def embeddings_cosine_sim(file, sp, df): print("Computing cosine similarity across image embeddings") create_embeddings(file[:file.rfind("/") + 1]) id_embeddings = utils.get_out_file(file) embeddings = id_embeddings["embedding"].tolist() print("Normalizing image embeddings") normalized_embeddings = normalize(embeddings) common.cosine_similarity(vectors=normalized_embeddings, df=df, sp=sp)
def similarity(other_post): return cosine_similarity(post.token_index, other_post.token_index)
def similarity(other_post): return cosine_similarity(this, __token_dist(other_post.title))
def bio_tfidf_cosine_sim(sp, df): vectors = common.tfidf(df, FEATURE) common.cosine_similarity(sp, df, vectors)
def text_tfidf_cosine_sim_length_constrained(sp, df, n): df[FEATURE] = common.constrain_length(df, FEATURE, n) vectors = common.tfidf(df, FEATURE) common.cosine_similarity(sp, df, vectors)