pyPreservica 3.0.1__py3-none-any.whl → 3.0.2__py3-none-any.whl

pyPreservica/__init__.py

@@ -23,6 +23,6 @@ from .mdformsAPI import MetadataGroupsAPI, Group, GroupField, GroupFieldType
 __author__ = "James Carr (drjamescarr@gmail.com)"
 
 # Version of the pyPreservica package
-__version__ = "3.0.1"
+__version__ = "3.0.2"
 
 __license__ = "Apache License Version 2.0"

pyPreservica/common.py

@@ -405,6 +405,9 @@ class Bitstream:
         self.length = int(length)
         self.fixity = fixity
         self.content_url = content_url
+        self.bs_index = None
+        self.gen_index = None
+        self.co_ref = None
 
     def __str__(self):
         return f"""

pyPreservica/entityAPI.py

@@ -8,7 +8,7 @@ author:     James Carr
 licence:    Apache License 2.0
 
 """
-import hashlib
+
 import os.path
 import uuid
 import xml.etree.ElementTree
@@ -118,6 +118,38 @@ class EntityAPI(AuthenticatedAPI):
                 logger.error(exception)
                 raise exception
 
+    def bitstream_location(self, bitstream: Bitstream):
+        """"
+        Retrieves information about a bitstreams storage locations
+        """
+        if not isinstance(bitstream, Bitstream):
+            logger.error("bitstream argument is not a Bitstream object")
+            raise RuntimeError("bitstream argument is not a Bitstream object")
+
+        storage_locations = []
+
+        url: str = f'{self.protocol}://{self.server}/api/entity/content-objects/{bitstream.co_ref}/generations/{bitstream.gen_index}/bitstreams/{bitstream.bs_index}/storage-locations'
+
+        with self.session.get(url, headers={HEADER_TOKEN: self.token}, stream=True) as request:
+            if request.status_code == requests.codes.unauthorized:
+                self.token = self.__token__()
+                return self.bitstream_location(bitstream)
+            elif request.status_code == requests.codes.ok:
+                xml_response = str(request.content.decode('utf-8'))
+                entity_response = xml.etree.ElementTree.fromstring(xml_response)
+                logger.debug(xml_response)
+                locations = entity_response.find(f'.//{{{self.entity_ns}}}StorageLocation')
+                for adapter in locations:
+                    storage_locations.append(adapter.attrib['name'])
+            else:
+                exception = HTTPException(bitstream.filename, request.status_code, request.url, "bitstream_location",
+                                          request.content.decode('utf-8'))
+                logger.error(exception)
+                raise exception
+
+        return storage_locations
+
+
     def bitstream_content(self, bitstream: Bitstream, filename: str, chunk_size: int = CHUNK_SIZE) -> Union[int, None]:
         """
         Download a file represented as a Bitstream to a local filename
@@ -1474,7 +1506,7 @@ class EntityAPI(AuthenticatedAPI):
             logger.error(exception)
             raise exception
 
-    def generation(self, url: str) -> Generation:
+    def generation(self, url: str, content_ref: str = None) -> Generation:
         """
          Retrieve a list of generation objects
 
@@ -1524,7 +1556,11 @@ class EntityAPI(AuthenticatedAPI):
             bitstreams = entity_response.findall(f'./{{{self.entity_ns}}}Bitstreams/{{{self.entity_ns}}}Bitstream')
             bitstream_list = []
             for bit in bitstreams:
-                bitstream_list.append(self.bitstream(bit.text))
+                bs: Bitstream = self.bitstream(bit.text)
+                bs.gen_index = index
+                if content_ref is not None:
+                    bs.co_ref = content_ref
+                bitstream_list.append(bs)
             generation = Generation(strtobool(ge.attrib['original']), strtobool(ge.attrib['active']),
                                     format_group.text if hasattr(format_group, 'text') else None,
                                     effective_date.text if hasattr(effective_date, 'text') else None,
@@ -1741,7 +1777,7 @@ class EntityAPI(AuthenticatedAPI):
             result = []
             for g in generations:
                 if hasattr(g, 'text'):
-                    generation = self.generation(g.text)
+                    generation = self.generation(g.text, content_object.reference)
                     generation.asset = content_object.asset
                     generation.content_object = content_object
                     generation.representation_type = content_object.representation_type

pyPreservica/uploadAPI.py

@@ -1659,7 +1659,22 @@ class UploadAPI(AuthenticatedAPI):
 
     def crawl_filesystem(self, filesystem_path, bucket_name, preservica_parent, callback: bool = False,
                          security_tag: str = "open",
-                         delete_after_upload: bool = True, max_MB_ingested: int = -1):
+                         delete_after_upload: bool = True, max_MB_ingested: int = -1, max_workflows = 8):
+        """
+        Crawl a filesystem and upload the contents to Preservica, the filesystem structure is mirrored in Preservica
+
+
+        :param str filesystem_path: The path to the filesystem to crawl
+        :param str bucket_name: The name of the bucket to upload to, use None to send directly to Preservica
+        :param str preservica_parent: The parent folder in Preservica to upload to
+        :param bool callback: Show upload progress bar
+        :param str security_tag: The security tag to apply to the uploaded assets
+        :param bool delete_after_upload: Delete the local copy of the package after the upload has completed
+        :param int max_MB_ingested: The maximum number of MB to ingest
+        :param int max_workflows: The maximum number of workflows to run concurrently
+
+
+        """
 
         def get_parent(client, identifier, parent_reference):
             id = str(os.path.dirname(identifier))
@@ -1684,12 +1699,17 @@ class UploadAPI(AuthenticatedAPI):
                 folder = entities.pop()
             return folder
 
-        from pyPreservica import EntityAPI
+        from pyPreservica import EntityAPI, WorkflowAPI
         entity_client = EntityAPI(username=self.username, password=self.password, server=self.server,
                                   tenant=self.tenant,
                                   two_fa_secret_key=self.two_fa_secret_key, use_shared_secret=self.shared_secret,
                                   protocol=self.protocol)
 
+        workflow_client = WorkflowAPI(username=self.username, password=self.password, server=self.server,
+                                  tenant=self.tenant,
+                                  two_fa_secret_key=self.two_fa_secret_key, use_shared_secret=self.shared_secret,
+                                  protocol=self.protocol)
+
         if preservica_parent:
             parent = entity_client.folder(preservica_parent)
             logger.info(f"Folders will be created inside Preservica collection {parent.title}")
@@ -1732,6 +1752,11 @@ class UploadAPI(AuthenticatedAPI):
                 else:
                     progress_display = None
 
+                workflow_queue_length = len(list(workflow_client.workflow_instances(workflow_state="Active", workflow_type="Ingest")))
+                while workflow_queue_length > max_workflows:
+                    sleep(30)
+                    workflow_queue_length = len(list(workflow_client.workflow_instances(workflow_state="Active", workflow_type="Ingest")))
+
                 if bucket_name is None:
                     self.upload_zip_package(path_to_zip_package=package, callback=progress_display,
                                             delete_after_upload=delete_after_upload)

{pyPreservica-3.0.1.dist-info → pyPreservica-3.0.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: pyPreservica
-Version: 3.0.1
+Version: 3.0.2
 Summary: Python library for the Preservica API
 Home-page: https://pypreservica.readthedocs.io/
 Author: James Carr
@@ -30,6 +30,17 @@ Requires-Dist: s3transfer
 Requires-Dist: azure-storage-blob
 Requires-Dist: tqdm
 Requires-Dist: pyotp
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license
+Dynamic: project-url
+Dynamic: requires-dist
+Dynamic: summary
 
 
 # pyPreservica

{pyPreservica-3.0.1.dist-info → pyPreservica-3.0.2.dist-info}/RECORD

@@ -1,19 +1,19 @@
-pyPreservica/__init__.py,sha256=6lYm_cDE5Z9aKXe_UfgNabaGq1MoIDPWUb16GgDGeS0,1177
+pyPreservica/__init__.py,sha256=HkIx9xYPB52tEV5TP_tH6MFW57o7yDtmSOU9DXOPW8o,1177
 pyPreservica/adminAPI.py,sha256=511bc5KtrCAXbDyBk39dmDnxUVDaOu6xaiyu0jYhxa4,37781
 pyPreservica/authorityAPI.py,sha256=Eule8g6LXr8c8SFcJgpRah4lH1FgevUItO5HhHDEaZE,9172
-pyPreservica/common.py,sha256=YvONehtYu2StoPY2oydWy86-zsyIK-r2FD79TnLi01Q,37707
+pyPreservica/common.py,sha256=nZeUObfypPXFauag5yYXEutvRC72sNXwVCN0wAFnssg,37796
 pyPreservica/contentAPI.py,sha256=z_IwndqzpTMtsDKUgneqWec5YPhi2Yb110Z-4tC42qU,22182
-pyPreservica/entityAPI.py,sha256=ljqiN2EjF20xprjrXjcb8OACtrgKn2K0Kl11uqh7uQU,119261
+pyPreservica/entityAPI.py,sha256=9BU55nvohydmwE4E3EviAkBrUcIj5ahmHQg2kLYZ8oY,121101
 pyPreservica/mdformsAPI.py,sha256=cTHxHgPV-EZAFN6C1JfnxrVNv3FLzqI1SV_fb5ZOxeE,19196
 pyPreservica/monitorAPI.py,sha256=HD-PUPdSI9wGAa07e2_2_-FLINH8PoWUwpFogz7F-j4,6269
 pyPreservica/opex.py,sha256=ccra1S4ojUXS3PlbU8WfxajOkJrwG4OykBnNrYP_jus,4875
 pyPreservica/parAPI.py,sha256=f0ZUxLd0U-BW6kBx5K7W2Pv7NjG3MkTNydmxQ3U1ZVE,9296
 pyPreservica/retentionAPI.py,sha256=F6okFSyqtnLhfMbcyChd_5V-D_PAxNLwyx8XohH2DEM,24840
-pyPreservica/uploadAPI.py,sha256=vIji6AUajiSvA2wuemHvU6RqHFd7PlKQToypIGx3KMw,98537
+pyPreservica/uploadAPI.py,sha256=_ldYoDWD6cOIlMXM98KTF2Xnc7GaZ8SKZNHPzbJyfQI,100105
 pyPreservica/webHooksAPI.py,sha256=_K3KUOsmwYf8qMa-mD47sAmNUW7Pzb9oKVpS0VoSbC0,6827
 pyPreservica/workflowAPI.py,sha256=OcOiiUdrQerbPllrkj1lWpmuW0jTuyyV0urwPSYcd_U,17561
-pyPreservica-3.0.1.dist-info/LICENSE.txt,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
-pyPreservica-3.0.1.dist-info/METADATA,sha256=SKMkfDkZMetZdqGhV_lxqqDT8meV1mJwTXLpzuSNWN8,2779
-pyPreservica-3.0.1.dist-info/WHEEL,sha256=YiKiUUeZQGmGJoR_0N1Y933DOBowq4AIvDe2-UIy8E4,91
-pyPreservica-3.0.1.dist-info/top_level.txt,sha256=iIBh6NAznYQHOV8mv_y_kGKSDITek9rANyFDwJsbU-c,13
-pyPreservica-3.0.1.dist-info/RECORD,,
+pyPreservica-3.0.2.dist-info/LICENSE.txt,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
+pyPreservica-3.0.2.dist-info/METADATA,sha256=aPMxioiNMabjITQ0OR-5cSUsk6VgNqSwfsg1QizXVNM,3018
+pyPreservica-3.0.2.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+pyPreservica-3.0.2.dist-info/top_level.txt,sha256=iIBh6NAznYQHOV8mv_y_kGKSDITek9rANyFDwJsbU-c,13
+pyPreservica-3.0.2.dist-info/RECORD,,

{pyPreservica-3.0.1.dist-info → pyPreservica-3.0.2.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (71.0.2)
+Generator: setuptools (75.8.0)
 Root-Is-Purelib: true
 Tag: py3-none-any