Source code for jexplore.sampling.base

"""
This module provides the definitions of the base
:py:attr:`jexplore.sampling.state.Sampling` class.
"""

from dataclasses import dataclass

import numpy as np

from .space import Box, Space




@dataclass
class Sampling[Tspace: Space]:
    """Markov sampling defining parameters.

    In this base class version the parameter are the dimension of the
    target space and the number of chains. Child classes may specialize
    to specific type of markov sampling (e.g.
    :py:attr:`jexplore.sampling.mh.SamplingMH`)

    :param dim: dimension of the target space.
    :param nchains: number of chains.
    :param space: :py:attr:`jexplore.sampling.space` object
                  describing the target space.

    the state will then be defined by a (nchain, dim) point"""

    dim: int
    """Dimension of the target space"""

    nchain: int
    """Number of chains"""

    space: Tspace | Box
    """Target space."""

    def __init__(
        self, nchain: int, dim: int | None = None, space: Tspace | None = None
    ):
        self.nchain = nchain

        if space is not None:
            self.space = space
            self.dim = self.space.dim

        elif dim is not None:
            self.dim = dim
            self.space = Box(dim=dim)

        else:
            raise ValueError(
                "You need at least to define the dimension of the chains space."
            )



    def to_backend(self) -> dict:
        """Save the attributes of the object in
        backend format."""
        return {
            "nchain": np.array(self.nchain),
            "dim": np.array(self.dim),
            "class": f"{self.__class__.__module__}.{self.__class__.__qualname__}",
        }