fix sms and ticket user frindly

apps/notifications/management/commands/process_notifications.py

@@ -6,6 +6,7 @@ import jdatetime
 from django.conf import settings
 from django.core.management.base import BaseCommand
 from django.db import transaction
+from django.db.models import Q
 from django.utils import timezone
 
 from apps.accounts.models import TicketMagicLink
@@ -130,6 +131,103 @@ def _prepare_customer_reply_bale(job):
             locked.save(update_fields=["status","next_attempt_at","last_error"]); return False
         return True
 
+
+def _record_failure(job, error, provider_code=""):
+    """Persist a provider failure using the channel-specific retry policy."""
+    job.attempts += 1
+    job.last_error = str(error)[:1000]
+    job.provider_status_code = str(provider_code or "")[:50]
+    job.provider_message_id = ""
+    if job.channel == NotificationJob.Channel.SMS:
+        if job.attempts < 1 + settings.SMS_MAX_RETRIES:
+            job.status = NotificationJob.Status.RETRYING
+            job.next_attempt_at = timezone.now() + timedelta(seconds=settings.SMS_RETRY_DELAY_SECONDS)
+        else:
+            job.status = NotificationJob.Status.FAILED
+    else:
+        job.status = NotificationJob.Status.FAILED
+        job.next_attempt_at = timezone.now() + timedelta(seconds=min(3600, 2**job.attempts * 10))
+
+
+def _record_success(job, result):
+    job.attempts += 1
+    job.status = NotificationJob.Status.SENT
+    job.sent_at = timezone.now()
+    job.last_error = ""
+    job.provider_message_id = getattr(result, "provider_id", "") if result is not None else ""
+    job.provider_status_code = getattr(result, "provider_code", "") if result is not None else ""
+    if job.channel == NotificationJob.Channel.SMS and job.provider_message_id:
+        job.delivery_status = NotificationJob.DeliveryStatus.PENDING
+        job.delivery_next_check_at = timezone.now() + timedelta(seconds=settings.SMS_DELIVERY_CHECK_DELAY_SECONDS)
+
+
+def _process_delivery_checks(options):
+    if options.get("channel") not in (None, NotificationJob.Channel.SMS):
+        return
+    now = timezone.now()
+    with transaction.atomic():
+        pollable_statuses = [
+            NotificationJob.DeliveryStatus.PENDING,
+            NotificationJob.DeliveryStatus.CHECK_FAILED,
+            NotificationJob.DeliveryStatus.UNKNOWN,
+            NotificationJob.DeliveryStatus.UNDELIVERED,
+        ]
+        # Resume provisional results created by older worker versions, but do
+        # not revive messages beyond the configured 45-minute monitoring span.
+        monitor_cutoff = now - timedelta(seconds=settings.SMS_DELIVERY_CHECK_DELAY_SECONDS * settings.SMS_DELIVERY_MAX_CHECKS)
+        NotificationJob.objects.filter(
+            channel=NotificationJob.Channel.SMS,
+            status=NotificationJob.Status.SENT,
+            delivery_status__in=[NotificationJob.DeliveryStatus.UNKNOWN, NotificationJob.DeliveryStatus.UNDELIVERED],
+            delivery_next_check_at__isnull=True,
+            delivery_check_attempts__lt=settings.SMS_DELIVERY_MAX_CHECKS,
+            sent_at__gte=monitor_cutoff,
+        ).update(delivery_next_check_at=now)
+        queryset = NotificationJob.objects.select_for_update(skip_locked=True).filter(
+            channel=NotificationJob.Channel.SMS,
+            status=NotificationJob.Status.SENT,
+            delivery_status__in=pollable_statuses,
+            delivery_next_check_at__lte=now,
+        )
+        if options.get("job_id"):
+            queryset = queryset.filter(pk=options["job_id"])
+        elif options.get("min_id"):
+            queryset = queryset.filter(pk__gte=options["min_id"])
+        jobs = list(queryset.order_by("delivery_next_check_at")[:20])
+        # Lease each status check so parallel workers cannot query it twice.
+        NotificationJob.objects.filter(pk__in=[job.pk for job in jobs]).update(
+            delivery_next_check_at=now + timedelta(minutes=10),
+        )
+    provider = get_sms_provider()
+    for job in jobs:
+        result = provider.get_delivery_status(job.provider_message_id)
+        if result is None:
+            job.delivery_status = ""
+            job.delivery_next_check_at = None
+            job.delivery_description = "این ارائه‌دهنده از استعلام دلیوری پشتیبانی نمی‌کند."
+        else:
+            job.delivery_check_attempts += 1
+            job.delivery_checked_at = timezone.now()
+            job.delivery_status_code = result.provider_code[:50]
+            job.delivery_description = result.description[:500]
+            if result.is_final:
+                job.delivery_status = result.status
+                job.delivery_next_check_at = None
+            elif job.delivery_check_attempts < settings.SMS_DELIVERY_MAX_CHECKS:
+                job.delivery_status = result.status
+                job.delivery_next_check_at = timezone.now() + timedelta(seconds=settings.SMS_DELIVERY_CHECK_DELAY_SECONDS)
+            else:
+                job.delivery_status = result.status if result.status in NotificationJob.DeliveryStatus.values else NotificationJob.DeliveryStatus.ERROR
+                job.delivery_next_check_at = None
+                job.delivery_description = f"{result.description} نتیجه پس از {job.delivery_check_attempts} نوبت استعلام در بازه پایش نهایی شد."[:500]
+        job.save(update_fields=[
+            "delivery_status", "delivery_status_code", "delivery_description",
+            "delivery_checked_at", "delivery_check_attempts", "delivery_next_check_at",
+        ])
+        if job.ticket_id and job.ticket.group_dispatch_id:
+            from apps.tickets.services import refresh_group_dispatch_status
+            refresh_group_dispatch_status(job.ticket.group_dispatch_id)
+
 class Command(BaseCommand):
     help = "Process notification outbox"
     def add_arguments(self, parser):
@@ -140,9 +238,15 @@ class Command(BaseCommand):
     def handle(self, *args, **options):
         while True:
             stale = timezone.now() - timedelta(minutes=10)
-            NotificationJob.objects.filter(status="PROCESSING", next_attempt_at__lt=stale).update(status="FAILED", next_attempt_at=timezone.now(), last_error="Recovered stale processing job")
+            NotificationJob.objects.filter(status="PROCESSING", next_attempt_at__lt=stale, channel=NotificationJob.Channel.SMS).update(status=NotificationJob.Status.RETRYING, next_attempt_at=timezone.now(), last_error="بازیابی تلاش ارسال متوقف‌شده")
+            NotificationJob.objects.filter(status="PROCESSING", next_attempt_at__lt=stale, channel=NotificationJob.Channel.BALE).update(status=NotificationJob.Status.FAILED, next_attempt_at=timezone.now(), last_error="Recovered stale processing job")
             with transaction.atomic():
-                queryset = NotificationJob.objects.select_for_update(skip_locked=True).filter(status__in=["PENDING", "FAILED"], attempts__lt=settings.MAX_NOTIFICATION_ATTEMPTS, next_attempt_at__lte=timezone.now())
+                sms_limit = 1 + settings.SMS_MAX_RETRIES
+                queryset = NotificationJob.objects.select_for_update(skip_locked=True).filter(
+                    Q(channel=NotificationJob.Channel.SMS, attempts__lt=sms_limit, status__in=[NotificationJob.Status.PENDING, NotificationJob.Status.RETRYING])
+                    | Q(channel=NotificationJob.Channel.BALE, attempts__lt=settings.MAX_NOTIFICATION_ATTEMPTS, status__in=[NotificationJob.Status.PENDING, NotificationJob.Status.FAILED]),
+                    next_attempt_at__lte=timezone.now(),
+                )
                 if options.get("channel"):
                     queryset = queryset.filter(channel=options["channel"])
                 if options.get("job_id"):
@@ -160,15 +264,22 @@ class Command(BaseCommand):
                         continue
                     if job.channel == NotificationJob.Channel.SMS: result, issued_link = _send_sms(job)
                     else: result = _send_bale(job)
-                    if result is not None and not result.success: raise RuntimeError(result.error or "Provider rejected notification")
-                    job.attempts += 1; job.status = "SENT"; job.sent_at = timezone.now(); job.last_error = ""; job.provider_message_id = getattr(result,"provider_id","") if result is not None else ""
+                    if result is not None and not result.success:
+                        _record_failure(job, result.error or "ارائه‌دهنده پیام را نپذیرفت.", getattr(result, "provider_code", ""))
+                        if issued_link:
+                            issued_link.revoked_at = timezone.now()
+                            issued_link.save(update_fields=["revoked_at"])
+                    else:
+                        _record_success(job, result)
                 except Exception as exc:
                     if issued_link:
                         issued_link.revoked_at = timezone.now(); issued_link.save(update_fields=["revoked_at"])
-                    job.attempts += 1; job.last_error = str(exc)[:1000]; job.provider_message_id = ""; job.status = "FAILED"; job.next_attempt_at = timezone.now() + timedelta(seconds=min(3600, 2**job.attempts * 10)); logger.exception("notification failed id=%s", job.pk)
+                    _record_failure(job, exc)
+                    logger.exception("notification failed id=%s", job.pk)
                 job.save()
                 if job.ticket_id and job.ticket.group_dispatch_id:
                     from apps.tickets.services import refresh_group_dispatch_status
                     refresh_group_dispatch_status(job.ticket.group_dispatch_id)
+            _process_delivery_checks(options)
             if options["once"] or options.get("job_id"): break
             time.sleep(5)