def name_dupe_similarity(
cls,
a1_name,
a2_name,
word_index,
languages=None,
likely_dupe_threshold=DedupeResponse.default_name_dupe_threshold,
needs_review_threshold=DedupeResponse.default_name_review_threshold
):
a1_name_tokens = Name.content_tokens(a1_name, languages=languages)
a2_name_tokens = Name.content_tokens(a2_name, languages=languages)
if not a1_name_tokens or not a2_name_tokens:
return None, 0.0
a1_scores = cls.word_scores_normalized(a1_name_tokens, word_index)
a2_scores = cls.word_scores_normalized(a2_name_tokens, word_index)
return is_name_duplicate_fuzzy(
a1_name_tokens,
a1_scores,
a2_name_tokens,
a2_scores,
languages=languages,
likely_dupe_threshold=likely_dupe_threshold,
needs_review_threshold=needs_review_threshold)
def name_dupe_fuzzy(cls, a1_name_tokens, a1_scores_norm, a2_name_tokens, a2_scores_norm, languages=None, likely_dupe_threshold=DedupeResponse.default_name_dupe_threshold,
needs_review_threshold=DedupeResponse.default_name_review_threshold):
if not a1_name_tokens or not a2_name_tokens:
return None, 0.0
return is_name_duplicate_fuzzy(a1_name_tokens, a1_scores_norm, a2_name_tokens, a2_scores_norm, languages=languages,
likely_dupe_threshold=likely_dupe_threshold, needs_review_threshold=needs_review_threshold)
def name_dupe_similarity(
cls,
a1_name,
a2_name,
tfidf,
languages=None,
likely_dupe_threshold=DedupeResponse.default_name_dupe_threshold,
needs_review_threshold=DedupeResponse.default_name_review_threshold
):
a1_name_tokens = Name.content_tokens(a1_name)
a2_name_tokens = Name.content_tokens(a2_name)
if not a1_name_tokens or not a2_name_tokens:
return None, 0.0
a1_name_tokens, a1_tfidf_norm = list(
zip(*cls.tfidf_vector_normalized(a1_name_tokens, tfidf)))
a2_name_tokens, a2_tfidf_norm = list(
zip(*cls.tfidf_vector_normalized(a2_name_tokens, tfidf)))
return is_name_duplicate_fuzzy(
a1_name_tokens,
a1_tfidf_norm,
a2_name_tokens,
a2_tfidf_norm,
languages=languages,
likely_dupe_threshold=likely_dupe_threshold,
needs_review_threshold=needs_review_threshold)