Logging dictionaries with datetime objects without Serialization error

Question

I'm trying to log python dictionary as JSON like so logging.info(dict(msg="foo", data=obj)) . However in some cases, obj may contain datetime objects and in those cases I get a JSON serialization error.

Is there a simple way to log dictionaries where objects (such as datetime objects) are serialized automatically, such as having it call the "str" function on these objects?

Answer 1

You may define your own serializer for datetime object:

def to_json(python_object):
    if isinstance(python_object, datetime.datetime):
        return {'__class__': 'datetime.datetime',
                '__value__': time.asctime(python_object.timetuple())}
    if isinstance(python_object, time.struct_time):
        return {'__class__': 'time.asctime',
                '__value__': time.asctime(python_object)}
    if isinstance(python_object, bytes):
        return {'__class__': 'bytes',
                '__value__': list(python_object)}
    raise TypeError(repr(python_object) + ' is not JSON serializable')

def from_json(json_object):
    if '__class__' in json_object:
        if json_object['__class__'] == 'datetime.datetime':
            return datetime.datetime.fromtimestamp(time.mktime(time.strptime(json_object['__value__'])))
        if json_object['__class__'] == 'time.asctime':
            return time.strptime(json_object['__value__'])
        if json_object['__class__'] == 'bytes':
            return bytes(json_object['__value__'])
    return json_object

and then call it this way:

json.dump(entry, f, default=to_json)

or for reading back

entry = json.load(f, object_hook=from_json)