preprocess/ade_dosage.py [17:36]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        self.task_type = "text to text"
        self.license = "unknown"

    def get_train_test_lines(self, dataset):
        # only train set, manually split 20% data as test
        lines = self.map_hf_dataset_to_list(dataset, "train")

        np.random.seed(42)
        np.random.shuffle(lines)
        
        n = len(lines)

        train_lines = lines[:int(0.8*n)]
        test_lines = lines[int(0.8*n):]

        return train_lines, test_lines

    def map_hf_dataset_to_list(self, hf_dataset, split_name):
        lines = []
        for datapoint in hf_dataset[split_name]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



preprocess/numer_sense.py [17:37]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        self.task_type = "text to text"
        self.license = "unknown"

    def get_train_test_lines(self, dataset):

        lines = self.map_hf_dataset_to_list(dataset, "train")

        np.random.seed(42)
        np.random.shuffle(lines)
        
        n = len(lines)

        train_lines = lines[:int(0.8*n)]
        test_lines = lines[int(0.8*n):]

        return train_lines, test_lines


    def map_hf_dataset_to_list(self, hf_dataset, split_name):
        lines = []
        for datapoint in hf_dataset[split_name]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -