Feature request: Support UUID data type

[havardthom]

Would love support for UUID data type. Current situation:

import pprint
import uuid

from deepdiff import DeepDiff

t1 = {"id": uuid.uuid4()}
t2 = {"id": uuid.uuid4()}

result = DeepDiff(t1, t2)

pprint.pprint(result)

{'values_changed': {"root['id'].int": {'new_value': 117850254076634266241219085205683222228,
                                       'old_value': 91358784609380289177662098655200634313}}}

This causes problems for me since an unexpected int is appended to path, and dumping result as json will crash due to Integer exceeds 64-bit range. Would be great if result could rather be something like:

{'values_changed': {"root['id']": {'new_value': UUID(...),
                                   'old_value': UUID(...)}}}

I've added exclude_types={UUID} to work around my problems for now.

Docs mentions support for custom objects, but I didn't find any examples

Supported data types
int, string, unicode, dictionary, list, tuple, set, frozenset, OrderedDict, NamedTuple, Numpy, custom objects and more!