Copilot commented on code in PR #9715: URL: https://github.com/apache/paimon/pull/9715#discussion_r3975902065
########## paimon-python/pypaimon/filesystem/oss_file_io.py: ########## @@ -0,0 +1,124 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +"""OSS conditional metadata creation, reusing the existing Arrow/Jindo FileIO.""" + +import re +from urllib.parse import urlparse + +from pypaimon.common.options.config import OssOptions +from pypaimon.filesystem.pyarrow_file_io import PyArrowFileIO + + +class OssFileIO(PyArrowFileIO): + """Override atomic metadata creation; inherit all ordinary file operations.""" + + def try_to_write_atomic(self, path: str, content: str) -> bool: + uri = urlparse(path) + if uri.scheme: + if uri.scheme != 'oss' or self._extract_oss_bucket(path) != self._oss_bucket: + raise ValueError("Atomic write must target the configured OSS bucket") + key = re.sub(r'/+', '/', uri.path).lstrip('/') + else: + key = path + if not self._use_jindo and not self._oss_bucket_in_endpoint: + bucket, _, key = path.partition('/') + if bucket != self._oss_bucket: + raise ValueError("Atomic write must target the configured OSS bucket") Review Comment: For oss:// URIs with embedded credentials (e.g. oss://AK:SK@endpoint/bucket/key), `_extract_oss_bucket` treats the first path segment as the bucket. The current key derivation uses the full `uri.path`, so `put_object` will write to a key prefixed with the bucket (bucket/key) even though the bucket is already provided to `oss2.Bucket`, resulting in the object being written to the wrong location for this URI form. ########## paimon-python/pypaimon/filesystem/oss_file_io.py: ########## @@ -0,0 +1,124 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +"""OSS conditional metadata creation, reusing the existing Arrow/Jindo FileIO.""" + +import re +from urllib.parse import urlparse + +from pypaimon.common.options.config import OssOptions +from pypaimon.filesystem.pyarrow_file_io import PyArrowFileIO + + +class OssFileIO(PyArrowFileIO): + """Override atomic metadata creation; inherit all ordinary file operations.""" + + def try_to_write_atomic(self, path: str, content: str) -> bool: + uri = urlparse(path) + if uri.scheme: + if uri.scheme != 'oss' or self._extract_oss_bucket(path) != self._oss_bucket: + raise ValueError("Atomic write must target the configured OSS bucket") + key = re.sub(r'/+', '/', uri.path).lstrip('/') + else: + key = path + if not self._use_jindo and not self._oss_bucket_in_endpoint: + bucket, _, key = path.partition('/') + if bucket != self._oss_bucket: + raise ValueError("Atomic write must target the configured OSS bucket") + if not key or key == '.' or key.endswith('/'): + return False + + try: + import oss2 + except ImportError as error: + raise ImportError( + "OSS atomic writes require oss2. Install pypaimon[oss] or pypaimon[jindo]." + ) from error + + access_key = self.properties.get(OssOptions.OSS_ACCESS_KEY_ID) + secret_key = self.properties.get(OssOptions.OSS_ACCESS_KEY_SECRET) + token = self.properties.get(OssOptions.OSS_SECURITY_TOKEN) + endpoint = self.properties.get(OssOptions.OSS_ENDPOINT) + if not access_key or not secret_key or not endpoint: + raise ValueError( + "OSS atomic writes require fs.oss.accessKeyId, fs.oss.accessKeySecret " + "and fs.oss.endpoint; pass fs.oss.securityToken for STS credentials.") + if '://' not in endpoint: + endpoint = 'https://' + endpoint + headers = self._sse_headers() + headers['x-oss-forbid-overwrite'] = 'true' + auth = oss2.StsAuth(access_key, secret_key, token) if token else oss2.Auth(access_key, secret_key) + + session = oss2.Session() + try: + bucket = oss2.Bucket(auth, endpoint, self._oss_bucket, session=session, + region=self.properties.get(OssOptions.OSS_REGION)) + try: + versioning = bucket.get_bucket_versioning().status + except oss2.exceptions.ServerError as error: + if error.status != 403 or error.code != 'AccessDenied': + raise + versioning = 'unknown (GetBucketVersioning denied)' + if versioning is not None: + self.logger.warning( + "Using legacy temporary-file-and-rename writes for OSS bucket %s " + "(versioning: %s). Concurrent commits are not protected against overwrites.", + self._oss_bucket, versioning) + return super().try_to_write_atomic(path, content) + try: + bucket.put_object(key, content.encode('utf-8'), headers=headers) + return True + except oss2.exceptions.ServerError as error: + if error.code == 'FileAlreadyExists': + return False + raise + except oss2.exceptions.OssError as error: + raise OSError("Failed to atomically write oss://{}/{}".format(self._oss_bucket, key)) from error + finally: + session.session.close() + + def _sse_headers(self): + """Match Java OSSFileIO's SSE resolution, including the native option fallback.""" + settings = (OssOptions.OSS_SSE_METHOD, OssOptions.OSS_SSE_KMS_KEY_ID, + OssOptions.OSS_SSE_DATA_ENCRYPTION) + values = [self.properties.get(setting) for setting in settings] + for setting, value in zip(settings, values): + if value is not None and not value.strip(): + raise ValueError("'{}' is set but blank".format(setting.key())) + method, key_id, data_encryption = [value.strip() if value is not None else None for value in values] + if method is None and key_id is None and data_encryption is None: + # Java forwards a nonempty native value as-is; the OSS service validates it. + algorithm = self.properties.get(OssOptions.OSS_SSE_ALGORITHM) + return {'x-oss-server-side-encryption': algorithm} if algorithm else {} Review Comment: `fs.oss.server-side-encryption-algorithm` is forwarded without stripping or blank validation. A whitespace-only value is truthy and will produce an invalid `x-oss-server-side-encryption` header, leading to hard-to-diagnose OSS request failures. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
