def test_generator(self):
from ambry_sources.sources import GeneratorSource, SourceSpec
from ambry_sources import head, tail
cache_fs = fsopendir(self.setup_temp_dir())
def gen():
yield list('abcde')
for i in range(10):
yield [i, i + 1, i + 2, i + 3, i + 4]
f = HDFPartition(cache_fs, 'foobar')
s = GeneratorSource(SourceSpec('foobar'), gen())
ri = RowIntuiter().run(head(s, 100), tail(s, 100))
row_spec = self._row_intuiter_to_dict(ri)
ti = TypeIntuiter().process_header(ri.headers).run(GeneratorSource(SourceSpec('foobar'), gen()))
with f.writer as w:
w.set_row_spec(row_spec, ri.headers)
w.set_types(ti)
f.load_rows(GeneratorSource(SourceSpec('foobar'), gen()))
self.assertEqual(f.headers, list('abcde'))
rows = []
for row in f.select():
rows.append(row.dict)
self.assertEqual(len(rows), 10)
self.assertEqual(rows[0], {'a': 0, 'b': 1, 'c': 2, 'd': 3, 'e': 4})
self.assertEqual(rows[-1], {'a': 9, 'b': 10, 'c': 11, 'd': 12, 'e': 13})
def test_creates_virtual_tables_for_partition_with_segment_without_errors(
self):
fs = fsopendir('temp://')
def gen():
# generate header
yield ['col1', 'col2']
# generate rows
yield [0, 0]
yield [1, 1]
mprows = MPRowsFile(fs, 'example.com/simple-0.1.3/1.mpr')
mprows.load_rows(GeneratorSource(SourceSpec('foobar'), gen()))
# create virtual tables. This should not raise an error.
#
connection = apsw.Connection(':memory:')
try:
add_partition(connection, mprows, 'vid1')
except Exception as exc:
raise AssertionError(
'partition adding unexpectadly failed with {} error.'.format(
exc))
# check selected rows
#
cursor = connection.cursor()
result = cursor.execute('SELECT * FROM {}'.format('vid1')).fetchall()
self.assertEqual(result, [(0, 0), (1, 1)])
def test_executes_select_query_without_any_error(self, fake_shares):
fake_shares.return_value = True
def gen():
# generate header
yield ['col1', 'col2']
# generate first row
yield [0, 0]
fs = fsopendir('temp://')
datafile = MPRowsFile(fs, 'vid1')
datafile.load_rows(GeneratorSource(SourceSpec('foobar'), gen()))
connection = None
try:
PostgreSQLTestBase._create_postgres_test_db()
connection = psycopg2.connect(**PostgreSQLTestBase.pg_test_db_data)
# create foreign table for partition
with connection.cursor() as cursor:
# we have to close opened transaction.
cursor.execute('COMMIT;')
add_partition(cursor, datafile, 'vid1')
# query just created foreign table.
with connection.cursor() as cursor:
cursor.execute('SELECT * FROM partitions.vid1;')
finally:
if connection:
connection.close()
PostgreSQLTestBase._drop_postgres_test_db()
def test_generator(self):
from ambry_sources.sources import GeneratorSource, SourceSpec
cache_fs = fsopendir(self.setup_temp_dir())
def gen():
yield list('abcde')
for i in range(10):
yield [i, i + 1, i + 2, i + 3, i + 4]
f = MPRowsFile(cache_fs, 'foobar').load_rows(
GeneratorSource(SourceSpec('foobar'), gen()))
self.assertEqual(1, f.info['data_start_row'])
self.assertEqual(11, f.info['data_end_row'])
self.assertEqual([0], f.info['header_rows'])
self.assertEqual(f.headers, list('abcde'))
rows = list(f.select())
self.assertEqual(len(rows), 10)
self.assertEqual(sorted(rows[0].keys()), sorted(list('abcde')))
self.assertTrue(f.is_finalized)
def test_uses_url_as_table(self):
fake_execute = Mock(return_value=iter([[1], [2]]))
connection = AttrDict({'execute': fake_execute})
spec = SourceSpec('table1')
relation_source = DatabaseRelationSource(spec, 'sqlite', connection)
rows = [x for x in relation_source]
self.assertEqual(rows, [[1], [2]])
fake_execute.assert_called_once_with(
'SELECT * FROM {};'.format('table1'))
def spec(self):
"""Return a SourceSpec to describe this source"""
from ambry_sources.sources import SourceSpec
d = self.dict
d['url'] = self.url
# Will get the URL twice; once as ref and once as URL, but the ref is ignored
return SourceSpec(**d)
def _get_generator_source(self, header, rows):
def gen():
# generate header
yield header
# generate rows
for row in rows:
yield row
return GeneratorSource(SourceSpec('foobar'), gen())
def load_sources(cls, file_name='sources.csv'):
import tests
import csv
from os.path import join, dirname
from ambry_sources.sources import ColumnSpec, SourceSpec
test_data = fsopendir(join(dirname(tests.__file__), 'test_data'))
sources = {}
fixed_widths = (
('id', 1, 6),
('uuid', 7, 34),
('int', 41, 3),
('float', 44, 14),
)
fw_columns = [
ColumnSpec(**dict(list(zip('name start width'.split(), e))))
for e in fixed_widths
]
with test_data.open(file_name) as f:
r = csv.DictReader(f)
for row in r:
if row['name'] == 'simple_fixed':
row['columns'] = fw_columns
ss = SourceSpec(**row)
if 'expect_headers' in row or 'expect_start' in row:
ss.expect_headers = row.get('expect_headers')
try:
ss.expect_start = int(row.get('expect_start'))
except ValueError:
ss.expect_start = None
sources[ss.name] = ss
return sources
def test_reads_layer_specified_by_segment(self, fake_open, fake_shape,
fake_dumps):
fake_collection = self._get_fake_collection()
fake_open.return_value = fake_collection
spec = SourceSpec('http://example.com', segment=5)
fstor = Mock(spec=DelayedOpen)
fstor._fs = Mock()
source = ShapefileSource(spec, fstor)
next(source._get_row_gen())
self.assertEqual(len(fake_open.mock_calls), 1)
self.assertEqual(fake_open.call_args_list[0][1]['layer'], 5,
'open function was called with wrong layer.')
def _get_generator_source(rows=None):
if not rows:
rows = [[0, 0], [1, 1], [2, 2]]
def gen(rows=rows):
# generate header
yield ['col1', 'col2']
# generate some rows
for row in rows:
yield row
return GeneratorSource(SourceSpec('foobar'), gen())
def load_sources(cls, file_name='sources.csv'):
import tests
import csv
from os.path import join, dirname
from ambry_sources.sources import ColumnSpec, SourceSpec
test_data = fsopendir(join(dirname(tests.__file__), 'test_data'))
sources = {}
fixed_widths = (('id', 1, 6),
('uuid', 7, 34),
('int', 41, 3),
('float', 44, 14),
)
fw_columns = [ColumnSpec(**dict(list(zip('name start width'.split(), e)))) for e in fixed_widths]
with test_data.open(file_name) as f:
r = csv.DictReader(f)
for row in r:
if row['name'] == 'simple_fixed':
row['columns'] = fw_columns
ss = SourceSpec(**row)
if 'expect_headers' in row or 'expect_start' in row:
ss.expect_headers = row.get('expect_headers')
try:
ss.expect_start = int(row.get('expect_start'))
except ValueError:
ss.expect_start = None
sources[ss.name] = ss
return sources
def test_populates_columns_of_the_spec(self, fake_open, fake_get,
fake_shape, fake_dumps):
fake_collection = self._get_fake_collection()
fake_open.return_value = fake_collection
fake_get.return_value = [{'name': 'col1', 'type': 'int'}]
spec = SourceSpec('http://example.com')
fstor = Mock(spec=DelayedOpen)
fstor._fs = Mock()
source = ShapefileSource(spec, fstor)
next(source._get_row_gen())
self.assertEqual(len(source.spec.columns), 1)
self.assertEqual(source.spec.columns[0].name, 'col1')
self.assertEqual(len(fake_open.mock_calls), 1)
self.assertEqual(len(fake_get.mock_calls), 2)
def test_converts_row_id_to_integer(self, fake_open, fake_get, fake_shape,
fake_dumps):
fake_collection = self._get_fake_collection()
fake_open.return_value = fake_collection
fake_shape.expects_call().is_a_stub()
fake_dumps.expects_call().is_a_stub()
fake_get.return_value = [{'name': 'col1', 'type': 'int'}]
spec = SourceSpec('http://example.com')
fstor = Mock(spec=DelayedOpen)
fstor._fs = Mock()
source = ShapefileSource(spec, fstor)
row_gen = source._get_row_gen()
first_row = next(row_gen)
self.assertEqual(first_row[0], 0)
self.assertEqual(len(fake_open.mock_calls), 1)
self.assertEqual(len(fake_get.mock_calls), 2)
def test_last_element_in_the_row_is_wkt(self, fake_open, fake_get,
fake_shape, fake_dumps):
fake_collection = self._get_fake_collection()
fake_open.return_value = fake_collection
fake_shape.expects_call().is_a_stub()
fake_dumps.return_value = 'I AM FAKE WKT'
fake_get.return_value = [{'name': 'col1', 'type': 'int'}]
spec = SourceSpec('http://example.com')
fstor = Mock(spec=DelayedOpen)
fstor._fs = Mock()
source = ShapefileSource(spec, fstor)
row_gen = source._get_row_gen()
first_row = next(row_gen)
self.assertEqual(first_row[-1], 'I AM FAKE WKT')
self.assertEqual(len(fake_open.mock_calls), 1)
self.assertEqual(len(fake_get.mock_calls), 2)
def test_saves_header(self, fake_open, fake_get, fake_shape, fake_dumps):
fake_collection = self._get_fake_collection()
fake_open.return_value = fake_collection
fake_get.return_value = [{
'name': 'id',
'type': 'int'
}, {
'name': 'col1',
'type': 'int'
}, {
'name': 'geometry',
'type': 'geometry_type'
}]
spec = SourceSpec('http://example.com')
fstor = Mock(spec=DelayedOpen)
fstor._fs = Mock()
source = ShapefileSource(spec, fstor)
next(source._get_row_gen())
self.assertEqual(source._headers, ['id', 'col1', 'geometry'])
self.assertEqual(len(fake_open.mock_calls), 1)
self.assertEqual(len(fake_get.mock_calls), 2)
def test_allow_source_data_to_start_from_0_row(self):
columns = [
ColumnSpec(name='col1', position=1, start=1, width=1),
ColumnSpec(name='col2', position=2, start=3, width=1)
]
spec = SourceSpec('http://example.com',
header_lines=[],
start_line=0,
columns=columns)
assert spec.start_line == 0
fs = fsopendir('temp://')
with fs.open('temp.txt', 'w') as f:
f.write(u'1 1\n')
f.write(u'2 2\n')
f.write(u'3 3\n')
fstor = DelayedOpen(fs, 'temp.txt')
fixed_source = FixedSource(spec, fstor)
source_data = [x for x in fixed_source]
self.assertEqual(len(source_data), 3)
self.assertEqual(source_data, [['1', '1'], ['2', '2'], ['3', '3']])
def test_creates_virtual_table_for_source_with_header_containing_sql_reserved_words(
self):
# build rows reader
cache_fs = fsopendir(self.setup_temp_dir())
spec = SourceSpec('foobar')
def gen():
# yield header
yield ['create', 'index', 'where', 'select', 'distinct']
# yield rows
for i in range(10):
yield [i, i + 1, i + 2, i + 3, i + 4]
s = GeneratorSource(spec, gen())
mprows = MPRowsFile(cache_fs, spec.name).load_rows(s)
connection = apsw.Connection(':memory:')
table = 'table1'
add_partition(connection, mprows, table)
# check all columns and some rows.
cursor = connection.cursor()
query = 'SELECT count(*) FROM {};'.format(table)
result = cursor.execute(query).fetchall()
self.assertEqual(result, [(10, )])
with mprows.reader as r:
expected_first_row = next(iter(r)).row
# query by columns.
query = 'SELECT "create", "index", "where", "select", "distinct" FROM {} LIMIT 1;'.format(
table)
result = cursor.execute(query).fetchall()
self.assertEqual(len(result), 1)
self.assertEqual(result[0], expected_first_row)
class Rows(object):
spec = SourceSpec(None, header_lines=(3, 4), start_line=5)
def __iter__(self):
return iter(rows)
