d3ec733c425e7d9d066380b95e04605843339bd0,dirty_cat/similarity_encoder.py,,get_prototype_sorted_by_frequencies,#,47

Before Change

    Returns a numpy array of the values without their frequencies
    
    uniques, counts = np.unique(prototypes, return_counts=True)
    frequencies = list(zip(uniques, counts))
    frequencies, _ = sorted(frequencies, key=lambda elt: elt[1], reverse=True)
    return np.array([f for f in frequencies])


_VECTORIZED_EDIT_DISTANCES = {
    "levenshtein-ratio": np.vectorize(string_distances.levenshtein_ratio),

After Change

    Returns a numpy array of the values without their frequencies
    
    uniques, counts = np.unique(prototypes, return_counts=True)
    sorted_indexes = np.argsort(counts)[::-1]
    return uniques[sorted_indexes], counts[sorted_indexes]


_VECTORIZED_EDIT_DISTANCES = {
    "levenshtein-ratio": np.vectorize(string_distances.levenshtein_ratio),

Instances