def test_fixed_named_type():
"""https://github.com/fastavro/fastavro/issues/450"""
schema = {
"type":
"record",
"name":
"test_fixed_named_type",
"fields": [
{
"name": "test1",
"type": {
"type": "fixed",
"name": "my_fixed",
"size": 4,
},
},
{
"name": "test2",
"type": "my_fixed",
},
],
}
record = {"test1": b"1234", "test2": b"4321"}
parsed_schema = parse_schema(schema)
validate(record, parsed_schema)
def test_no_data(tmpdir):
filepath = tmpdir + "no_data.avro"
schema = {
"name":
"Weather",
"type":
"record",
"fields": [
{
"name": "station",
"type": "string"
},
{
"name": "time",
"type": "long"
},
{
"name": "temp",
"type": "int"
},
],
}
parsed_schema = fa.parse_schema(schema)
with open(filepath, "wb") as out:
fa.writer(out, parsed_schema, [])
df = cudf.read_avro(filepath)
# fastavro returns an empty dataframe, need to verify manually
assert_eq(df.shape, (0, 3))
dtypes = df.dtypes.values.tolist()
assert_eq(dtypes, [np.dtype("O"), np.dtype("int64"), np.dtype("int32")])
col_names = df.columns.tolist()
assert_eq(col_names, ["station", "time", "temp"])
class ClickEvent:
email: str = field(default_factory=faker.email)
timestamp: str = field(default_factory=faker.iso8601)
uri: str = field(default_factory=faker.uri)
number: int = field(default_factory=lambda: random.randint(0, 999))
attributes: dict = field(default_factory=ClickAttribute.attributes)
schema = parse_schema({
"type":
"record",
"name":
"click_event",
"namespace":
"com.udacity.lesson3.exercise2",
"fields": [
{
"name": "email",
"type": "string"
},
{
"name": "timestamp",
"type": "string"
},
{
"name": "uri",
"type": "string"
},
{
"name": "number",
"type": "int"
},
#
# TODO: Add the attributes map!
{
"name": "attributes",
"type": {
"type": "map",
"values": {
"type":
"record",
"name":
"attribute",
"fields": [{
"name": "element",
"type": "string"
}, {
"name": "content",
"type": "string"
}]
}
}
}
]
})
def serialize(self):
"""Serializes the ClickEvent for sending to Kafka"""
out = BytesIO()
writer(out, ClickEvent.schema, [asdict(self)])
return out.getvalue()
def test_enum_named_type():
"""https://github.com/fastavro/fastavro/issues/450"""
schema = {
"type":
"record",
"name":
"test_enum_named_type",
"fields": [
{
"name": "test1",
"type": {
"type": "enum",
"name": "my_enum",
"symbols": ["FOO", "BAR"],
},
},
{
"name": "test2",
"type": "my_enum",
},
],
}
record = {"test1": "FOO", "test2": "BAR"}
parsed_schema = parse_schema(schema)
validate(record, parsed_schema)
def write_avro():
schema = {
"namespace": "sample.avro",
"type": "record",
"name": "Cars",
"fields": [
{"name": "model", "type": "string"},
{"name": "make", "type": ["string", "null"]},
{"name": "year", "type": ["int", "null"]}
]
}
records = [
{"model": "ABC", "make": "Audi", "year": 2010},
{"model": "ABC", "make": "Audi", "year": 2010},
{"model": "ABC", "make": "Audi", "year": 2010},
{"model": "ABC", "make": "Audi", "year": 2010},
{"model": "ABC", "make": "Audi", "year": 2010},
{"model": "EDF", "make": "BEWM"},
{"model": "EDF", "make": "BEWM"},
{"model": "EDF", "make": "BEWM"},
{"model": "EDF", "make": "BEWM"},
{"model": "EDF", "make": "BEWM"},
{"model": "XCV", "year": 2010},
{"model": "XCV", "year": 2010},
{"model": "XCV", "year": 2010},
{"model": "XCV", "year": 2010},
{"model": "XCV", "year": 2010}
]
os.makedirs(os.path.join('.', 'data'), exist_ok=True)
with open(os.path.join('.', 'data', 'cars.avro'), 'wb') as avro_file:
writer(avro_file, parse_schema(schema), records)
with open(os.path.join('.', 'data', 'cars.json'), 'w') as json_file:
json.dump(records, json_file)
def encode_into_avro(alert: dict, schema_file: str) -> str:
"""Encode a dict record into avro bytes
Parameters
----------
alert: dict
A Dictionary of alert data
schema_file: str
Path of avro schema file
Returns
----------
value: str
a bytes string with avro encoded alert data
Examples
----------
>>> r = AlertReader(avro_single_alert)
>>> alert = r.to_list(size=1)[0]
>>> avro_encoded = encode_into_avro(alert, schema_path)
"""
with open(schema_file) as f:
schema = json.load(f)
parsed_schema = fastavro.parse_schema(schema)
b = io.BytesIO()
fastavro.schemaless_writer(b, parsed_schema, alert)
return b.getvalue()
def load_schema(name, suffix=""):
"""Load an Avro schema from the local app data.
This function is memoized so that repeated calls are fast.
Parameters
----------
name : `str`
Name of the schema. This should be a fully-qualified name that matches
the file name of schemas in the ``schemas/`` directory of the source
repository.
suffix : `str`, optional
A suffix to add to the schema's name. This is typically used to create
"staging" schemas, therefore "staging subjects" in the Schema Registry.
Returns
-------
schema : `dict`
A schema object.
"""
schemas_dir = Path(__file__).parent / "schemas"
schema_path = schemas_dir / f"{name}.json"
schema = json.loads(schema_path.read_text())
if suffix:
schema["name"] = "".join((schema["name"], suffix))
return fastavro.parse_schema(schema)
class ClickEvent:
email: str = field(default_factory=faker.email)
timestamp: str = field(default_factory=faker.iso8601)
uri: str = field(default_factory=faker.uri)
number: int = field(default_factory=lambda: random.randint(0, 999))
#
# TODO: Define an Avro Schema for this ClickEvent
# See: https://avro.apache.org/docs/1.8.2/spec.html#schema_record
# See: https://fastavro.readthedocs.io/en/latest/schema.html?highlight=parse_schema#fastavro-schema
#
schema = parse_schema({
"type"
})
def serialize(self):
"""Serializes the ClickEvent for sending to Kafka"""
#
# TODO: Rewrite the serializer to send data in Avro format
# See: https://fastavro.readthedocs.io/en/latest/schema.html?highlight=parse_schema#fastavro-schema
#
# HINT: Python dataclasses provide an `asdict` method that can quickly transform this
# instance into a dictionary!
# See: https://docs.python.org/3/library/dataclasses.html#dataclasses.asdict
#
# HINT: Use BytesIO for your output buffer. Once you have an output buffer instance, call
# `getvalue() to retrieve the data inside the buffer.
# See: https://docs.python.org/3/library/io.html?highlight=bytesio#io.BytesIO
#
return json.dumps(
{"uri": self.uri, "timestamp": self.timestamp, "email": self.email}
)
def __write(self, data):
"""Write data into the avro file."""
if not self.__schema:
self.__schema = parse_schema(
self.__avro_translator.deduce_scheme(data[0]))
fast_writer(self.__descriptor, self.__schema, data, self.__codec)
def cudf_from_avro_util(schema, records):
schema = [] if schema is None else fastavro.parse_schema(schema)
buffer = io.BytesIO()
fastavro.writer(buffer, schema, records)
buffer.seek(0)
return cudf.read_avro(buffer)
def __init__(self, file_, data=None, schema=None, codec: str = 'snappy'):
"""Create an avro archive.
Note:
Specification: https://avro.apache.org/docs/1.8.2/spec.html
Args:
file_ (str, BytesIO, IOBase): the output file_
data (dict, list(dict)): the data to write
schema (dict): the avro schema as dictionary
Returns:
AvroDataFileWriter: the instance of this object
"""
self.__descriptor = None
if isinstance(file_, str):
self.__descriptor = open(file_, 'wb')
elif isinstance(file_, (BytesIO, IOBase)):
self.__descriptor = file_
else:
raise Exception("Type '{}' for file_ is not supported...".format(
type(file_)))
self.__schema = None
self.__codec = codec
self.__avro_translator = AvroObjectTranslator()
if schema:
self.__schema = parse_schema(schema)
if data is not None:
self.append(data)
def _write_inference_result(self, sample_ids, labels, weights, scores,
scores_and_offsets, task_index, schema_params, output_dir):
""" Write inference results. """
photon_ml_writer = PhotonMLWriter(schema_params=schema_params)
output_avro_schema = photon_ml_writer.get_inference_output_avro_schema(
self.metadata,
self._has_label(schema_params[constants.LABEL]),
True,
has_weight=self._has_feature(schema_params[constants.SAMPLE_WEIGHT]))
parsed_schema = parse_schema(output_avro_schema)
records = []
for rec_id, rec_label, rec_weight, rec_score, rec_score_and_offset in \
zip(sample_ids, labels, weights, scores, scores_and_offsets):
rec = {schema_params[constants.SAMPLE_ID]: int(rec_id),
schema_params[constants.PREDICTION_SCORE]: float(rec_score),
schema_params[constants.PREDICTION_SCORE_PER_COORDINATE]: float(rec_score_and_offset)
}
if self._has_label(schema_params[constants.LABEL]):
rec[schema_params[constants.LABEL]] = int(rec_label)
if self._has_feature(schema_params[constants.SAMPLE_WEIGHT]):
rec[schema_params[constants.SAMPLE_WEIGHT]] = int(rec_weight)
records.append(rec)
output_file = os.path.join(output_dir, "part-{0:05d}.avro".format(task_index))
error_msg = "worker {} encountered error in writing inference results".format(task_index)
with tf1.gfile.GFile(output_file, 'wb') as f:
try_write_avro_blocks(f, parsed_schema, records, None, error_msg)
logging("Worker {} saved inference result to {}".format(task_index, output_file))
def _write_toavro(table,
target,
mode,
schema,
sample,
codec='deflate',
compression_level=None,
**avro_args):
if table is None or len(table) <= 0:
return
# build a schema when not defined by user
if not schema:
schema, table2 = _build_schema_from_values(table, sample)
else:
table2 = _fix_missing_headers(table, schema)
# fastavro expects a iterator of dicts
rows = dicts(table2) if PY3 else _ordered_dict_iterator(table2)
with target.open(mode) as target_file:
# delay the import of fastavro for not breaking when unused
import fastavro
parsed_schema = fastavro.parse_schema(schema)
# this could raise a error when any value is not of supported tupe
fastavro.writer(fo=target_file,
schema=parsed_schema,
records=rows,
codec=codec,
codec_compression_level=compression_level,
**avro_args)
class ClickEvent:
email: str = field(default_factory=faker.email)
timestamp: str = field(default_factory=faker.iso8601)
uri: str = field(default_factory=faker.uri)
number: int = field(default_factory=lambda: random.randint(0, 999))
attributes: dict = field(default_factory=ClickAttribute.attributes)
#
# TODO: Update this Avro schema to include a map of attributes
# See: https://avro.apache.org/docs/1.8.2/spec.html#Maps
#
schema = parse_schema(
{
"type": "record",
"name": "click_event",
"namespace": "com.udacity.lesson3.exercise2",
"fields": [
{"name": "email", "type": "string"},
{"name": "timestamp", "type": "string"},
{"name": "uri", "type": "string"},
{"name": "number", "type": "int"}
#
# TODO: Add the attributes map!
#
],
}
)
def serialize(self):
"""Serializes the ClickEvent for sending to Kafka"""
out = BytesIO()
writer(out, ClickEvent.schema, [asdict(self)])
return out.getvalue()
class Purchase:
username: str = field(default_factory=faker.user_name)
currency: str = field(default_factory=faker.currency_code)
amount: int = field(default_factory=lambda: random.randint(100, 200000))
schema = parse_schema({
"type":
"record",
"name":
"purchase",
"namespace":
"com.udacity.lesson3.sample2",
"fields": [
{
"name": "username",
"type": "string"
},
{
"name": "currency",
"type": "string"
},
{
"name": "amount",
"type": "int"
},
],
})
def serialize(self):
#
# TODO: Modify the following sample to use Avro instead of JSON
#
out = io.BytesIO()
writer(out, Purchase.schema, [asdict(self)])
return out.getvalue()
def read_schema() -> fastavro.schema:
with open("avro/schema/schema-SKDB.public.sdcocdmst.json", "r") as file:
schema = json.load(file)
schema = json.loads(schema["schema"])
schema = fastavro.parse_schema(schema)
return schema
def fix_record_for_avro(record, avro_schema):
for field in avro_schema['fields']:
field_name = field['name']
datatype = field['type']
if isinstance(datatype, dict):
# This is a record type definition so we need to recurse a level deeper.
record[field_name] = fix_record_for_avro(
record[field_name], fastavro.parse_schema(datatype))[0]
elif isinstance(datatype, list) and isinstance(datatype[1], dict):
logical_type = datatype[1].get('logicalType', None)
if logical_type:
if logical_type.find('-') > -1:
logical_prefix, precision = logical_type.split('-')
else:
logical_prefix = logical_type
precision = None
if logical_prefix == 'timestamp':
is_micros = (precision == 'micros')
record[field_name] = datetime_to_epoch_timestamp(
record[field_name], micros=is_micros)
elif logical_type == 'date':
record[field_name] = date_to_epoch_date(record[field_name])
elif logical_prefix == 'time':
is_micros = (precision == 'micros')
record[field_name] = time_to_epoch_time(record[field_name],
micros=is_micros)
return [record]
def gen_schema():
schema = {
'doc':
'Twitter',
'name':
'Twitter',
'namespace':
'twitter',
'type':
'record',
'fields': [
{
'name': 'twitter_name',
'type': 'string'
},
{
'name': 'twitter_token',
'type': 'float'
},
{
'name': 'uuid',
'type': 'string'
},
],
}
parsed_schema = parse_schema(schema)
return parsed_schema
def test_serde_to_avro(pytestconfig: PytestConfig, json_filename: str) -> None:
# In this test, we want to read in from JSON -> MCE object.
# Next we serialize from MCE to Avro and then deserialize back to MCE.
# Finally, we want to compare the two MCE objects.
json_path = pytestconfig.rootpath / json_filename
mces = list(iterate_mce_file(str(json_path)))
# Serialize to Avro.
parsed_schema = fastavro.parse_schema(
json.loads(getMetadataChangeEventSchema()))
fo = io.BytesIO()
out_records = [mce.to_obj(tuples=True) for mce in mces]
fastavro.writer(fo, parsed_schema, out_records)
# Deserialized from Avro.
fo.seek(0)
in_records = list(fastavro.reader(fo, return_record_name=True))
in_mces = [
MetadataChangeEventClass.from_obj(record, tuples=True)
for record in in_records
]
# Check diff
assert len(mces) == len(in_mces)
for i in range(len(mces)):
assert mces[i] == in_mces[i]
class ClickEvent:
email: str = field(default_factory=faker.email)
timestamp: str = field(default_factory=faker.iso8601)
uri: str = field(default_factory=faker.uri)
number: int = field(default_factory=lambda: random.randint(0, 999))
schema = parse_schema({
"type":
"record",
"name":
"click_event",
"namespace":
"com.hcvn.bi",
"fields": [{
"name": "email",
"type": "string"
}, {
"name": "timestamp",
"type": "string"
}, {
"name": "uri",
"type": "string"
}, {
"name": "number",
"type": "int"
}]
})
def serialize(self):
out = BytesIO()
writer(out, ClickEvent.schema, [asdict(self)])
return out.getvalue()
def test_parse_schema_accepts_nested_records_from_arrays():
parsed_schema = fastavro.parse_schema({
"fields": [{
"type": {
"items": {
"type": "record",
"fields": [{
"type": "string",
"name": "text"
}],
"name": "Nested"
},
"type": "array",
},
"name": "multiple"
}, {
"type": {
"type": "array",
"items": "Nested"
},
"name": "single"
}],
"type":
"record",
"name":
"test_parse_schema_accepts_nested_records_from_arrays",
})
assert "Nested" == parsed_schema["fields"][1]["type"]["items"]
def _write_inference_result(self, sample_ids, labels, weights, prediction_score,
prediction_score_per_coordinate, task_index, schema_params: SchemaParams, output_dir):
""" Write inference results. """
output_avro_schema = get_inference_output_avro_schema(
self.metadata,
True,
schema_params,
has_weight=self._has_feature(schema_params.weight_column_name))
parsed_schema = parse_schema(output_avro_schema)
records = []
for rec_id, rec_label, rec_weight, rec_prediction_score, rec_prediction_score_per_coordinate in \
zip(sample_ids, labels, weights, prediction_score, prediction_score_per_coordinate):
rec = {schema_params.uid_column_name: int(rec_id),
schema_params.prediction_score_column_name: float(rec_prediction_score),
schema_params.prediction_score_per_coordinate_column_name: float(rec_prediction_score_per_coordinate)}
if self._has_label(schema_params.label_column_name):
rec[schema_params.label_column_name] = int(rec_label)
if self._has_feature(schema_params.weight_column_name):
rec[schema_params.weight_column_name] = int(rec_weight)
records.append(rec)
output_file = os.path.join(output_dir, f"part-{task_index:05d}.avro")
error_msg = f"worker {task_index} encountered error in writing inference results"
with tf1.gfile.GFile(output_file, 'wb') as f:
try_write_avro_blocks(f, parsed_schema, records, None, error_msg)
logging(f"Worker {task_index} saved inference result to {output_file}")
def test_parse_schema_resolves_references_from_unions():
parsed_schema = fastavro.parse_schema({
"namespace":
"com.other",
"name":
"Outer",
"type":
"record",
"fields": [
{
"name":
"a",
"type": [
"null", {
"type": "record",
"name": "Inner",
"fields": [{
"name": "the_thing",
"type": "string"
}]
}
]
},
{
"name": "b",
# This should resolve to com.example.Inner because of the
# `namespace` of the enclosing record.
"type": ["null", "Inner"]
}
]
})
assert "com.other.Inner" == parsed_schema["fields"][1]["type"][1]
def test_order_of_values_in_map():
"""https://github.com/fastavro/fastavro/issues/303"""
schema = {
'doc':
'A weather reading.',
'name':
'Weather',
'namespace':
'test',
'type':
'record',
'fields': [{
'name': 'metadata',
'type': {
'type':
'map',
'values': [{
'type': 'array',
'items': 'string'
}, {
'type': 'map',
'values': ['string']
}]
}
}],
}
parsed_schema = fastavro.parse_schema(schema)
records = [{'metadata': {'map1': {'map2': 'str'}}}]
assert records == roundtrip(parsed_schema, records)
def __init__(
self,
file_path_prefix=None,
location=None,
schema=parse_schema(KLIO_SCHEMA_OBJ),
codec="deflate",
file_name_suffix="",
num_shards=0,
shard_name_template=None,
mime_type="application/x-avro",
):
file_path = self._get_file_path(file_path_prefix, location)
super(KlioWriteToAvro, self).__init__(
file_path_prefix=file_path,
schema=schema,
codec=codec,
file_name_suffix=file_name_suffix,
num_shards=num_shards,
shard_name_template=shard_name_template,
mime_type=mime_type,
use_fastavro=True,
)
self._sink = _KlioFastAvroSink(
file_path,
schema,
codec,
file_name_suffix,
num_shards,
shard_name_template,
mime_type,
)
def test_parse_schema_accepts_nested_namespaces():
parsed_schema = fastavro.parse_schema({
"namespace":
"com.example",
"name":
"Outer",
"type":
"record",
"fields": [
{
"name": "a",
"type": {
"type": "record",
"name": "Inner",
"fields": [{
"name": "the_thing",
"type": "string"
}]
}
},
{
"name": "b",
# This should resolve to com.example.Inner because of the
# `namespace` of the enclosing record.
"type": "Inner"
},
{
"name": "b",
"type": "com.example.Inner"
}
]
})
assert "com.example.Inner" == parsed_schema["fields"][0]["type"]["name"]
assert "com.example.Inner" == parsed_schema["fields"][1]["type"]
def test_newer_versions_of_named_schemas_2():
"""https://github.com/fastavro/fastavro/issues/450"""
schema = {
"name":
"Weather",
"type":
"record",
"fields": [
{
"name": "place1",
"type": {
"name": "Location",
"type": "record",
"fields": [{
"name": "city",
"type": "string"
}],
},
},
{
"name": "place2",
"type": "Location",
},
],
}
example_1 = {"place1": {"city": "London"}, "place2": {"city": "Berlin"}}
parsed_schema = fastavro.parse_schema(schema)
assert example_1 == roundtrip(parsed_schema, example_1)
def main(args):
schema_row_size = struct.calcsize(args.format)
parsed_schema = fastavro.parse_schema(schema(args.format))
file_path = Path(args.bin_input)
file_size = file_path.stat().st_size
out_filename = (args.output if args.output else filename_from_args(
args, prefix='sima.exported.', postfix='.avro'))
if file_size % schema_row_size != 0:
msg = 'Format {} does not fit into file {}'.format(
args.format, args.bin_input)
raise ExportError(msg)
with open(str(file_path), 'rb') as binf, open(out_filename, 'wb') as avrf:
records = ({
"col_{}".format(i): v
for i, v in enumerate(struct.unpack(args.format, piece))
} for piece in read_in_chunks(binf, schema_row_size))
fastavro.writer(avrf,
parsed_schema,
records,
codec='deflate',
codec_compression_level=4)
def test_record_named_type():
"""https://github.com/fastavro/fastavro/issues/450"""
schema = {
"type":
"record",
"name":
"test_record_named_type",
"fields": [{
"name": "test1",
"type": {
"type": "record",
"name": "my_record",
"fields": [{
"name": "field1",
"type": "string",
}]
},
}, {
"name": "test2",
"type": "my_record",
}]
}
record = {"test1": {"field1": "foo"}, "test2": {"field1": "bar"}}
parsed_schema = parse_schema(schema)
validate(record, parsed_schema)
def write_garmin_file_object_to_file(gfile, avro_file):
parsed_schema = fastavro.parse_schema(GarminFile._avro_schema)
js = [gfile.to_dict()]
with gzip.open('%s.tmp' % avro_file, 'wb') as f:
fastavro.writer(f, parsed_schema, js, validator=True)
os.rename('%s.tmp' % avro_file, avro_file)
return True
def validater(schema, records, runs=1):
times = []
valid = []
schema = parse_schema(schema)
for _ in range(runs):
start = time.time()
valid = validate_many(records, schema)
end = time.time()
times.append(end - start)
print('... {0} runs averaged {1} seconds'.format(runs, (sum(times) / runs)))
return valid
def write(schema, records, runs=1):
times = []
schema = parse_schema(schema)
for _ in range(runs):
iostream = BytesIO()
start = time.time()
writer(iostream, schema, records)
end = time.time()
times.append(end - start)
print('... {0} runs averaged {1} seconds'.format(runs, (sum(times) / runs)))
return iostream
def read_schemaless(iostream, schema, num_records, runs=1):
times = []
schema = parse_schema(schema)
for _ in range(runs):
for _ in range(num_records):
iostream.seek(0)
start = time.time()
record = schemaless_reader(iostream, schema)
end = time.time()
times.append(end - start)
print('... {0} runs averaged {1} seconds'.format(runs, (sum(times) / runs)))
return records
def _bq_to_avro_blocks(bq_blocks, avro_schema_json):
avro_schema = fastavro.parse_schema(avro_schema_json)
avro_blocks = []
for block in bq_blocks:
blockio = six.BytesIO()
for row in block:
fastavro.schemaless_writer(blockio, avro_schema, row)
response = bigquery_storage_v1beta1.types.ReadRowsResponse()
response.avro_rows.row_count = len(block)
response.avro_rows.serialized_binary_rows = blockio.getvalue()
avro_blocks.append(response)
return avro_blocks
def _avro_schema(read_session):
"""Extract and parse Avro schema from a read session.
Args:
read_session ( \
~google.cloud.bigquery_storage_v1beta1.types.ReadSession \
):
The read session associated with this read rows stream. This
contains the schema, which is required to parse the data
blocks.
Returns:
Tuple[fastavro.schema, Tuple[str]]:
A parsed Avro schema, using :func:`fastavro.schema.parse_schema`
and the column names for a read session.
"""
json_schema = json.loads(read_session.avro_schema.schema)
column_names = tuple((field["name"] for field in json_schema["fields"]))
return fastavro.parse_schema(json_schema), column_names
SCHEMA = {
"namespace": "big_obj",
"type": "record",
"name": "DirFolders",
"fields": [
{"name": "base_folder", "type": {
"type": "record",
"name": "Folder",
"fields": [
{"name": "path", "type": "string"},
{"name": "last_mod", "type": ["null", "string"], "default": "null"},
{"name": "files", "type": ["null", {"type": "array", "items": {
"type": "record",
"name": "File",
"fields": [
{"name": "file_name", "type": "string"},
{"name": "content", "type": "bytes"},
{"name": "last_mod", "type": ["null", "string"], "default": "null"}
]
}}], "default": "null"},
{"name": "folders", "type":
{"type": "array", "items": "Folder"}}
]}},
{"name": "details", "type": "string"},
]
}
avro_schema = fastavro.parse_schema(SCHEMA)