[ENG-8515] - Add management command to manual archive (#11361)

* Add management command to manual archive

* fix test

* Fix lints

* Fix tests

Author: Vlad0n20

admin/nodes/views.py

@@ -1,33 +1,31 @@
-import pytz
-from enum import Enum
 from datetime import datetime
-from framework import status
+from enum import Enum
 
-from django.utils import timezone
-from django.core.exceptions import PermissionDenied, ValidationError
-from django.urls import NoReverseMatch
+import pytz
 from django.db import transaction
-from django.db.models import F, Case, When, IntegerField
 from django.contrib import messages
 from django.contrib.auth.mixins import PermissionRequiredMixin
+from django.core.exceptions import PermissionDenied, ValidationError
+from django.db.models import F, Case, When, IntegerField
 from django.http import HttpResponse
+from django.shortcuts import redirect, reverse, get_object_or_404
+from django.urls import NoReverseMatch
+from django.urls import reverse_lazy
+from django.utils import timezone
 from django.views.generic import (
     View,
     FormView,
     ListView,
 )
-from django.shortcuts import redirect, reverse, get_object_or_404
-from django.urls import reverse_lazy
 
+from admin.base.forms import GuidForm
 from admin.base.utils import change_embargo_date
 from admin.base.views import GuidView
-from admin.base.forms import GuidForm
-from admin.notifications.views import delete_selected_notifications
 from admin.nodes.forms import AddSystemTagForm, RegistrationDateForm
-
-from api.share.utils import update_share
+from admin.notifications.views import delete_selected_notifications
 from api.caching.tasks import update_storage_usage_cache
-
+from api.share.utils import update_share
+from framework import status
 from osf.exceptions import NodeStateError, RegistrationStuckError
 from osf.management.commands.change_node_region import _update_schema_meta
 from osf.models import (
@@ -53,9 +51,7 @@
     REINDEX_ELASTIC,
 )
 from osf.utils.permissions import ADMIN, API_CONTRIBUTOR_PERMISSIONS
-
 from scripts.approve_registrations import approve_past_pendings
-
 from website import settings, search
 from website.archiver.tasks import force_archive
 
@@ -149,7 +145,8 @@ def get_context_data(self, **kwargs):
             'STORAGE_LIMITS': settings.StorageLimits,
             'node': node,
             # to edit contributors we should have guid as django prohibits _id usage as it starts with an underscore
-            'annotated_contributors': node.contributor_set.prefetch_related('user__guids').annotate(guid=F('user__guids___id')),
+            'annotated_contributors': node.contributor_set.prefetch_related('user__guids').annotate(
+                guid=F('user__guids___id')),
             'children': children,
             'permissions': API_CONTRIBUTOR_PERMISSIONS,
             'has_update_permission': self.request.user.has_perm('osf.change_node'),
@@ -209,7 +206,9 @@ class NodeRemoveContributorView(NodeMixin, View):
     def post(self, request, *args, **kwargs):
         node = self.get_object()
         user = OSFUser.objects.get(id=self.kwargs.get('user_id'))
-        if node.has_permission(user, ADMIN) and not node._get_admin_contributors_query(node._contributors.all(), require_active=False).exclude(user=user).exists():
+        if node.has_permission(user, ADMIN) and not node._get_admin_contributors_query(node._contributors.all(),
+                                                                                       require_active=False).exclude(
+                user=user).exists():
             messages.error(self.request, 'Must be at least one admin on this node.')
             return redirect(self.get_success_url())
 
@@ -906,6 +905,7 @@ class ForceArchiveRegistrationsView(NodeMixin, View):
     def post(self, request, *args, **kwargs):
         # Prevents circular imports that cause admin app to hang at startup
         from osf.management.commands.force_archive import verify, DEFAULT_PERMISSIBLE_ADDONS
+        from osf.models.admin_log_entry import update_admin_log, MANUAL_ARCHIVE_RESTART
 
         registration = self.get_object()
         force_archive_params = request.POST
@@ -933,16 +933,31 @@ def post(self, request, *args, **kwargs):
                 messages.error(request, str(exc))
                 return redirect(self.get_success_url())
         else:
-            # For actual archiving, skip synchronous verification to avoid 502 timeouts
-            # Verification will be performed asynchronously in the task
-            force_archive_task = force_archive.delay(
-                str(registration._id),
-                permissible_addons=list(addons),
-                allow_unconfigured=allow_unconfigured,
-                skip_collisions=skip_collision,
-                delete_collisions=delete_collision,
-            )
-            messages.success(request, f'Registration archive process has started. Task id: {force_archive_task.id}.')
+            try:
+                update_admin_log(
+                    user_id=request.user.id,
+                    object_id=registration.pk,
+                    object_repr=str(registration),
+                    message=f'Manual archive restart initiated for registration {registration._id}',
+                    action_flag=MANUAL_ARCHIVE_RESTART
+                )
+                # For actual archiving, skip synchronous verification to avoid 502 timeouts
+                # Verification will be performed asynchronously in the task
+                force_archive_task = force_archive.delay(
+                    str(registration._id),
+                    permissible_addons=list(addons),
+                    allow_unconfigured=allow_unconfigured,
+                    skip_collisions=skip_collision,
+                    delete_collisions=delete_collision,
+                )
+                messages.success(
+                    request,
+                    f'Registration archive process has started. Task id: {force_archive_task.id}.'
+                )
+            except Exception as exc:
+                messages.error(request,
+                               f'This registration cannot be archived due to {exc.__class__.__name__}: {str(exc)}. '
+                               f'If the problem persists get a developer to fix it.')
 
         return redirect(self.get_success_url())
 

api/institutions/serializers.py

@@ -231,7 +231,9 @@ class Meta:
     number_of_users = ser.IntegerField(read_only=True)
 
     def get_absolute_url(self, obj):
-        raise NotImplementedError()
+        institution_id = self.context['request'].parser_context['kwargs']['institution_id']
+        dept_id = obj['name'].replace(' ', '-')
+        return f'/institutions/{institution_id}/metrics/departments/{dept_id}/'
 
 
 class InstitutionUserMetricsSerializer(JSONAPISerializer):
@@ -289,7 +291,8 @@ def get_contacts(self, obj):
         return list(results)
 
     def get_absolute_url(self, obj):
-        raise NotImplementedError()
+        institution_id = self.context['request'].parser_context['kwargs']['institution_id']
+        return f'/institutions/{institution_id}/metrics/users/'
 
 
 class InstitutionSummaryMetricsSerializer(JSONAPISerializer):
@@ -323,7 +326,8 @@ class Meta:
     )
 
     def get_absolute_url(self, obj):
-        raise NotImplementedError()
+        institution_id = self.context['request'].parser_context['kwargs']['institution_id']
+        return f'/institutions/{institution_id}/metrics/summary/'
 
 
 class InstitutionRelated(JSONAPIRelationshipSerializer):

osf/management/commands/process_manual_restart_approvals.py

@@ -0,0 +1,145 @@
+import logging
+from datetime import timedelta
+from django.core.management.base import BaseCommand
+from django.utils import timezone
+from osf.models import Registration
+from osf.models.admin_log_entry import AdminLogEntry, MANUAL_ARCHIVE_RESTART
+from website import settings
+from scripts.approve_registrations import approve_past_pendings
+
+logger = logging.getLogger(__name__)
+
+
+class Command(BaseCommand):
+    help = 'Process registrations that were manually restarted and may need approval'
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            '--dry-run',
+            action='store_true',
+            help='Show what would be done without actually doing it',
+        )
+        parser.add_argument(
+            '--hours-back',
+            type=int,
+            default=72,
+            help='How many hours back to look for manual restarts (default: 72)',
+        )
+        parser.add_argument(
+            '--registration-id',
+            type=str,
+            help='Process a specific registration ID only',
+        )
+
+    def handle(self, *args, **options):
+        dry_run = options['dry_run']
+        hours_back = options['hours_back']
+        specific_registration = options.get('registration_id')
+
+        if dry_run:
+            self.stdout.write(self.style.WARNING('Running in DRY RUN mode - no changes will be made'))
+
+        since = timezone.now() - timedelta(hours=hours_back)
+
+        query = AdminLogEntry.objects.filter(
+            action_flag=MANUAL_ARCHIVE_RESTART,
+            action_time__gte=since
+        )
+
+        if specific_registration:
+            try:
+                reg = Registration.objects.get(_id=specific_registration)
+                query = query.filter(object_id=reg.pk)
+                self.stdout.write(f"Processing specific registration: {specific_registration}")
+            except Registration.DoesNotExist:
+                self.stdout.write(self.style.ERROR(f"Registration {specific_registration} not found"))
+                return
+
+        manual_restart_logs = query.values_list('object_id', flat=True).distinct()
+
+        registrations_to_check = Registration.objects.filter(
+            pk__in=manual_restart_logs,
+        )
+
+        self.stdout.write(f"Found {registrations_to_check.count()} manually restarted registrations to check")
+
+        approvals_ready = []
+        skipped_registrations = []
+
+        for registration in registrations_to_check:
+            status = self.should_auto_approve(registration)
+
+            if status == 'ready':
+                approval = registration.registration_approval
+                if approval:
+                    approvals_ready.append(approval)
+                    self.stdout.write(
+                        self.style.SUCCESS(f"✓ Queuing registration {registration._id} for approval")
+                    )
+            else:
+                skipped_registrations.append((registration._id, status))
+                self.stdout.write(
+                    self.style.WARNING(f"⚠ Skipping registration {registration._id}: {status}")
+                )
+
+        if approvals_ready:
+            if dry_run:
+                self.stdout.write(
+                    self.style.WARNING(f"DRY RUN: Would approve {len(approvals_ready)} registrations")
+                )
+            else:
+                try:
+                    approve_past_pendings(approvals_ready, dry_run=False)
+                    self.stdout.write(
+                        self.style.SUCCESS(f"✓ Successfully approved {len(approvals_ready)} manually restarted registrations")
+                    )
+                except Exception as e:
+                    self.stdout.write(
+                        self.style.ERROR(f"✗ Error approving registrations: {e}")
+                    )
+        else:
+            self.stdout.write('No registrations ready for approval')
+
+        self.stdout.write(f"Total checked: {registrations_to_check.count()}")
+        self.stdout.write(f"Ready for approval: {len(approvals_ready)}")
+        self.stdout.write(f"Skipped: {len(skipped_registrations)}")
+
+        if skipped_registrations:
+            self.stdout.write('\nSkipped registrations:')
+            for reg_id, reason in skipped_registrations:
+                self.stdout.write(f"  - {reg_id}: {reason}")
+
+    def should_auto_approve(self, registration):
+        if registration.is_public:
+            return 'already public'
+
+        if registration.is_registration_approved:
+            return 'already approved'
+
+        if registration.archiving:
+            return 'still archiving'
+
+        archive_job = registration.archive_job
+        if archive_job and hasattr(archive_job, 'status'):
+            if archive_job.status not in ['SUCCESS', None]:
+                return f'archive status: {archive_job.status}'
+
+        approval = registration.registration_approval
+        if not approval:
+            return 'no approval object'
+
+        if approval.is_approved:
+            return 'approval already approved'
+
+        if approval.is_rejected:
+            return 'approval was rejected'
+
+        time_since_initiation = timezone.now() - approval.initiation_date
+        if time_since_initiation < settings.REGISTRATION_APPROVAL_TIME:
+            remaining = settings.REGISTRATION_APPROVAL_TIME - time_since_initiation
+            return f'not ready yet ({remaining} remaining)'
+
+        if registration.is_stuck_registration:
+            return 'registration still stuck'
+
+        return 'ready'

osf/models/admin_log_entry.py

@@ -37,6 +37,8 @@
 DOI_CREATION_FAILED = 80
 DOI_UPDATE_FAILED = 81
 
+MANUAL_ARCHIVE_RESTART = 90
+
 def update_admin_log(user_id, object_id, object_repr, message, action_flag=UNKNOWN):
     AdminLogEntry.objects.log_action(
         user_id=user_id,

scripts/check_manual_restart_approval.py

@@ -0,0 +1,74 @@
+import logging
+from framework.celery_tasks import app as celery_app
+from django.core.management import call_command
+from osf.models import Registration
+
+logger = logging.getLogger(__name__)
+
+
+@celery_app.task(name='scripts.check_manual_restart_approval')
+def check_manual_restart_approval(registration_id):
+    try:
+        try:
+            registration = Registration.objects.get(_id=registration_id)
+        except Registration.DoesNotExist:
+            logger.error(f"Registration {registration_id} not found")
+            return f"Registration {registration_id} not found"
+
+        if registration.is_public or registration.is_registration_approved:
+            return f"Registration {registration_id} already approved/public"
+
+        if registration.archiving:
+            logger.info(f"Registration {registration_id} still archiving, retrying in 10 minutes")
+            check_manual_restart_approval.apply_async(
+                args=[registration_id],
+                countdown=600
+            )
+            return f"Registration {registration_id} still archiving, scheduled retry"
+
+        logger.info(f"Processing manual restart approval for registration {registration_id}")
+
+        call_command(
+            'process_manual_restart_approvals',
+            registration_id=registration_id,
+            dry_run=False,
+            hours_back=24,
+            verbosity=1
+        )
+
+        return f"Processed manual restart approval check for registration {registration_id}"
+
+    except Exception as e:
+        logger.error(f"Error processing manual restart approval for {registration_id}: {e}")
+        raise
+
+
+@celery_app.task(name='scripts.check_manual_restart_approvals_batch')
+def check_manual_restart_approvals_batch(hours_back=24):
+    try:
+        logger.info(f"Running batch check for manual restart approvals (last {hours_back} hours)")
+
+        call_command(
+            'process_manual_restart_approvals',
+            dry_run=False,
+            hours_back=hours_back,
+            verbosity=1
+        )
+
+        return f"Completed batch manual restart approval check for last {hours_back} hours"
+
+    except Exception as e:
+        logger.error(f"Error in batch manual restart approval check: {e}")
+        raise
+
+
+@celery_app.task(name='scripts.delayed_manual_restart_approval')
+def delayed_manual_restart_approval(registration_id, delay_minutes=30):
+    logger.info(f"Scheduling delayed manual restart approval check for {registration_id} in {delay_minutes} minutes")
+
+    check_manual_restart_approval.apply_async(
+        args=[registration_id],
+        countdown=delay_minutes * 60
+    )
+
+    return f"Scheduled manual restart approval check for {registration_id} in {delay_minutes} minutes"