cache usage state in api

common/views.py

@@ -1,6 +1,7 @@
 from django.conf import settings
 from django.contrib import messages
 from django.contrib.auth.decorators import login_required
+from django.core.cache import cache
 from django.db import connection
 from django.http import JsonResponse
 from django.shortcuts import redirect, render
@@ -37,16 +38,7 @@ def ap_redirect(request, uri):
 
 
 def nodeinfo2(request):
-    usage = {"users": {"total": User.objects.count()}}
-    # return estimated number of marks as posts, since count the whole table is slow
-    # TODO filter local with SQL function in https://wiki.postgresql.org/wiki/Count_estimate
-    with connection.cursor() as cursor:
-        cursor.execute(
-            "SELECT n_live_tup FROM pg_stat_all_tables WHERE relname = 'journal_shelflogentry';"
-        )
-        row = cursor.fetchone()
-        if row:
-            usage["localPosts"] = row[0]
+    usage = cache.get("nodeinfo_usage") or {}
     return JsonResponse(
         {
             "version": "2.0",

takahe/apps.py

@@ -4,3 +4,7 @@ from django.apps import AppConfig
 class TakaheConfig(AppConfig):
     default_auto_field = "django.db.models.BigAutoField"
     name = "takahe"
+
+    def ready(self):
+        # register cron jobs
+        from .jobs import TakaheStats  # isort:skip

takahe/jobs.py

@@ -0,0 +1,62 @@
+from datetime import timedelta
+
+from django.core.cache import cache
+from django.db.models import Count, Q
+from django.utils import timezone
+from loguru import logger
+
+from common.models import BaseJob, JobManager
+from journal.models import Comment, Review, ShelfMember
+from mastodon.api import detect_server_info
+from mastodon.models import MastodonApplication
+from takahe.models import Domain, Identity, Post
+
+
+@JobManager.register
+class TakaheStats(BaseJob):
+    interval = timedelta(hours=6)
+    max_unreachable_days = 31
+
+    def active_users(self, d: int) -> int:
+        return (
+            ShelfMember.objects.filter(
+                created_time__gte=timezone.now() - timedelta(days=d), local=True
+            )
+            .values("owner_id")
+            .distinct()
+            .count()
+        )
+
+    def run(self):
+        logger.info("Updating Tahake stats.")
+        # for /api/v1/instance
+        stats = {
+            "user_count": Identity.objects.filter(
+                local=True, deleted__isnull=True
+            ).count(),
+            "status_count": Post.objects.filter(local=True)
+            .exclude(state__in=["deleted", "deleted_fanned_out"])
+            .count(),
+            "domain_count": Domain.objects.count(),
+        }
+        cache.set("instance_info_stats", stats)
+        # for /api/v2/instance
+        usage = {
+            "users": {
+                "active_month": self.active_users(30),
+            }
+        }
+        cache.set("instance_info_usage", usage)
+        # for NodeInfo
+        nodeinfo_usage = {
+            "users": {
+                "total": stats["user_count"],
+                "activeHalfyear": usage["users"]["active_month"],
+                "activeMonth": self.active_users(180),
+            },
+            "localPosts": ShelfMember.objects.filter(local=True).count(),
+            "localComments": Comment.objects.filter(local=True).count()
+            + Review.objects.filter(local=True).count(),
+        }
+        cache.set("nodeinfo_usage", nodeinfo_usage)
+        logger.info(f"Tahake stats updated. {stats} {nodeinfo_usage}")