Created script for bulk access

src/dspace/impl/client.py

@@ -231,7 +231,7 @@ def api_post(self, url, params, data, retry=False, content_type='application/jso
         check_response(r, "api post")
         return r
 
-    def api_put(self, url, params, json_p, retry=False):
+    def api_put(self, url, params, json_p, retry=False, content_type='application/json'):
         """
         Perform a PUT request. Refresh XSRF token if necessary.
         PUTs are typically used to update objects.
@@ -243,15 +243,16 @@ def api_put(self, url, params, json_p, retry=False):
                         Used if we need to refresh XSRF.
         @return:        Response from API
         """
-        h = {'Content-type': 'application/json'}
+        h = {'Content-type': content_type}
         r = self.session.put(url, params=params, json=json_p, headers=h)
         if 'DSPACE-XSRF-TOKEN' in r.headers:
             t = r.headers['DSPACE-XSRF-TOKEN']
-            logging.debug('Updating token to ' + t)
+            logging.debug('API Put: Updating token to ' + t)
             self.session.headers.update({'X-XSRF-Token': t})
             self.session.cookies.update({'X-XSRF-Token': t})
 
         if r.status_code == 403:
+            self.exception401Counter = 0
             # 403 Forbidden
             # If we had a CSRF failure, retry the request with the updated token
             # After speaking in #dev it seems that these
@@ -264,8 +265,26 @@ def api_put(self, url, params, json_p, retry=False):
                     logging.error('Already retried... something must be wrong')
                 else:
                     logging.debug("Retrying request with updated CSRF token")
-                    return self.api_put(url, params=params, json_p=json_p, retry=True)
-
+                    return self.api_put(url, params=params, json_p=json_p, retry=True, content_type=content_type)
+        elif r.status_code == 401:
+            r_json = r.json()
+            if 'message' in r_json and 'Authentication is required' in r_json[
+                    'message']:
+                if retry:
+                    logging.error(
+                        'API Post: Already retried... something must be wrong')
+                    self.exception401Counter = 0
+                else:
+                    logging.debug("API Post: Retrying request with updated CSRF token")
+                    # try to authenticate
+                    self.authenticate()
+                    # Try to authenticate and repeat the request 3 times -
+                    # if it won't happen log error
+                    self.exception401Counter = self.exception401Counter + 1
+                    retry_value = False
+                    if self.exception401Counter > 3:
+                        retry_value = True
+                    return self.api_put(url, params=params, json_p=json_p, retry=True, content_type=content_type)
         return r
 
     def api_delete(self, url, params, retry=False):
@@ -556,6 +575,24 @@ def delete_dso(self, dso=None, url=None, params=None):
             logging.error(f'{e}')
             return None
 
+
+    def get_bundle_by_name(self, name, item_uuid):
+        """
+        Get a bundle by name for a specific item
+        @param name:    Name of the bundle
+        @param item_uuid: UUID of the item
+        @return:        Bundle object
+        """
+        url = f'{self.API_ENDPOINT}core/items/{item_uuid}/bundles'
+        r_json = self.fetch_resource(url, params=None)
+        if '_embedded' in r_json:
+            if 'bundles' in r_json['_embedded']:
+                for bundle in r_json['_embedded']['bundles']:
+                    if bundle['name'] == name:
+                        return Bundle(bundle)
+        return None
+
+
     def get_bundles(self, parent=None, uuid=None):
         """
         Get bundles for an item
@@ -839,6 +876,47 @@ def get_items(self):
                     items.append(Item(item_resource))
         return items
 
+    def get_items_from_collection(self, collection_id, page=0, size=1000):
+        """
+        Get all items
+        @return:        list of Item objects
+        """
+        url = f'{self.API_ENDPOINT}discover/search/objects?sort=dc.date.accessioned,DESC&page={page}&size={size}&scope={collection_id}&dsoType=ITEM&embed=thumbnail'
+
+        items = list()
+        r = self.api_get(url)
+        r_json = parse_json(r)
+        if '_embedded' in r_json:
+            if 'searchResult' in r_json['_embedded']:
+                if '_embedded' in r_json['_embedded']['searchResult']:
+                    for item_resource in r_json['_embedded']['searchResult']['_embedded']['objects']:
+                        items.append(Item(item_resource['_embedded']['indexableObject']))
+
+        return items
+
+
+    def get_resource_policy(self, bundle_uuid):
+        """
+        Get a resource policy for a specific bundle
+        """
+        url = f'{self.API_ENDPOINT}authz/resourcepolicies/search/resource?uuid={bundle_uuid}&embed=eperson&embed=group'
+        r = self.api_get(url)
+        r_json = parse_json(r)
+        if '_embedded' in r_json:
+            if 'resourcepolicies' in r_json['_embedded']:
+                return r_json['_embedded']['resourcepolicies'][0]
+
+
+    def update_resource_policy_group(self, policy_id, group_uuid):
+        """
+        Update a resource policy with a new group
+        """
+        url = f'{self.API_ENDPOINT}authz/resourcepolicies/{policy_id}/group'
+        body = f'{self.API_ENDPOINT}eperson/groups/{group_uuid}'
+        r = self.api_put(url, None, body, content_type='text/uri-list')
+        return r
+
+
     def get_item(self, uuid):
         """
         Get an item, given its UUID

src/project_settings.py

@@ -9,7 +9,7 @@
     "resume_dir": "__temp/resume/",
 
     "backend": {
-        "endpoint": "http://dev-5.pc:85/server/api/",
+        "endpoint": "http://localhost:8080/server/api/",
         "user": "[email protected]",
         "password": "admin",
         "authentication": True,

tools/change_item_policies.py

@@ -0,0 +1,112 @@
+###
+# This script changes the policy of items in a community to a specific group. Bulk access.
+###
+import logging
+import os
+import sys
+
+
+_this_dir = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, os.path.join(_this_dir, "../src"))
+
+
+import dspace  # noqa
+import settings  # noqa
+import project_settings  # noqa
+from dspace.impl.models import Item  # noqa
+from dspace.impl.models import Community  # noqa
+from utils import init_logging, update_settings  # noqa
+
+_logger = logging.getLogger()
+
+# env settings, update with project_settings
+env = update_settings(settings.env, project_settings.settings)
+init_logging(_logger, env["log_file"])
+
+if "DSPACE_REST_API" in os.environ:
+    env["backend"]["endpoint"] = os.getenv("DSPACE_REST_API")
+    env_backend_endpoint = env["backend"]["endpoint"]
+    _logger.info(f"Loaded env.backend.endpoint from env DSPACE_REST_API."
+                 f" Current value: {env_backend_endpoint}")
+
+def get_all_items_from_collection(coll):
+    """
+    Get all items from collection
+    @param coll:
+    @return:
+    """
+    # Pagination and size because BE has a limit of 100 items per page and if the size is set to 1000 it will return
+    # only 100 items
+    page = 0
+    size = 5
+    all_collections = list()
+    has_more = True
+    while has_more:
+        collections_on_page = dspace_be.client.get_items_from_collection(coll.uuid, page=page, size=size)
+        if not collections_on_page:
+            has_more = False
+            break
+        page += 1
+        all_collections.extend(collections_on_page)
+    return all_collections
+
+
+
+if __name__ == '__main__':
+    dspace_be = dspace.rest(
+        env["backend"]["endpoint"],
+        env["backend"]["user"],
+        env["backend"]["password"],
+        env["backend"]["authentication"]
+    )
+
+    # Group ID of the group to which the policy will be changed e.g. admin group
+    GROUP_ID = "59ca14ed-0380-4655-bfa2-ca0711d1f1d0"
+
+    # Community UUID of the community whose items of collections will be updated
+    COM_UPDATE_ITEMS_UUID = 'e640c622-f0de-43e1-8446-bd6007737022'
+    COL_SUBCOLLS_URL = f'{dspace_be.endpoint}/core/communities/{COM_UPDATE_ITEMS_UUID}/collections'
+
+    COMMUNITY = Community({
+        "id": COM_UPDATE_ITEMS_UUID,
+        "type": "community",
+        "_links": {
+            "collections": {
+                "href": COL_SUBCOLLS_URL
+            }
+        },
+    })
+
+    # How many items were updated
+    counter = 0
+    # How many items were without file
+    without_file = 0
+    # Get all collections of the community
+    subcolls = dspace_be.client.get_collections(community=COMMUNITY)
+    for coll in subcolls:
+        # Counter for items in collection
+        collection_counter = 0
+        # Get all items of the collection
+        items_of_collection = get_all_items_from_collection(coll)
+        _logger.info(f'*******************Collection: {coll.name}*******************')
+        _logger.info(f'Items length: {len(items_of_collection)}')
+        for item in items_of_collection:
+            collection_counter += 1
+            _logger.debug(f'Item: {item.uuid}')
+            # Get bundle of the item - ORIGINAL
+            bundle = dspace_be.client.get_bundle_by_name('ORIGINAL', item.uuid)
+            # If there is no bundle, skip the item - there is no file
+            if not bundle:
+                _logger.debug(f'No ORIGINAL bundle for item uuid={item.uuid}')
+                without_file += 1
+                continue
+            counter += 1
+            resource_policy = dspace_be.client.get_resource_policy(bundle.uuid)
+            _logger.debug(
+                f'Changing policy uuid={resource_policy["id"]} for item uuid={item.uuid} to group uuid={GROUP_ID}')
+            r = dspace_be.client.update_resource_policy_group(resource_policy["id"], GROUP_ID)
+            _logger.debug('Response: ' + str(r))
+        _logger.info(f'===================Updated Items: {collection_counter}=====================')
+
+    _logger.info(f'Items Without file: {without_file}')
+    _logger.info(f'Total updated Items: {counter}')