import re
from datasets import load_dataset, Dataset, DatasetDict


def main():
    for config in ['high', 'middle']:
        datasets = load_dataset('ehovy/race', config)

        _datasets = {}
        for split in ['train', 'validation', 'test']:
            data = []
            for example in datasets[split]:
                context = str(example['article'])
                question = str(example['question'])
                choices = example['options']
                answer = example['answer']
                answer_index = ord(answer) - ord('A')
                data.append({
                    'context': context.strip(), 
                    'question': question.strip(), 
                    'choices': choices, 
                    'answer': answer.strip(), 
                    'answer_index': answer_index
                })
            dataset = Dataset.from_list(data)
            _datasets[split] = dataset

        datasets = DatasetDict({
            'train': _datasets['train'], 
            'validation': _datasets['validation'], 
            'test': _datasets['test'], 
        })
        datasets.push_to_hub('extraordinarylab/race', config)


if __name__ == '__main__':
    main()