Convert Lambda / AWS Stream Dictionary to Regular Dictionary

recursive_get_dict.py

from decimal import *


def recursive_get_dict(my_dict, use_decimal=True):
    """
    :param my_dict: The dict to be converted, which is assumed to come from the record from the DynamoDB stream
    :param use_decimal: Whether or not to use Decimal or int for the type for the number
    :return: A dictionary converted from this format {"username" : {"S" : "beto"}} to this format {"username": "beto"}
    """
    if type(my_dict) is not dict:
        return my_dict
    if my_dict.get('L', None) is not None:
        my_dict = list(my_dict['L'])
        new_list = []
        for item in my_dict:
            new_list.append(recursive_get_dict(item))
        return new_list
    elif my_dict.get('BOOL', None) is not None:
        return bool(my_dict['BOOL'])
    elif my_dict.get('S', None) is not None:
        return str(my_dict['S'])
    elif my_dict.get('SS', None) is not None:
        return set(my_dict['SS'])
    elif my_dict.get('NS', None) is not None:
        return set(my_dict['NS'])
    elif my_dict.get('N', None) is not None:
        return Decimal(my_dict['N']) if use_decimal is True else int(my_dict['N'])
    elif my_dict.get('M', None) is not None or type(my_dict) is dict:
        my_dict = my_dict.get('M', my_dict)
        new_dict = {}
        for key in my_dict.keys():
            new_dict[key] = recursive_get_dict(my_dict[key])
        return new_dict
    return {}

In AWS, sometimes streams are formatted to indicate the type.

e.g. New and old item streams from DynamoDB to Lambda will be formatted like {"username" : {"S" : "beto"}} and {"count" : {"N" : "0"}}. This removes the type strings and converts the dictionaries to dictionaries without the type strings: in this case, {"username" : "beto"} and {"count" : "0"}, respectively.