This module defines the PartitionProposal class, which implements various operations
for proposing modifications to partitions in a structure learning context.
 
Classes:
    PartitionProposal: Implements operations such as swapping, splitting, merging,
    and moving nodes between partitions.
 
Methods:
    propose: Proposes a new state by applying one of the defined operations.
    compute_acceptance_ratio: Computes the acceptance ratio for the proposed state.
    get_nodes_to_rescore: Returns the nodes that need to be rescored.
    accept: Accepts the proposed state and updates the current state.
    compute_neighborhoods: Computes the number of neighborhoods for a given state.
    _calculate_move_probs: Calculates probabilities for different moves.
    _swap_adjacent: Swaps nodes between adjacent partitions.
    _swap_global: Swaps nodes between non-adjacent partitions.
    _split_or_merge_move: Performs either a split or merge operation.
    _split_move: Splits a partition into two.
    _join_move: Merges two adjacent partitions.
    _node_to_existing_partition: Moves a node to an existing partition.
    _node_to_new_partition: Moves a node to a new partition.
    _move_node_to_existing_or_new_partition: Moves a node to either an existing or new partition.
    _compute_neighborhoods_new_existing_partition: Computes neighborhoods for new and existing partitions.
    _calculate_join_possibilities: Calculates join possibilities for nodes.
    _calculate_partition_transitions: Calculates partition transition possibilities.
    _possible_splits_joins: Calculates possible splits and joins.
    _possible_permutations_neighbors: Calculates permutations for neighboring partitions.
    _possible_permutations: Calculates permutations for all partitions.

Modules
		math numpy matplotlib.patches matplotlib.pyplot seaborn

Classes
		structure_learning.proposals.proposal.StructureLearningProposal(abc.ABC) PartitionProposal   class PartitionProposal(structure_learning.proposals.proposal.StructureLearningProposal)     PartitionProposal(initial_state: structure_learning.data_structures.partition.OrderedPartition, whitelist: numpy.ndarray = None, blacklist: numpy.ndarray = None, seed=32)       Method resolution order: PartitionProposal structure_learning.proposals.proposal.StructureLearningProposal abc.ABC builtins.object Methods defined here: __init__(self, initial_state: structure_learning.data_structures.partition.OrderedPartition, whitelist: numpy.ndarray = None, blacklist: numpy.ndarray = None, seed=32) Initialise StructureLearningProposal instance.   Parameters:     graph (networkx.DiGraph): graph     blacklist (numpy.ndarray): mask for edges to ignore in the proposal     whitelist (numpy.ndarray): mask for edges to include in the proposal accept(self) Accepts the proposed state and updates the current state. compute_acceptance_ratio(self, current_state_score, proposed_state_score, current_state_prior=0, proposed_state_prior=0) Computes the acceptance ratio for the proposed state.   Args:     current_state_score (float): The score of the current state.     proposed_state_score (float): The score of the proposed state.   Returns:     float: The acceptance ratio, capped at 0. compute_neighborhoods(self, state) Computes the number of neighborhoods of an ordered partiion by using the equation sum_{i=1}^{m} ( sum_{c=1}^{k_i-1} ( comb( k_i, c) ) ) + m - 1   Args:     state (OrderedPartition): The state for which neighborhoods are computed.   Returns:     tuple: A tuple containing the total number of neighborhoods and a list of combinations. get_nodes_to_rescore(self) Returns the nodes that need to be rescored.   Returns:     set: A set of nodes to be rescored. propose(self) Proposes a new state by applying one of the defined operations.   Returns:     tuple: A tuple containing the proposed state and the operation applied. Data and other attributes defined here: MERGE_PARTITIONS = 'merge_partitions' MOVE_NODE_TO_EXISTING_PARTITION = 'move_node_to_existing_partition' MOVE_NODE_TO_NEW_OR_EXISTING = 'move_node_to_new_or_existing' MOVE_NODE_TO_NEW_PARTITION = 'move_node_to_new_partition' SPLIT_OR_MERGE = 'split_or_merge' SPLIT_PARTITIONS = 'split_partitions' STAY_STILL = 'stay_still' SWAP_ADJACENT = 'swap_adjacent' SWAP_GLOBAL = 'swap_global' __abstractmethods__ = frozenset() operations = ['swap_adjacent', 'swap_global', 'split_or_merge', 'move_node_to_new_or_existing', 'stay_still'] Data descriptors inherited from structure_learning.proposals.proposal.StructureLearningProposal: __dict__ dictionary for instance variables __weakref__ list of weak references to the object Data and other attributes inherited from structure_learning.proposals.proposal.StructureLearningProposal: INITIAL = 'initial'

Data
		List = typing.List