def test_convert_object_with_simple_list_with_join_serialization(self):
# given:
contain_list = {
'first': 'hello',
'list': [1, 2, 3, 4],
'list2': ['one', 'two'],
}
schema = skinfer.generate_schema(contain_list)
serialize_options = dict(method='join_values')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
# then:
self.assertEquals(['first', 'list', 'list2'], f.fieldnames)
self.assertEquals(['hello', '1,2,3,4', '["one","two"]'],
f.flatten(contain_list))
# and when:
schema['properties']['list']['flatson_serialize']['separator'] = '+'
f = Flatson(schema=schema)
# then:
self.assertEquals(['hello', '1+2+3+4', '["one","two"]'],
f.flatten(contain_list))
def test_convert_object_with_simple_list_with_default_serialization(self):
contain_list = {
'first': 'hello',
'list': [1, 2, 3, 4],
'list2': ['one', 'two'],
}
schema = skinfer.generate_schema(contain_list)
f = Flatson(schema=schema)
self.assertEquals(['first', 'list', 'list2'], f.fieldnames)
self.assertEquals(['hello', '[1,2,3,4]', '["one","two"]'], f.flatten(contain_list))
def test_lists_with_objects_with_default_serialization(self):
# given:
schema = skinfer.generate_schema(SAMPLE_WITH_LIST_OF_OBJECTS)
f = Flatson(schema=schema)
# when:
result = f.flatten(SAMPLE_WITH_LIST_OF_OBJECTS)
# then:
expected = '[{"key1":"value1","key2":"value2"},{"key1":"value3","key2":"value4"}]'
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', expected], result)
def test_convert_object_with_simple_list_with_default_serialization(self):
contain_list = {
'first': 'hello',
'list': [1, 2, 3, 4],
'list2': ['one', 'two'],
}
schema = skinfer.generate_schema(contain_list)
f = Flatson(schema=schema)
self.assertEquals(['first', 'list', 'list2'], f.fieldnames)
self.assertEquals(['hello', '[1,2,3,4]', '["one","two"]'],
f.flatten(contain_list))
def test_convert_nested_objects(self):
contain_nested_object = {
'first': 'hello',
'second': {
'one': 1,
'two': 2,
}
}
schema = skinfer.generate_schema(contain_nested_object)
f = Flatson(schema=schema)
self.assertEquals(['first', 'second.one', 'second.two'], f.fieldnames)
self.assertEquals(['hello', 1, 2], f.flatten(contain_nested_object))
def test_convert_nested_objects(self):
contain_nested_object = {
'first': 'hello',
'second': {
'one': 1,
'two': 2,
}
}
schema = skinfer.generate_schema(contain_nested_object)
f = Flatson(schema=schema)
self.assertEquals(['first', 'second.one', 'second.two'], f.fieldnames)
self.assertEquals(['hello', 1, 2], f.flatten(contain_nested_object))
def test_lists_with_objects_with_default_serialization(self):
# given:
schema = skinfer.generate_schema(SAMPLE_WITH_LIST_OF_OBJECTS)
f = Flatson(schema=schema)
# when:
result = f.flatten(SAMPLE_WITH_LIST_OF_OBJECTS)
# then:
expected = '[{"key1":"value1","key2":"value2"},{"key1":"value3","key2":"value4"}]'
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', expected], result)
def test_convert_object_with_nested_simple_list_with_default_serialization(self):
contain_list = {
'first': 'hello',
'second': {
'list1': [1, 2, 3, 4],
'word': 'world',
},
}
schema = skinfer.generate_schema(contain_list)
f = Flatson(schema=schema)
self.assertEquals(['first', 'second.list1', 'second.word'], f.fieldnames)
self.assertEquals(['hello', '[1,2,3,4]', 'world'], f.flatten(contain_list))
def test_array_serialization_with_extract_first(self):
# given:
sample = {'first': 'hello', 'list': ['one', 'two']}
schema = skinfer.generate_schema(sample)
serialize_options = dict(method='extract_first')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
result = f.flatten(sample)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', 'one'], result)
# and when:
sample2 = {'first': 'hello', 'list': []}
result = f.flatten(sample2)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', None], result)
def test_array_serialization_with_extract_first(self):
# given:
sample = {'first': 'hello', 'list': ['one', 'two']}
schema = skinfer.generate_schema(sample)
serialize_options = dict(method='extract_first')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
result = f.flatten(sample)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', 'one'], result)
# and when:
sample2 = {'first': 'hello', 'list': []}
result = f.flatten(sample2)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', None], result)
def test_array_serialization_with_extract_key_values(self):
# given:
schema = skinfer.generate_schema(SAMPLE_WITH_LIST_OF_OBJECTS)
serialize_options = dict(method='extract_key_values')
# when:
schema['properties']['list']['flatson_serialize'] = serialize_options
f = Flatson(schema=schema)
result = f.flatten(SAMPLE_WITH_LIST_OF_OBJECTS)
# then:
expected = 'key1:value1,key2:value2;key1:value3,key2:value4'
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', expected], result)
def test_array_serialization_with_extract_key_values(self):
# given:
schema = skinfer.generate_schema(SAMPLE_WITH_LIST_OF_OBJECTS)
serialize_options = dict(method='extract_key_values')
# when:
schema['properties']['list']['flatson_serialize'] = serialize_options
f = Flatson(schema=schema)
result = f.flatten(SAMPLE_WITH_LIST_OF_OBJECTS)
# then:
expected = 'key1:value1,key2:value2;key1:value3,key2:value4'
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', expected], result)
def test_register_custom_serialization_method(self):
# given:
sample = {'first': 'hello', 'list': ['one', 'two']}
schema = skinfer.generate_schema(sample)
serialize_options = dict(method='always_one')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
f.register_serialization_method('always_one', lambda _v, **kw: '1')
result = f.flatten(sample)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', '1'], result)
def test_convert_object_with_simple_list_with_join_serialization(self):
# given:
contain_list = {
'first': 'hello',
'list': [1, 2, 3, 4],
'list2': ['one', 'two'],
}
schema = skinfer.generate_schema(contain_list)
serialize_options = dict(method='join_values')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
# then:
self.assertEquals(['first', 'list', 'list2'], f.fieldnames)
self.assertEquals(['hello', '1,2,3,4', '["one","two"]'], f.flatten(contain_list))
# and when:
schema['properties']['list']['flatson_serialize']['separator'] = '+'
f = Flatson(schema=schema)
# then:
self.assertEquals(['hello', '1+2+3+4', '["one","two"]'], f.flatten(contain_list))
def test_convert_object_with_nested_simple_list_with_default_serialization(
self):
contain_list = {
'first': 'hello',
'second': {
'list1': [1, 2, 3, 4],
'word': 'world',
},
}
schema = skinfer.generate_schema(contain_list)
f = Flatson(schema=schema)
self.assertEquals(['first', 'second.list1', 'second.word'],
f.fieldnames)
self.assertEquals(['hello', '[1,2,3,4]', 'world'],
f.flatten(contain_list))
def test_register_custom_serialization_method(self):
# given:
sample = {'first': 'hello', 'list': ['one', 'two']}
schema = skinfer.generate_schema(sample)
serialize_options = dict(method='always_one')
schema['properties']['list']['flatson_serialize'] = serialize_options
# when:
f = Flatson(schema=schema)
f.register_serialization_method('always_one', lambda _v, **kw: '1')
result = f.flatten(sample)
# then:
self.assertEquals(['first', 'list'], f.fieldnames)
self.assertEquals(['hello', '1'], result)
class ODOWriter(BaseWriter):
"""
Writes items to a odo destination. https://odo.readthedocs.org/en/latest/
Needed parameters:
- schema (object)
schema object.
- odo_uri (str)
ODO valid destination uri.
"""
requirements = {
'schema': {
'type': object,
'required': True
},
'odo_uri': {
'type': six.string_types,
'required': True
}
}
def __init__(self, options):
super(ODOWriter, self).__init__(options)
from flatson import Flatson
schema = self.read_option('schema', None)
self.odo_uri = self.read_option('odo_uri', None)
self.flatson = Flatson(schema)
self.logger.info('ODOWriter has been initiated. Writing to: {}'.format(
self.odo_uri))
@retry_long
def write(self, dump_path, group_key=''):
from odo import odo, resource, discover
import pandas as pd
with gzip.open(dump_path) as f:
lines = [
json.loads(line.replace('\n', '')) for line in f.readlines()
]
flattened_lines = (self.flatson.flatten(line) for line in lines)
pf = pd.DataFrame(flattened_lines, columns=self.flatson.fieldnames)
dshape = discover(pf)
odo(pf, resource(self.odo_uri), dshape=dshape)
class ODOWriter(BaseWriter):
"""
Writes items to a odo destination. https://odo.readthedocs.org/en/latest/
Needed parameters:
- schema (object)
schema object.
- odo_uri (str)
ODO valid destination uri.
"""
requirements = {
'schema': {'type': object, 'required': True},
'odo_uri': {'type': basestring, 'required': True}
}
def __init__(self, options):
super(ODOWriter, self).__init__(options)
from flatson import Flatson
schema = self.read_option('schema', None)
self.odo_uri = self.read_option('odo_uri', None)
self.flatson = Flatson(schema)
self.logger.info('ODOWriter has been initiated. Writing to: {}'.format(self.odo_uri))
@retry_long
def write(self, dump_path, group_key=''):
from odo import odo, resource, discover
import pandas as pd
with gzip.open(dump_path) as f:
lines = [json.loads(line.replace('\n', '')) for line in f.readlines()]
flattened_lines = (self.flatson.flatten(line) for line in lines)
pf = pd.DataFrame(flattened_lines, columns=self.flatson.fieldnames)
dshape = discover(pf)
odo(pf, resource(self.odo_uri), dshape=dshape)
def test_when_no_declared_properties_flatten_empty_list(self):
f = Flatson(schema=EMPTY_SCHEMA)
result = f.flatten({'a_prop': 'a_value'})
self.assertEquals([], result)
def test_convert_simple_objects(self):
f = Flatson(schema=SIMPLE_SCHEMA)
self.assertEquals(['a_prop'], f.fieldnames)
self.assertEquals(['a_value'], f.flatten({'a_prop': 'a_value'}))
self.assertEquals([None], f.flatten({}))
def test_when_no_declared_properties_flatten_empty_list(self):
f = Flatson(schema=EMPTY_SCHEMA)
result = f.flatten({'a_prop': 'a_value'})
self.assertEquals([], result)
def test_convert_simple_objects(self):
f = Flatson(schema=SIMPLE_SCHEMA)
self.assertEquals(['a_prop'], f.fieldnames)
self.assertEquals(['a_value'], f.flatten({'a_prop': 'a_value'}))
self.assertEquals([None], f.flatten({}))
