Source code for textattack.search_methods.population_based_search

"""
Population based Search
==========================
"""

from abc import ABC, abstractmethod

from textattack.search_methods import SearchMethod


[docs]class PopulationBasedSearch(SearchMethod, ABC): """Abstract base class for population-based search methods. Examples include: genetic algorithm, particle swarm optimization """ def _check_constraints(self, transformed_text, current_text, original_text): """Check if `transformted_text` still passes the constraints with respect to `current_text` and `original_text`. This method is required because of a lot population-based methods does their own transformations apart from the actual `transformation`. Examples include `crossover` from `GeneticAlgorithm` and `move` from `ParticleSwarmOptimization`. Args: transformed_text (AttackedText): Resulting text after transformation current_text (AttackedText): Recent text from which `transformed_text` was produced from. original_text (AttackedText): Original text Returns `True` if constraints satisfied and `False` if otherwise. """ filtered = self.filter_transformations( [transformed_text], current_text, original_text=original_text ) return True if filtered else False @abstractmethod def _perturb(self, pop_member, original_result, **kwargs): """Perturb `pop_member` in-place. Must be overridden by specific population-based method Args: pop_member (PopulationMember): Population member to perturb\ original_result (GoalFunctionResult): Result for original text. Often needed for constraint checking. Returns `True` if perturbation occured. `False` if not. """ raise NotImplementedError() @abstractmethod def _initialize_population(self, initial_result, pop_size): """ Initialize a population of size `pop_size` with `initial_result` Args: initial_result (GoalFunctionResult): Original text pop_size (int): size of population Returns: population as `list[PopulationMember]` """ raise NotImplementedError
[docs]class PopulationMember: """Represent a single member of population.""" def __init__(self, attacked_text, result=None, attributes={}, **kwargs): self.attacked_text = attacked_text self.result = result self.attributes = attributes for key, value in kwargs.items(): setattr(self, key, value) @property def score(self): if not self.result: raise ValueError( "Result must be obtained for PopulationMember to get its score." ) return self.result.score @property def words(self): return self.attacked_text.words @property def num_words(self): return self.attacked_text.num_words