Function fast_deepcopy_dict

st2common/st2common/util/deep_copy.py:32–61

Perform a fast deep copy of the provided value. This function is designed primary to operate on values of a simple type (think JSON types - dicts, lists, arrays, strings, ints). It's up to 10x faster compared to copy.deepcopy(). In case the provided value contains non-simple

(value, fall_back_to_deepcopy=True)

def fast_deepcopy_dict(value, fall_back_to_deepcopy=True):
    """
    Perform a fast deep copy of the provided value.

    This function is designed primary to operate on values of a simple type (think JSON types -
    dicts, lists, arrays, strings, ints).

    It's up to 10x faster compared to copy.deepcopy().

    In case the provided value contains non-simple types, we simply fall back to "copy.deepcopy()".
    This means that we can still use it on values which sometimes, but not always contain complex
    types - in that case, when value doesn't contain complex types we will perform much faster copy
    and when it does, we will simply fall back to copy.deepcopy().

    :param fall_back_to_deepcopy: True to fall back to copy.deepcopy() in case we fail to fast deep
                                  copy the value because it contains complex types or similar
    :type fall_back_to_deepcopy: ``bool``
    """
    # NOTE: ujson / orjson round-trip is up to 10 times faster on smaller and larger dicts compared
    # to copy.deepcopy(), but it has some edge cases with non-simple types such as datetimes, class
    # instances, etc.
    try:
        value = orjson.loads(orjson.dumps(value, default=default))
    except (OverflowError, ValueError, TypeError) as e:
        if not fall_back_to_deepcopy:
            raise e

        value = copy.deepcopy(value)

    return value