Source code for pina.domain.intersection

"""Module for the Intersection operation."""

import random
from .base_operation import BaseOperation
from ..label_tensor import LabelTensor
from ..utils import check_consistency




class Intersection(BaseOperation):
    r"""
    Implementation of the intersection operation defined on a list of domains.

    Given multiple sets :math:`A_1, A_2, \ldots, A_n`, define their intersection
    as:

    .. math::

        \bigcap_{i=1}^{n} A_i = \{x \mid x \in A_i \forall i\}

    No check is performed to ensure that the resulting domain is non-empty.

    :Example:

        >>> cartesian1 = CartesianDomain({'x': [0, 1], 'y': [0, 1]})
        >>> cartesian2 = CartesianDomain({'x': [0, 1], 'y': [0.5, 1.5]})
        >>> intersection = Intersection([cartesian1, cartesian2])
    """



    def is_inside(self, point, check_border=False):
        """
        Check if a point is inside the intersection of the domains.

        :param LabelTensor point: The point to check.
        :param bool check_border: If ``True``, the boundary is considered inside
            the domain. Default is ``False``.
        :raises ValueError: If ``point`` is not a :class:`LabelTensor`.
        :raises ValueError: If the labels of ``point`` differ from the variables
            of the domain.
        :return: Whether the point is inside the domain or not.
        :rtype: bool
        """
        # Checks on point
        check_consistency(point, LabelTensor)
        if set(self.variables) != set(point.labels):
            raise ValueError(
                "Point labels differ from domain's dictionary labels. "
                f"Got {sorted(point.labels)}, expected {self.variables}."
            )

        return all(g.is_inside(point, check_border) for g in self.geometries)




    def sample(self, n, mode="random", variables="all"):
        """
        The sampling routine.

        .. note::

            This sampling method relies on rejection sampling. Points are drawn
            from the individual geometries, and only those that lie exclusively
            within one geometry are kept. When the exclusion domain is small
            relative to the combined area of the input domains, the method may
            become highly inefficient.

        :param int n: The number of samples to generate.
        :param str mode: The sampling method. Default is ``random``.
        :param variables: The list of variables to sample. If ``all``, all
            variables are sampled. Default is ``all``.
        :type variables: list[str] | str
        :raises AssertionError: If ``n`` is not a positive integer.
        :raises ValueError: If the sampling mode is invalid.
        :raises ValueError: If ``variables`` is neither ``all``, a string, nor a
            list/tuple of strings.
        :raises ValueError: If any of the specified variables is unknown.
        :return: The sampled points.
        :rtype: LabelTensor
        """
        # Validate sampling settings
        variables = self._validate_sampling(n, mode, variables)

        # Compute number of points per geometry and remainder
        num_pts, remainder = divmod(n, len(self.geometries))

        # Shuffle indices
        shuffled_geometries = random.sample(
            range(len(self.geometries)), len(self.geometries)
        )

        # Precompute per-geometry allocations following the shuffled order
        alloc = [num_pts + (i < remainder) for i in range(len(self.geometries))]
        samples = []

        # Iterate over geometries in shuffled order
        for idx, gi in enumerate(shuffled_geometries):

            # If no points to allocate (possible if len(self.geometries) > n)
            if alloc[idx] == 0:
                continue

            # Sampled points for the current geometry
            sampled_points = []

            # Sample until we have enough points
            while len(sampled_points) < alloc[idx]:

                # Sample a sufficiently large number of points
                batch_size = 2 * (alloc[idx] - len(sampled_points))
                pts = self.geometries[gi].sample(batch_size, mode)

                # Filter points inside the intersection
                for p in pts:
                    p = p.reshape(1, -1)
                    p.labels = pts.labels
                    if self.is_inside(p):
                        sampled_points.append(p[variables])
                        if len(sampled_points) >= alloc[idx]:
                            break

            # Sample points
            samples.append(LabelTensor.cat(sampled_points, dim=0))

        return LabelTensor.cat(samples, dim=0)




    def partial(self):
        """
        Return the boundary of the domain resulting from the operation.

        :raises NotImplementedError: The :meth:`partial` method is not
            implemented for intersection domains. Please operate on the
            individual domains instead.
        """
        raise NotImplementedError(
            "The partial method is not implemented for intersection domains. "
            "Please operate on the individual domains instead."
        )