Merge branch 'release-7.0.3'

CHANGELOG.md

@@ -1,3 +1,8 @@
+# 7.0.3, 2024-03-21
+
+* add support for zst writing (PR [#812](https://github.com/piskvorky/smart_open/pull/812), [@mpenkov](https://github.com/mpenkov))
+* roll back PR [#812](https://github.com/piskvorky/smart_open/pull/788), restore compatibility with built-in open function ([@mpenkov](https://github.com/mpenkov))
+
 # 7.0.2, 2024-03-21
 
 * Add `__next__` method to FileLikeProxy (PR [#811](https://github.com/piskvorky/smart_open/pull/811), [@ddelange](https://github.com/ddelange))

integration-tests/test_gcs.py

@@ -28,9 +28,6 @@ def write_read(key, content, write_mode, read_mode, **kwargs):
     with smart_open.open(key, read_mode, **kwargs) as fin:
         return fin.read()
 
-def open_only(key, read_mode, **kwargs) -> None:
-    with smart_open.open(key, read_mode, **kwargs):
-        pass
 
 def read_length_prefixed_messages(key, read_mode, **kwargs):
     result = io.BytesIO()
@@ -124,10 +121,3 @@ def test_gcs_performance_small_reads(benchmark):
 
     actual = benchmark(read_length_prefixed_messages, key, 'rb', buffering=ONE_MIB)
     assert actual == one_megabyte_of_msgs
-
-def test_gcs_performance_open(benchmark):
-    # we don't need to use a uri that actually exists in order to call GCS's open()
-    key = "gs://some-bucket/some_blob.txt"
-    transport_params = {'client': google.cloud.storage.Client()}
-    benchmark(open_only, key, 'rb', transport_params=transport_params)
-    assert True

setup.py

@@ -50,7 +50,6 @@ def read(fname):
     'boto3',
     'pytest',
     'pytest-rerunfailures',
-    'pytest-benchmark',
 ]
 
 setup(

smart_open/compression.py

@@ -106,8 +106,8 @@ def _handle_gzip(file_obj, mode):
 
 
 def _handle_zstd(file_obj, mode):
-    import zstandard as zstd
-    result = zstd.ZstdDecompressor().stream_reader(file_obj, closefd=True)
+    import zstandard  # type: ignore
+    result = zstandard.open(filename=file_obj, mode=mode)
     return result
 
 

smart_open/gcs.py

@@ -128,7 +128,10 @@ def Reader(bucket,
         warn_deprecated('line_terminator')
 
     bkt = client.bucket(bucket)
-    blob = bkt.blob(key)
+    blob = bkt.get_blob(key)
+
+    if blob is None:
+        raise google.cloud.exceptions.NotFound(f'blob {key} not found in {bucket}')
 
     return blob.open('rb', **blob_open_kwargs)
 

smart_open/tests/test_smart_open.py

@@ -77,6 +77,18 @@ def named_temporary_file(mode='w+b', prefix=None, suffix=None, delete=True):
             logger.error(e)
 
 
+def test_zst_write():
+    with named_temporary_file(suffix=".zst") as tmp:
+        with smart_open.open(tmp.name, "wt") as fout:
+            print("hello world", file=fout)
+            print("this is a test", file=fout)
+
+        with smart_open.open(tmp.name, "rt") as fin:
+            got = list(fin)
+
+    assert got == ["hello world\n", "this is a test\n"]
+
+
 class ParseUriTest(unittest.TestCase):
     """
     Test ParseUri class.

smart_open/version.py

@@ -1,4 +1,4 @@
-__version__ = '7.0.2'
+__version__ = '7.0.3'
 
 
 if __name__ == '__main__':