prepare.py

import re
from datasets import load_dataset, Dataset, DatasetDict


def main():
    datasets = load_dataset('openlifescienceai/medmcqa')

    _datasets = {}
    for split in ['train','validation']:
        data = []
        for example in datasets[split]:
            question = str(example['question'])
            choices = [example['opa'], example['opb'], example['opc'], example['opd']]
            answer_index = example['cop']
            answer = chr(ord('A') + answer_index)
            data.append({
                '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'], 
    })
    datasets.push_to_hub('extraordinarylab/medmcqa')


if __name__ == '__main__':
    main()