def make_data()

in grok/data.py [0:0]

• 19 lines of code
• 7 McCabe index (conditional complexity)

    def make_data(cls, operator, operands=None, shuffle=True, seed=0) -> List[str]:
        operator, noise_level = cls._get_operator_and_noise_level(operator)
        assert operator in VALID_OPERATORS

        if operator not in ["sort", "reverse", "copy"]:
            data = cls._make_binary_operation_data(operator)
        else:
            data = cls._make_unary_operation_data(operator, operands)

        rng = np.random.RandomState(seed=seed)
        if shuffle:
            rng.shuffle(data)

        if noise_level > 0:
            random_answer_eqns = rng.choice(data, size=noise_level)
            random_answers = [
                random_eq.split(" = ")[1] for random_eq in random_answer_eqns
            ]
            for i in range(noise_level):
                data[i] = data[i].split(" = ")[0] + " = " + random_answers[i]

        data = [EOS_TOKEN + " " + eq + " " + EOS_TOKEN for eq in data]

        return data