def is_avro(path_or_buffer):
"""Return True if path (or buffer) points to an Avro file.
Paramaters
----------
path_or_buffer: path to file or file line object
Path to file
"""
if is_str(path_or_buffer):
fp = open(path_or_buffer, 'rb')
close = True
else:
fp = path_or_buffer
close = False
try:
header = fp.read(len(MAGIC))
return header == MAGIC
finally:
if close:
fp.close()
def validate(datum, schema):
"""Determine if a python datum is an instance of a schema."""
record_type = extract_record_type(schema)
if record_type == 'null':
return datum is None
if record_type == 'boolean':
return isinstance(datum, bool)
if record_type == 'string':
return is_str(datum)
if record_type == 'bytes':
return isinstance(datum, (bytes, decimal.Decimal))
if record_type == 'int':
return ((isinstance(datum, (
int,
long,
)) and INT_MIN_VALUE <= datum <= INT_MAX_VALUE)
or isinstance(datum, (datetime.time, datetime.datetime)))
if record_type == 'long':
return ((isinstance(datum, (
int,
long,
)) and INT_MIN_VALUE <= datum <= INT_MAX_VALUE)
or isinstance(datum, (datetime.time, datetime.datetime)))
if record_type in ['float', 'double']:
return isinstance(datum, (int, long, float))
if record_type == 'fixed':
return isinstance(datum, bytes) and len(datum) == schema['size']
if record_type == 'union':
if isinstance(datum, tuple):
(name, datum) = datum
for candidate in schema:
if extract_record_type(candidate) == 'record':
if name == candidate["name"]:
return validate(datum, candidate)
else:
return False
return any(validate(datum, s) for s in schema)
# dict-y types from here on.
if record_type == 'enum':
return datum in schema['symbols']
if record_type == 'array':
return (isinstance(datum, Iterable)
and all(validate(d, schema['items']) for d in datum))
if record_type == 'map':
return (isinstance(datum, Mapping)
and all(is_str(k) for k in datum.keys())
and all(validate(v, schema['values']) for v in datum.values()))
if record_type in (
'record',
'error',
'request',
):
return (isinstance(datum, Mapping) and all(
validate(datum.get(f['name'], f.get('default')), f['type'])
for f in schema['fields']))
if record_type in SCHEMA_DEFS:
return validate(datum, SCHEMA_DEFS[record_type])
raise ValueError('unkown record type - %s' % record_type)
def validate(datum, schema):
"""Determine if a python datum is an instance of a schema."""
record_type = extract_record_type(schema)
if record_type == 'null':
return datum is None
if record_type == 'boolean':
return isinstance(datum, bool)
if record_type == 'string':
return is_str(datum)
if record_type == 'bytes':
return isinstance(datum, bytes)
if record_type == 'int':
return (
isinstance(datum, (int, long,)) and
INT_MIN_VALUE <= datum <= INT_MAX_VALUE
)
if record_type == 'long':
return (
isinstance(datum, (int, long,)) and
LONG_MIN_VALUE <= datum <= LONG_MAX_VALUE
)
if record_type in ['float', 'double']:
return isinstance(datum, (int, long, float))
if record_type == 'fixed':
return isinstance(datum, bytes) and len(datum) == schema['size']
if record_type == 'union':
return any(validate(datum, s) for s in schema)
# dict-y types from here on.
if record_type == 'enum':
return datum in schema['symbols']
if record_type == 'array':
return (
isinstance(datum, Iterable) and
all(validate(d, schema['items']) for d in datum)
)
if record_type == 'map':
return (
isinstance(datum, Mapping) and
all(is_str(k) for k in datum.keys()) and
all(validate(v, schema['values']) for v in datum.values())
)
if record_type in ('record', 'error', 'request',):
return (
isinstance(datum, Mapping) and
all(
validate(datum.get(f['name'], f.get('default')), f['type'])
for f in schema['fields']
)
)
if record_type in SCHEMA_DEFS:
return validate(datum, SCHEMA_DEFS[record_type])
raise ValueError('unkown record type - %s' % record_type)
