Removes project documents with no active users

Introduces an admin cleanup function that scans project documents
in the Redis database and deletes those associated with projects
that do not have any connected users via socketio. Helps reduce
unnecessary storage and maintains data consistency.

Author: matusdrobuliak66

services/web/server/src/simcore_service_webserver/projects/_project_document_service.py

@@ -3,13 +3,16 @@
 This module contains common utilities for building and versioning project documents.
 """
 
+import logging
+import re
 from typing import cast
 
 from aiohttp import web
 from models_library.api_schemas_webserver.projects import (
     ProjectDocument,
     ProjectDocumentVersion,
 )
+from models_library.api_schemas_webserver.socketio import SocketIORoomStr
 from models_library.projects import ProjectID, ProjectTemplateType
 from models_library.projects import ProjectType as ProjectTypeAPI
 from servicelib.redis import (
@@ -22,8 +25,11 @@
     get_redis_document_manager_client_sdk,
     get_redis_lock_manager_client_sdk,
 )
+from ..socketio._utils import get_socket_server
 from . import _projects_repository
 
+_logger = logging.getLogger(__name__)
+
 
 async def create_project_document_and_increment_version(
     app: web.Application, project_uuid: ProjectID
@@ -86,4 +92,70 @@ async def _create_project_document_and_increment_version() -> (
     return await _create_project_document_and_increment_version()
 
 
-async def remove_project_documents_as_admin(app: web.Application) -> None: ...
+async def remove_project_documents_as_admin(app: web.Application) -> None:
+    """Admin function to clean up project documents for projects with no connected users.
+
+    This function scans through all project documents in the Redis DOCUMENTS database,
+    checks if there are any users currently connected to the project room via socketio,
+    and removes documents that have no connected users.
+    """
+    # Get Redis document manager client to access the DOCUMENTS database
+    redis_client = get_redis_document_manager_client_sdk(app)
+
+    # Pattern to match project document keys - looking for keys that contain project UUIDs
+    project_document_pattern = "projects:*:version"
+
+    # Get socketio server instance
+    sio = get_socket_server(app)
+
+    projects_removed = 0
+
+    # Scan through all project document keys
+    async for key in redis_client.redis.scan_iter(
+        match=project_document_pattern, count=1000
+    ):
+        # Extract project UUID from the key pattern "projects:{project_uuid}:version"
+        key_str = key.decode("utf-8") if isinstance(key, bytes) else key
+        match = re.match(r"projects:([0-9a-f-]+):version", key_str)
+
+        if not match:
+            continue
+
+        project_uuid_str = match.group(1)
+        project_uuid = ProjectID(project_uuid_str)
+        project_room = SocketIORoomStr.from_project_id(project_uuid)
+
+        # Check if there are any users connected to this project room
+        try:
+            # Get all session IDs (socket IDs) in the project room
+            room_sessions = list(
+                sio.manager.get_participants(namespace="/", room=project_room)
+            )
+
+            # If no users are connected to this project room, remove the document
+            if not room_sessions:
+                await redis_client.redis.delete(key_str)
+                projects_removed += 1
+                _logger.info(
+                    "Removed project document for project %s (no connected users)",
+                    project_uuid,
+                )
+            else:
+                _logger.debug(
+                    "Project %s has %d connected users, keeping document",
+                    project_uuid,
+                    len(room_sessions),
+                )
+
+        except (KeyError, AttributeError, ValueError) as e:
+            _logger.warning(
+                "Failed to check room participants for project %s: %s",
+                project_uuid,
+                str(e),
+            )
+            continue
+
+    _logger.info(
+        "Project document cleanup completed: removed %d project documents",
+        projects_removed,
+    )