alpha_xu_neg), xpt.shape[1])
+
+    alpha_xu_pos = np.atleast_2d(
+        np.broadcast_to(xu, i_xu_pos.shape)[i_xu_pos] / xpt.T[i_xu_pos]
+    )
+    alpha_xu_pos = np.max(alpha_xu_pos, axis=1, initial=np.inf)
+    alpha_xu_pos = np.broadcast_to(np.atleast_1d(alpha_xu_pos), xpt.shape[1])
+
+    for k in range(xpt.shape[1]):
+        # Set alpha_tr to the step size for the trust-region constraint.
+        if s_norm[k] > TINY * delta:
+            alpha_tr = max(delta / s_norm[k], 0.0)
+        else:
+            # The current straight line is basically zero.
+            continue
+
+        alpha_bd_pos = max(min(alpha_xu_pos[k], alpha_xl_neg[k]), 0.0)
+        alpha_bd_neg = min(max(alpha_xl_pos[k], alpha_xu_neg[k]), 0.0)
+
+        # Set alpha_quad_pos and alpha_quad_neg to the step size to the extrema
+        # of the quadratic function along the positive and negative directions.
+        grad_step = grad @ xpt[:, k]
+        curv_step = curv(xpt[:, k])
+        if (
+            grad_step >= 0.0
+            and curv_step < -TINY * grad_step
+            or grad_step <= 0.0
+            and curv_step > -TINY * grad_step
+        ):
+            alpha_quad_pos = max(-grad_step / curv_step, 0.0)
+        else:
+            alpha_quad_pos = np.inf
+        if (
+            grad_step >= 0.0
+            and curv_step > TINY * grad_step
+            or grad_step <= 0.0
+            and curv_step < TINY * grad_step
+        ):
+            alpha_quad_neg = min(-grad_step / curv_step, 0.0)
+        else:
+            alpha_quad_neg = -np.inf
+
+        # Select the step that provides the largest value of the objective
+        # function if it improves the current best. The best positive step is
+        # either the one that reaches the constraints or the one that reaches
+        # the extremum of the objective function along the current direction
+        # (only possible if the resulting step is feasible). We test both, and
+        # we perform similar calculations along the negative step.
+        # N.B.: we select the largest possible step among all the ones that
+        # maximize the objective function. This is to avoid returning the zero
+        # step in some extreme cases.
+        alpha_pos = min(alpha_tr, alpha_bd_pos)
+        alpha_neg = max(-alpha_tr, alpha_bd_neg)
+        q_val_pos = (
+            const + alpha_pos * grad_step + 0.5 * alpha_pos**2.0 * curv_step
+        )
+        q_val_neg = (
+            const + alpha_neg * grad_step + 0.5 * alpha_neg**2.0 * curv_step
+        )
+        if alpha_quad_pos < alpha_pos:
+            q_val_quad_pos = (
+                const
+                + alpha_quad_pos * grad_step
+                + 0.5 * alpha_quad_pos**2.0 * curv_step
+            )
+            if abs(q_val_quad_pos) > abs(q_val_pos):
+                alpha_pos = alpha_quad_pos
+                q_val_pos = q_val_quad_pos
+        if alpha_quad_neg > alpha_neg:
+            q_val_quad_neg = (
+                const
+                + alpha_quad_neg * grad_step
+                + 0.5 * alpha_quad_neg**2.0 * curv_step
+            )
+            if abs(q_val_quad_neg) > abs(q_val_neg):
+                alpha_neg = alpha_quad_neg
+                q_val_neg = q_val_quad_neg
+        if abs(q_val_pos) >= abs(q_val_neg) and abs(q_val_pos) > abs(q_val):
+            step = np.clip(alpha_pos * xpt[:, k], xl, xu)
+            q_val = q_val_pos
+        elif abs(q_val_neg) > abs(q_val_pos) and abs(q_val_neg) > abs(q_val):
+            step = np.clip(alpha_neg * xpt[:, k], xl, xu)
+            q_val = q_val_neg
+
+    if debug:
+        assert np.all(xl <= step)
+        assert np.all(step <= xu)
+        assert np.linalg.norm(step) < 1.1 * delta
+    return step
+
+
+def _cauchy_geom(const, grad, curv, xl, xu, delta, debug):
+    """
+    Same as `bound_constrained_cauchy_step` without the absolute value.
+    """
+    # Calculate the initial active set.
+    fixed_xl = (xl < 0.0) & (grad > 0.0)
+    fixed_xu = (xu > 0.0) & (grad < 0.0)
+
+    # Calculate the Cauchy step.
+    cauchy_step = np.zeros_like(grad)
+    cauchy_step[fixed_xl] = xl[fixed_xl]
+    cauchy_step[fixed_xu] = xu[fixed_xu]
+    if np.linalg.norm(cauchy_step) > delta:
+        working = fixed_xl | fixed_xu
+        while True:
+            # Calculate the Cauchy step for the directions in the working set.
+            g_norm = np.linalg.norm(grad[working])
+            delta_reduced = np.sqrt(
+                delta**2.0 - cauchy_step[~working] @ cauchy_step[~working]
+            )
+            if g_norm > TINY * abs(delta_reduced):
+                mu = max(delta_reduced / g_norm, 0.0)
+            else:
+                break
+            cauchy_step[working] = mu * grad[working]
+
+            # Update the working set.
+            fixed_xl = working & (cauchy_step < xl)
+            fixed_xu = working & (cauchy_step > xu)
+            if not np.any(fixed_xl) and not np.any(fixed_xu):
+                # Stop the calculations as the Cauchy step is now feasible.
+                break
+            cauchy_step[fixed_xl] = xl[fixed_xl]
+            cauchy_step[fixed_xu] = xu[fixed_xu]
+            working = working & ~(fixed_xl | fixed_xu)
+
+    # Calculate the step that maximizes the quadratic along the Cauchy step.
+    grad_step = grad @ cauchy_step
+    if grad_step >= 0.0:
+        # Set alpha_tr to the step size for the trust-region constraint.
+        s_norm = np.linalg.norm(cauchy_step)
+        if s_norm > TINY * delta:
+            alpha_tr = max(delta / s_norm, 0.0)
+        else:
+            # The Cauchy step is basically zero.
+            alpha_tr = 0.0
+
+        # Set alpha_quad to the step size for the maximization problem.
+        curv_step = curv(cauchy_step)
+        if curv_step < -TINY * grad_step:
+            alpha_quad = max(-grad_step / curv_step, 0.0)
+        else:
+            alpha_quad = np.inf
+
+        # Set alpha_bd to the step size for the bound constraints.
+        i_xl = (xl > -np.inf) & (cauchy_step < TINY * xl)
+        i_xu = (xu < np.inf) & (cauchy_step > TINY * xu)
+        alpha_xl = np.min(xl[i_xl] / cauchy_step[i_xl], initial=np.inf)
+        alpha_xu = np.min(xu[i_xu] / cauchy_step[i_xu], initial=np.inf)
+        alpha_bd = min(alpha_xl, alpha_xu)
+
+        # Calculate the solution and the corresponding function value.
+        alpha = min(alpha_tr, alpha_quad, alpha_bd)
+        step = np.clip(alpha * cauchy_step, xl, xu)
+        q_val = const + alpha * grad_step + 0.5 * alpha**2.0 * curv_step
+    else:
+        # This case is never reached in exact arithmetic. It prevents this
+        # function to return a step that decreases the objective function.
+        step = np.zeros_like(grad)
+        q_val = const
+
+    if debug:
+        assert np.all(xl <= step)
+        assert np.all(step <= xu)
+        assert np.linalg.norm(step) < 1.1 * delta
+    return step, q_val

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/utils/__init__.py

@@ -0,0 +1,18 @@
+from .exceptions import (
+    MaxEvalError,
+    TargetSuccess,
+    CallbackSuccess,
+    FeasibleSuccess,
+)
+from .math import get_arrays_tol, exact_1d_array
+from .versions import show_versions
+
+__all__ = [
+    "MaxEvalError",
+    "TargetSuccess",
+    "CallbackSuccess",
+    "FeasibleSuccess",
+    "get_arrays_tol",
+    "exact_1d_array",
+    "show_versions",
+]

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/utils/exceptions.py

@@ -0,0 +1,22 @@
+class MaxEvalError(Exception):
+    """
+    Exception raised when the maximum number of evaluations is reached.
+    """
+
+
+class TargetSuccess(Exception):
+    """
+    Exception raised when the target value is reached.
+    """
+
+
+class CallbackSuccess(StopIteration):
+    """
+    Exception raised when the callback function raises a ``StopIteration``.
+    """
+
+
+class FeasibleSuccess(Exception):
+    """
+    Exception raised when a feasible point of a feasible problem is found.
+    """

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/utils/math.py

@@ -0,0 +1,77 @@
+import numpy as np
+
+
+EPS = np.finfo(float).eps
+
+
+def get_arrays_tol(*arrays):
+    """
+    Get a relative tolerance for a set of arrays.
+
+    Parameters
+    ----------
+    *arrays: tuple
+        Set of `numpy.ndarray` to get the tolerance for.
+
+    Returns
+    -------
+    float
+        Relative tolerance for the set of arrays.
+
+    Raises
+    ------
+    ValueError
+        If no array is provided.
+    """
+    if len(arrays) == 0:
+        raise ValueError("At least one array must be provided.")
+    size = max(array.size for array in arrays)
+    weight = max(
+        np.max(np.abs(array[np.isfinite(array)]), initial=1.0)
+        for array in arrays
+    )
+    return 10.0 * EPS * max(size, 1.0) * weight
+
+
+def exact_1d_array(x, message):
+    """
+    Preprocess a 1-dimensional array.
+
+    Parameters
+    ----------
+    x : array_like
+        Array to be preprocessed.
+    message : str
+        Error message if `x` cannot be interpreter as a 1-dimensional array.
+
+    Returns
+    -------
+    `numpy.ndarray`
+        Preprocessed array.
+    """
+    x = np.atleast_1d(np.squeeze(x)).astype(float)
+    if x.ndim != 1:
+        raise ValueError(message)
+    return x
+
+
+def exact_2d_array(x, message):
+    """
+    Preprocess a 2-dimensional array.
+
+    Parameters
+    ----------
+    x : array_like
+        Array to be preprocessed.
+    message : str
+        Error message if `x` cannot be interpreter as a 2-dimensional array.
+
+    Returns
+    -------
+    `numpy.ndarray`
+        Preprocessed array.
+    """
+    x = np.atleast_2d(x).astype(float)
+    if x.ndim != 2:
+        raise ValueError(message)
+    return x

linedance-app/venv/lib/python3.12/site-packages/scipy/_lib/cobyqa/utils/versions.py

@@ -0,0 +1,67 @@
+import os
+import platform
+import sys
+from importlib.metadata import PackageNotFoundError, version
+
+
+def _get_sys_info():
+    """
+    Get useful system information.
+
+    Returns
+    -------
+    dict
+        Useful system information.
+    """
+    return {
+        "python": sys.version.replace(os.linesep, " "),
+        "executable": sys.executable,
+        "machine": platform.platform(),
+    }
+
+
+def _get_deps_info():
+    """
+    Get the versions of the dependencies.
+
+    Returns
+    -------
+    dict
+        Versions of the dependencies.
+    """
+    deps = ["cobyqa", "numpy", "scipy", "setuptools", "pip"]
+    deps_info = {}
+    for module in deps:
+        try:
+            deps_info[module] = version(module)
+        except PackageNotFoundError:
+            deps_info[module] = None
+    return deps_info
+
+
+def show_versions():
+    """
+    Display useful system and dependencies information.
+
+    When reporting issues, please include this information.
+    """
+    print("System settings")
+    print("---------------")
+    sys_info = _get_sys_info()
+    print(
+        "\n".join(
+            f"{k:>{max(map(len, sys_info.keys())) + 1}}: {v}"
+            for k, v in sys_info.items()
+        )
+    )
+
+    print()
+    print("Python dependencies")
+    print("-------------------")
+    deps_info = _get_deps_info()
+    print(
+        "\n".join(
+            f"{k:>{max(map(len, deps_info.keys())) + 1}}: {v}"
+            for k, v in deps_info.items()
+        )
+    )