Exclude users by Tag name or ID

.github/workflows/omero_plugin.yml

@@ -23,7 +23,7 @@ jobs:
     env:
       STAGE: cli
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v3
       - name: Checkout omero-test-infra
         uses: actions/checkout@master
         with:

.pre-commit-config.yaml

@@ -7,13 +7,13 @@ repos:
       - id: seed-isort-config
 
   - repo: https://github.com/PyCQA/isort
-    rev: 5.10.1
+    rev: 5.11.5
     hooks:
       - id: isort
         args: ["--profile", "black"]
 
   - repo: https://github.com/psf/black
-    rev: 22.1.0
+    rev: 22.3.0
     hooks:
       - id: black
         args: [--target-version=py36]

README.rst

@@ -15,6 +15,22 @@ system without running the deletion::
 
     $ omero demo-cleanup --gigabytes 300
 
+To ignore users who's data whose data must not be deleted, you can Tag those users
+and specify the `Tag` or parent `Tag Group` by `Name` or `ID`.
+This is enabled by default using a Tag named "NO DELETE".
+So it is preferable to Tag users on the server with a `Tag` named "NO DELETE" or create
+a `Tag Group` named "NO DELETE" containing Tags linked to users.
+
+    # Add a Tag to a User via CLI (not possible to see this in the clients)
+    $ omero obj new ExperimenterAnnotationLink child=TagAnnotation:123 parent=Experimenter:52
+
+    # Choose a non-default Tag or Tag Group (by ID or Name) to ignore the tagged users
+    $ omero demo-cleanup --gigabytes 300 --ignore-tag "Tag Name"
+
+You can also specify individual users by ID or user name, e.g:
+
+    --ignore-users 123,user-1,ben,234
+
 To generate the list of users which data must be deleted to free 300GB on the
 system and running the deletion (WARNING: data belonging to these users will
 be removed permanently)::

src/omero_demo_cleanup/cli.py

@@ -25,7 +25,13 @@
 
 from omero.cli import BaseControl, Parser
 from omero.gateway import BlitzGateway
-from omero_demo_cleanup.library import choose_users, delete_data, resource_usage
+from omero_demo_cleanup.library import (
+    choose_users,
+    delete_data,
+    resource_usage,
+    users_by_id_or_username,
+    users_by_tag,
+)
 
 HELP = """Cleanup disk space on OMERO.server """
 
@@ -85,11 +91,20 @@ def _configure(self, parser: Parser) -> None:
             help="Perform the data deletion rather than running in dry-run mode."
             " Default: false.",
         )
+        parser.add_argument(
+            "--ignore-tag",
+            "-t",
+            default="NO DELETE",
+            help="Members tagged with this Tag (Name or ID) or child Tags are ignored.",
+        )
+        parser.add_argument(
+            "--ignore-users",
+            help="Ingore users: Comma-separated IDs and/or user-names.",
+        )
         parser.set_defaults(func=self.cleanup)
 
     @gateway_required
     def cleanup(self, args: argparse.Namespace) -> None:
-
         if args.inodes == 0 and args.gigabytes == 0:
             self.ctx.die(23, "Please specify how much to delete")
 
@@ -111,7 +126,11 @@ def cleanup(self, args: argparse.Namespace) -> None:
                     )
                 )
 
-            stats = resource_usage(self.gateway, minimum_days=args.days)
+            ignore = users_by_tag(self.gateway, args.ignore_tag)
+            ignore.extend(users_by_id_or_username(self.gateway, args.ignore_users))
+            stats = resource_usage(
+                self.gateway, minimum_days=args.days, ignore_users=ignore
+            )
             users = choose_users(args.inodes, args.gigabytes * 1000**3, stats)
             self.ctx.err(f"Found {len(users)} user(s) for deletion.")
             for user in users:

src/omero_demo_cleanup/library.py

@@ -36,7 +36,7 @@
     LegalGraphTargetsResponse,
 )
 from omero.gateway import BlitzGateway
-from omero.rtypes import rlong
+from omero.rtypes import rlong, unwrap
 from omero.sys import ParametersI
 
 # If adjusting UserStats, find_worst, choose_users then check with unit tests.
@@ -181,8 +181,78 @@ def delete_data(conn: BlitzGateway, user_id: int, dry_run: bool = True) -> None:
         submit(conn, delete, Delete2Response)
 
 
+def exp_to_str(exp):
+    # "user-3" (#6) Charles Darwin
+    full_name = f"{unwrap(exp.firstName)} {unwrap(exp.lastName)}"
+    return f'"{exp.omeName.val}" (#{exp.id.val}) {full_name}'
+
+
+def users_by_id_or_username(conn: BlitzGateway, ignore_users: str) -> List[int]:
+
+    if not ignore_users:
+        return []
+    exclude = []
+    users = ignore_users.split(",")
+    print(f"Ignoring {len(users)} users by ID or Username:")
+    for user_str in users:
+        if user_str.isnumeric():
+            exp = conn.getQueryService().get("Experimenter", int(user_str))
+            print("  " + exp_to_str(exp))
+            exclude.append(exp.id.val)
+        else:
+            exp = conn.getObject("Experimenter", attributes={"omeName": user_str})
+            if exp is None:
+                raise ValueError("Experimenter: %s not found" % user_str)
+            print("  " + exp_to_str(exp._obj))
+            exclude.append(exp.id)
+    return exclude
+
+
+def users_by_tag(conn: BlitzGateway, tag_name: str) -> List[int]:
+    # Get users linked to Tag (Name or ID) or linked to child Tags.
+    if not tag_name or tag_name == "None":
+        print("No Tag chosen for ingoring users")
+        return []
+    exclude = []
+    if tag_name.isnumeric():
+        tag = conn.getObject("Annotation", tag_name)
+    else:
+        tags = list(
+            conn.getObjects("TagAnnotation", attributes={"textValue": tag_name})
+        )
+        tag = tags[0] if len(tags) > 0 else None
+        if len(tags) > 1:
+            ids = [tag.id for tag in tags]
+            raise ValueError(f"Multiple Tags with name: {tag_name} ({ids})")
+    if tag is None:
+        raise ValueError("Tag: %s not found" % tag_name)
+    # Check if this is a Tag Group
+    tag_links = list(conn.getAnnotationLinks("Annotation", parent_ids=[tag.id]))
+
+    # Handle Tagged Experimenters first...
+    links = list(conn.getAnnotationLinks("Experimenter", ann_ids=[tag.id]))
+    exclude.extend([link.parent.id.val for link in links])
+    # If we have NO child Tags, then always print:
+    if len(links) > 0 or len(tag_links) == 0:
+        print(
+            "Ignoring %s users linked to Tag:%s %s:"
+            % (len(links), tag.id, tag.textValue)
+        )
+
+    for link in links:
+        print("  " + exp_to_str(link.parent))
+
+    # Then recursively check any child Tags...
+    if len(tag_links) > 0 or len(links) == 0:
+        print(f"Tag:{tag.id} {tag.textValue} has {len(tag_links)} child Tags...")
+    for link in tag_links:
+        exclude.extend(users_by_tag(conn, str(link.child.id.val)))
+
+    return exclude
+
+
 def find_users(
-    conn: BlitzGateway, minimum_days: int = 0
+    conn: BlitzGateway, minimum_days: int = 0, ignore_users: List[int] = []
 ) -> Tuple[Dict[int, str], Dict[int, int]]:
     # Determine which users' data to consider deleting.
 
@@ -194,7 +264,8 @@ def find_users(
         user_id = result[0].val
         user_name = result[1].val
         if user_name not in ("PUBLIC", "guest", "root", "monitoring"):
-            users[user_id] = user_name
+            if user_id not in ignore_users:
+                users[user_id] = user_name
 
     for result in conn.getQueryService().projection(
         "SELECT DISTINCT owner.id FROM Session WHERE closed IS NULL", None
@@ -235,12 +306,16 @@ def find_users(
     return users, logouts
 
 
-def resource_usage(conn: BlitzGateway, minimum_days: int = 0) -> List[UserStats]:
+def resource_usage(
+    conn: BlitzGateway, minimum_days: int = 0, ignore_users: List[int] = []
+) -> List[UserStats]:
     # Note users' resource usage.
     # DiskUsage2.targetClasses remains too inefficient so iterate.
 
     user_stats = []
-    users, logouts = find_users(conn, minimum_days=minimum_days)
+    users, logouts = find_users(
+        conn, minimum_days=minimum_days, ignore_users=ignore_users
+    )
     for user_id, user_name in users.items():
         print(f'Finding disk usage of "{user_name}" (#{user_id}).')
         user = {"Experimenter": [user_id]}
@@ -286,7 +361,6 @@ def perform_delete(
 
 
 def main() -> None:
-
     with omero.cli.cli_login() as cli:
         conn = omero.gateway.BlitzGateway(client_obj=cli.get_client())
         try: