def _file_on_s3(self, relpath):
path = Path(__file__).parent / 'test_data' / relpath
try:
minio.fput_file(bucket, key, path)
yield
finally:
minio.remove(bucket, key)
def test_upload_xlsx(self):
path = Path(mock_xlsx_path)
minio.fput_file(minio.UserFilesBucket,
'eb785452-f0f2-4ebe-97ce-e225e346148e.xlsx', path)
expected_table = pd.DataFrame({
'Month': ['Jan', 'Feb'],
'Amount': [10, 20]
})
self._test_upload(uuid='eb785452-f0f2-4ebe-97ce-e225e346148e',
filename='test.xlsx',
ext='xlsx',
size=path.stat().st_size,
expected_result=ProcessResult(expected_table))
def test_upload_xls(self):
path = (Path(__file__).parent.parent / 'test_data' / 'example.xls')
minio.fput_file(minio.UserFilesBucket,
'eb785452-f0f2-4ebe-97ce-e225e346148e.xls', path)
expected_table = pd.DataFrame({
'foo': [1, 2],
'bar': [2, 3],
})
self._test_upload(uuid='eb785452-f0f2-4ebe-97ce-e225e346148e',
filename='test.xls',
ext='xls',
size=path.stat().st_size,
expected_result=ProcessResult(expected_table))
def test_read_file_fastparquet_issue_375(self):
path = (
Path(__file__).parent.parent
/ 'test_data' / 'fastparquet-issue-375-snappy.par'
)
minio.fput_file(minio.StoredObjectsBucket,
'fastparquet-issue-375-snappy.par', path)
so = StoredObject(
size=10,
bucket=minio.StoredObjectsBucket,
key='fastparquet-issue-375-snappy.par'
)
assert_frame_equal(so.get_table(), pd.DataFrame())
def write(bucket: str, key: str, table: pandas.DataFrame) -> int:
"""
Write a Pandas DataFrame to a minio file, overwriting if needed.
Return number of bytes written.
We aim to keep the file format "stable": all future versions of
parquet.read() should support all files written by today's version of this
function.
"""
with tempfile.NamedTemporaryFile() as tf:
fastparquet.write(tf.name, table, compression='SNAPPY',
object_encoding='utf8')
minio.fput_file(bucket, key, Path(tf.name))
tf.seek(0, io.SEEK_END)
return tf.tell()