def test_xopen_local(text_path):
from datasets.utils.streaming_download_manager import xopen
with xopen(text_path,
encoding="utf-8") as f, open(text_path,
encoding="utf-8") as expected_file:
assert list(f) == list(expected_file)
def test_streaming_gg_drive_zipped():
urlpath = StreamingDownloadManager().download_and_extract(
TEST_GG_DRIVE_ZIPPED_URL)
all_files = list(xglob(xjoin(urlpath, "*")))
assert len(all_files) == 1
assert xbasename(all_files[0]) == TEST_GG_DRIVE_FILENAME
with xopen(all_files[0]) as f:
assert f.read() == TEST_GG_DRIVE_CONTENT
def test_streaming_dl_manager_download(text_path):
dl_manager = StreamingDownloadManager()
out = dl_manager.download(text_path)
assert out == text_path
with xopen(out,
encoding="utf-8") as f, open(text_path,
encoding="utf-8") as expected_file:
assert f.read() == expected_file.read()
def test_xopen_local(text_path):
with xopen(text_path, "r",
encoding="utf-8") as f, open(text_path,
encoding="utf-8") as expected_file:
assert list(f) == list(expected_file)
with xpathopen(Path(text_path), "r", encoding="utf-8") as f, open(
text_path, encoding="utf-8") as expected_file:
assert list(f) == list(expected_file)
def test_streaming_dl_manager_download_and_extract_with_extraction(text_gz_path, text_path):
dl_manager = StreamingDownloadManager()
output_path = dl_manager.download_and_extract(text_gz_path)
path = os.path.basename(text_gz_path)
path = path[: path.rindex(".")]
assert output_path == f"gzip://{path}::{text_gz_path}"
fsspec_open_file = xopen(output_path, encoding="utf-8")
with fsspec_open_file as f, open(text_path, encoding="utf-8") as expected_file:
assert f.read() == expected_file.read()
def test_streaming_dl_manager_download(text_path):
from datasets.utils.streaming_download_manager import StreamingDownloadManager, xopen
dl_manager = StreamingDownloadManager()
out = dl_manager.download(text_path)
assert out == text_path
with xopen(out,
encoding="utf-8") as f, open(text_path,
encoding="utf-8") as expected_file:
assert f.read() == expected_file.read()
def test_streaming_dl_manager_download_and_extract_with_extraction(
text_gz_path, text_path):
from datasets.utils.streaming_download_manager import StreamingDownloadManager, xopen
dl_manager = StreamingDownloadManager()
filename = os.path.basename(text_gz_path).rstrip(".gz")
out = dl_manager.download_and_extract(text_gz_path)
assert out == f"gzip://{filename}::{text_gz_path}"
with xopen(out,
encoding="utf-8") as f, open(text_path,
encoding="utf-8") as expected_file:
assert f.read() == expected_file.read()
def test_streaming_dl_manager_extract_all_supported_single_file_compression_types(
compression_fs_class, gz_file, xz_file, zstd_file, bz2_file, lz4_file, text_file
):
input_paths = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bz2_file, "lz4": lz4_file}
input_path = str(input_paths[compression_fs_class.protocol])
dl_manager = StreamingDownloadManager()
output_path = dl_manager.extract(input_path)
path = os.path.basename(input_path)
path = path[: path.rindex(".")]
assert output_path == f"{compression_fs_class.protocol}://{path}::{input_path}"
fsspec_open_file = xopen(output_path, encoding="utf-8")
with fsspec_open_file as f, open(text_file, encoding="utf-8") as expected_file:
assert f.read() == expected_file.read()
def test_streaming_gg_drive_gzipped():
urlpath = StreamingDownloadManager().download_and_extract(
TEST_GG_DRIVE_GZIPPED_URL)
with xopen(urlpath) as f:
assert f.read() == TEST_GG_DRIVE_CONTENT
def test_streaming_gg_drive():
with xopen(TEST_GG_DRIVE_URL) as f:
assert f.read() == TEST_GG_DRIVE_CONTENT
def test_xopen_remote():
with xopen(TEST_URL, "r", encoding="utf-8") as f:
assert list(f) == TEST_URL_CONTENT.splitlines(keepends=True)
with xpathopen(Path(TEST_URL), "r", encoding="utf-8") as f:
assert list(f) == TEST_URL_CONTENT.splitlines(keepends=True)
def test_xopen_remote():
from datasets.utils.streaming_download_manager import xopen
with xopen(TEST_URL, encoding="utf-8") as f:
assert list(f) == TEST_URL_CONTENT.splitlines(keepends=True)
