def calculate_merged_string(string1, string2):
merged_templates = Template.merge_templates_wagner_fischer(
Template.from_string(string1, slot_token="[SLOT]"),
Template.from_string(string2, slot_token="[SLOT]"),
allow_longer_template=False
)
return next(merged_templates).to_flat_string(detokenizer=lambda x: " ".join(x))
def _merge_templates(t1: Template,
t2: Template,
minimal_variables: bool,
allow_empty_string=True) -> Template:
return next(
Template.merge_templates_wagner_fischer(
t1,
t2,
minimal_variables=minimal_variables,
allow_empty_string=allow_empty_string,
))
def _get_any_merge_candidate(self, t1: Template,
t2: Template) -> MergeCandidate:
""" Unused version of _get_best_merge_candidate, but might be prefered for performance gains """
max_length = max(t1.get_number_of_non_slots(),
t2.get_number_of_non_slots())
min_slots = min(t1.get_number_of_slots(), t2.get_number_of_slots())
merged_template = next(
Template.merge_templates_wagner_fischer(
t1, t2, minimal_variables=self._minimal_variables))
return MergeCandidate(
t1, t2,
_get_distance_of_merged(merged_template, max_length, min_slots))
def _get_best_merge_candidate(self, t1: Template,
t2: Template) -> MergeCandidate:
"""
Calculates the distance between two given templates, that can contain slots
"""
max_length = max(t1.get_number_of_non_slots(),
t2.get_number_of_non_slots())
min_slots = min(t1.get_number_of_slots(), t2.get_number_of_slots())
merged_templates = set(
Template.merge_templates_wagner_fischer(
t1, t2, minimal_variables=self._minimal_variables))
merge_candidates = []
for merged_template in merged_templates:
distance = _get_distance_of_merged(merged_template, max_length,
min_slots)
merge_candidates.append(
MergeCandidate(t1, t2, distance, merged=merged_template))
return min(merge_candidates)
def _merge_templates(t1: Template, t2: Template,
minimal_variables: bool) -> Template:
return next(
Template.merge_templates_wagner_fischer(
t1, t2, minimal_variables=minimal_variables))