30a84d87f185185a7256231097a461e2a255247d,dirty_cat/test/test_dirty_cat.py,,test_similarity_encoder,#,7

Before Change

    X = np.array(["aa", "aaa", "aaab"]).reshape(-1, 1)
    X_test = np.array([["aa", "aaa", "aaa", "aaab", "aaac"]]).reshape(-1, 1)
    model.fit(X)
    encoder = model.transform(X_test)
    ans = np.zeros((len(X_test), len(X)))
    for i, x_t in enumerate(X_test.reshape(-1)):
        for j, x in enumerate(X.reshape(-1)):
            ans[i, j] = lev.ratio(x_t, x)

After Change

    X = np.array(["aa", "aaa", "aaab"]).reshape(-1, 1)
    X_test = np.array([["aa", "aaa", "aaa", "aaab", "aaac"]]).reshape(-1, 1)

    similarity_types = [
        "levenshtein-ratio",
        "jaro-winkler"
        ]

    for similarity_type in similarity_types:
        model = similarity_encoder.SimilarityEncoder(
            similarity_type=similarity_type, handle_unknown="ignore")

        encoder = model.fit(X).transform(X_test)

        if similarity_type == "levenshtein-ratio":
            ans = np.zeros((len(X_test), len(X)))
            for i, x_t in enumerate(X_test.reshape(-1)):
                for j, x in enumerate(X.reshape(-1)):
                    ans[i, j] = lev.ratio(x_t, x)
            assert np.array_equal(encoder, ans)

        if similarity_type == "jaro-winkler":
            ans = np.zeros((len(X_test), len(X)))
            for i, x_t in enumerate(X_test.reshape(-1)):
                for j, x in enumerate(X.reshape(-1)):
                    ans[i, j] = jaro_distance(x_t, x)
            assert np.array_equal(encoder, ans)

Instances