Merge pull request #383 from reef-technologies/cache-control

Support for cache control headers while uploading
reef-technologies · Apr 20, 2023 · df4a1db · df4a1db
2 parents f7668d6 + e5d680e
commit df4a1db
Show file tree

Hide file tree

Showing 19 changed files with 206 additions and 12 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 * Add support for custom upload timestamp
+* Add support for cache control header while uploading
 
 ### Infrastructure
 * Remove dependency from `arrow`

diff --git a/b2sdk/bucket.py b/b2sdk/bucket.py
@@ -495,6 +495,7 @@ def upload_bytes(
         legal_hold: Optional[LegalHold] = None,
         large_file_sha1: Optional[Sha1HexDigest] = None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Upload bytes in memory to a B2 file.
@@ -512,6 +513,7 @@ def upload_bytes(
         :param bool legal_hold: legal hold setting
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         :rtype: b2sdk.v2.FileVersion
         """
         upload_source = UploadSourceBytes(data_bytes)
@@ -526,6 +528,7 @@ def upload_bytes(
             legal_hold=legal_hold,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def upload_local_file(
@@ -542,6 +545,7 @@ def upload_local_file(
         legal_hold: Optional[LegalHold] = None,
         upload_mode: UploadMode = UploadMode.FULL,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Upload a file on local disk to a B2 file.
@@ -565,6 +569,7 @@ def upload_local_file(
         :param bool legal_hold: legal hold setting
         :param b2sdk.v2.UploadMode upload_mode: desired upload mode
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         :rtype: b2sdk.v2.FileVersion
         """
         upload_source = UploadSourceLocalFile(local_path=local_file, content_sha1=sha1_sum)
@@ -596,6 +601,7 @@ def upload_local_file(
             legal_hold=legal_hold,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def upload_unbound_stream(
@@ -617,6 +623,7 @@ def upload_unbound_stream(
         read_size: int = 8192,
         unused_buffer_timeout_seconds: float = 3600.0,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Upload an unbound file-like read-only object to a B2 file.
@@ -680,6 +687,7 @@ def upload_unbound_stream(
         :param read_size: size of a single read operation performed on the ``read_only_object``
         :param unused_buffer_timeout_seconds: amount of time that a buffer can be idle before returning error
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         :rtype: b2sdk.v2.FileVersion
         """
         if buffers_count <= 1:
@@ -719,6 +727,7 @@ def upload_unbound_stream(
             max_queue_size=buffers_count - 1,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def upload(
@@ -734,6 +743,7 @@ def upload(
         legal_hold: Optional[LegalHold] = None,
         large_file_sha1: Optional[Sha1HexDigest] = None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Upload a file to B2, retrying as needed.
@@ -760,6 +770,7 @@ def upload(
         :param bool legal_hold: legal hold setting
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         :rtype: b2sdk.v2.FileVersion
         """
         return self.create_file(
@@ -775,6 +786,7 @@ def upload(
             legal_hold=legal_hold,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def create_file(
@@ -793,6 +805,7 @@ def create_file(
         max_part_size=None,
         large_file_sha1=None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Creates a new file in this bucket using an iterable (list, tuple etc) of remote or local sources.
@@ -824,6 +837,7 @@ def create_file(
         :param int max_part_size: upper limit of part size for the transfer planner, in bytes
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         """
         return self._create_file(
             self.api.services.emerger.emerge,
@@ -841,6 +855,7 @@ def create_file(
             max_part_size=max_part_size,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def create_file_stream(
@@ -859,6 +874,7 @@ def create_file_stream(
         max_part_size=None,
         large_file_sha1=None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Creates a new file in this bucket using a stream of multiple remote or local sources.
@@ -892,6 +908,7 @@ def create_file_stream(
         :param int max_part_size: upper limit of part size for the transfer planner, in bytes
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         """
         return self._create_file(
             self.api.services.emerger.emerge_stream,
@@ -909,6 +926,7 @@ def create_file_stream(
             max_part_size=max_part_size,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def _create_file(
@@ -966,6 +984,7 @@ def concatenate(
         max_part_size=None,
         large_file_sha1=None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Creates a new file in this bucket by concatenating multiple remote or local sources.
@@ -994,6 +1013,7 @@ def concatenate(
         :param int max_part_size: upper limit of part size for the transfer planner, in bytes
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         """
         return self.create_file(
             list(WriteIntent.wrap_sources_iterator(outbound_sources)),
@@ -1010,6 +1030,7 @@ def concatenate(
             max_part_size=max_part_size,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def concatenate_stream(
@@ -1026,6 +1047,7 @@ def concatenate_stream(
         legal_hold: Optional[LegalHold] = None,
         large_file_sha1: Optional[Sha1HexDigest] = None,
         custom_upload_timestamp: Optional[int] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Creates a new file in this bucket by concatenating stream of multiple remote or local sources.
@@ -1050,6 +1072,7 @@ def concatenate_stream(
         :param bool legal_hold: legal hold setting
         :param Sha1HexDigest,None large_file_sha1: SHA-1 hash of the result file or ``None`` if unknown
         :param int,None custom_upload_timestamp: override object creation date, expressed as a number of milliseconds since epoch
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         """
         return self.create_file_stream(
             WriteIntent.wrap_sources_iterator(outbound_sources_iterator),
@@ -1064,6 +1087,7 @@ def concatenate_stream(
             legal_hold=legal_hold,
             large_file_sha1=large_file_sha1,
             custom_upload_timestamp=custom_upload_timestamp,
+            cache_control=cache_control,
         )
 
     def get_download_url(self, filename):
@@ -1104,6 +1128,7 @@ def copy(
         source_content_type: Optional[str] = None,
         file_retention: Optional[FileRetentionSetting] = None,
         legal_hold: Optional[LegalHold] = None,
+        cache_control: Optional[str] = None,
         min_part_size=None,
         max_part_size=None,
     ):
@@ -1131,6 +1156,7 @@ def copy(
         :param str,None source_content_type: source file's content type, useful when copying files with SSE-C
         :param b2sdk.v2.FileRetentionSetting file_retention: file retention setting for the new file.
         :param bool legal_hold: legal hold setting for the new file.
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         :param int min_part_size: lower limit of part size for the transfer planner, in bytes
         :param int max_part_size: upper limit of part size for the transfer planner, in bytes
         """
@@ -1159,6 +1185,7 @@ def copy(
                     source_encryption=source_encryption,
                     file_retention=file_retention,
                     legal_hold=legal_hold,
+                    cache_control=cache_control,
                 ).result()
             except CopySourceTooBig as e:
                 copy_source.length = e.size
@@ -1175,6 +1202,7 @@ def copy(
             encryption=destination_encryption,
             file_retention=file_retention,
             legal_hold=legal_hold,
+            cache_control=cache_control,
             min_part_size=min_part_size,
             max_part_size=max_part_size,
         )

diff --git a/b2sdk/file_version.py b/b2sdk/file_version.py
@@ -48,6 +48,7 @@ class BaseFileVersion:
         'file_retention',
         'mod_time_millis',
         'replication_status',
+        'cache_control',
     ]
     _TYPE_MATCHER = re.compile('[a-z0-9]+_[a-z0-9]+_f([0-9]).*')
     _FILE_TYPE = {
@@ -71,6 +72,7 @@ def __init__(
         file_retention: FileRetentionSetting = NO_RETENTION_FILE_SETTING,
         legal_hold: LegalHold = LegalHold.UNSET,
         replication_status: Optional[ReplicationStatus] = None,
+        cache_control: Optional[str] = None,
     ):
         self.api = api
         self.id_ = id_
@@ -84,6 +86,7 @@ def __init__(
         self.file_retention = file_retention
         self.legal_hold = legal_hold
         self.replication_status = replication_status
+        self.cache_control = cache_control
 
         if SRC_LAST_MODIFIED_MILLIS in self.file_info:
             self.mod_time_millis = int(self.file_info[SRC_LAST_MODIFIED_MILLIS])
@@ -124,6 +127,7 @@ def _get_args_for_clone(self):
             'file_retention': self.file_retention,
             'legal_hold': self.legal_hold,
             'replication_status': self.replication_status,
+            'cache_control': self.cache_control,
         }  # yapf: disable
 
     def as_dict(self):
@@ -135,6 +139,7 @@ def as_dict(self):
             'serverSideEncryption': self.server_side_encryption.as_dict(),
             'legalHold': self.legal_hold.value,
             'fileRetention': self.file_retention.as_dict(),
+            'cacheControl': self.cache_control,
         }
 
         if self.size is not None:
@@ -251,6 +256,7 @@ def __init__(
         file_retention: FileRetentionSetting = NO_RETENTION_FILE_SETTING,
         legal_hold: LegalHold = LegalHold.UNSET,
         replication_status: Optional[ReplicationStatus] = None,
+        cache_control: Optional[str] = None,
     ):
         self.account_id = account_id
         self.bucket_id = bucket_id
@@ -270,6 +276,7 @@ def __init__(
             file_retention=file_retention,
             legal_hold=legal_hold,
             replication_status=replication_status,
+            cache_control=cache_control,
         )
 
     def _get_args_for_clone(self):
@@ -342,6 +349,7 @@ def _get_upload_headers(self) -> bytes:
             server_side_encryption=sse,
             file_retention=self.file_retention,
             legal_hold=self.legal_hold,
+            cache_control=self.cache_control,
         )
 
         headers_str = ''.join(
@@ -418,6 +426,7 @@ def __init__(
             file_retention=file_retention,
             legal_hold=legal_hold,
             replication_status=replication_status,
+            cache_control=cache_control,
         )
 
     def _get_args_for_clone(self):
@@ -501,10 +510,10 @@ def from_api_response(self, file_version_dict, force_action=None):
         file_retention = FileRetentionSetting.from_file_version_dict(file_version_dict)
 
         legal_hold = LegalHold.from_file_version_dict(file_version_dict)
-
         replication_status_value = file_version_dict.get('replicationStatus')
         replication_status = replication_status_value and ReplicationStatus[
             replication_status_value.upper()]
+        cache_control = file_version_dict.get('cacheControl')
 
         return self.FILE_VERSION_CLASS(
             self.api,
@@ -523,6 +532,7 @@ def from_api_response(self, file_version_dict, force_action=None):
             file_retention,
             legal_hold,
             replication_status,
+            cache_control,
         )
 
 
@@ -554,13 +564,19 @@ def file_info_from_headers(cls, headers: dict) -> dict:
 
     def from_response_headers(self, headers):
         file_info = self.file_info_from_headers(headers)
+
         if 'Content-Range' in headers:
             range_, size = self.range_and_size_from_header(headers['Content-Range'])
             content_length = int(headers['Content-Length'])
         else:
             size = content_length = int(headers['Content-Length'])
             range_ = Range(0, max(size - 1, 0))
 
+        if 'Cache-Control' in headers:
+            cache_control = b2_url_decode(headers['Cache-Control'])
+        else:
+            cache_control = None
+
         return DownloadVersion(
             api=self.api,
             id_=headers['x-bz-file-id'],
@@ -576,7 +592,7 @@ def from_response_headers(self, headers):
             content_length=content_length,
             content_language=headers.get('Content-Language'),
             expires=headers.get('Expires'),
-            cache_control=headers.get('Cache-Control'),
+            cache_control=cache_control,
             content_encoding=headers.get('Content-Encoding'),
             file_retention=FileRetentionSetting.from_response_headers(headers),
             legal_hold=LegalHold.from_response_headers(headers),

diff --git a/b2sdk/large_file/services.py b/b2sdk/large_file/services.py
@@ -87,6 +87,7 @@ def start_large_file(
         encryption: Optional[EncryptionSetting] = None,
         file_retention: Optional[FileRetentionSetting] = None,
         legal_hold: Optional[LegalHold] = None,
+        cache_control: Optional[str] = None,
     ):
         """
         Start a large file transfer.
@@ -95,8 +96,9 @@ def start_large_file(
         :param str,None content_type: the MIME type, or ``None`` to accept the default based on file extension of the B2 file name
         :param dict,None file_info: a file info to store with the file or ``None`` to not store anything
         :param b2sdk.v2.EncryptionSetting encryption: encryption settings (``None`` if unknown)
-        :param b2sdk.v2.LegalHold legal_hold: legal hold setting
         :param b2sdk.v2.FileRetentionSetting file_retention: file retention setting
+        :param b2sdk.v2.LegalHold legal_hold: legal hold setting
+        :param str,None cache_control: an optional cache control setting. Syntax based on the section 14.9 of RFC 2616. Example string value: 'public, max-age=86400, s-maxage=3600, no-transform'.
         """
         return UnfinishedLargeFile(
             self.services.session.start_large_file(
@@ -107,6 +109,7 @@ def start_large_file(
                 server_side_encryption=encryption,
                 file_retention=file_retention,
                 legal_hold=legal_hold,
+                cache_control=cache_control,
             )
         )
 

diff --git a/b2sdk/large_file/unfinished_large_file.py b/b2sdk/large_file/unfinished_large_file.py
@@ -37,6 +37,7 @@ def __init__(self, file_dict):
         self.encryption = EncryptionSettingFactory.from_file_version_dict(file_dict)
         self.file_retention = FileRetentionSetting.from_file_version_dict(file_dict)
         self.legal_hold = LegalHold.from_file_version_dict(file_dict)
+        self.cache_control = file_dict.get('cacheControl')
 
     def __repr__(self):
         return '<%s %s %s>' % (self.__class__.__name__, self.bucket_id, self.file_name)