more discover mix

catalog/management/commands/discover.py

@@ -2,14 +2,16 @@ from django.core.management.base import BaseCommand
 from django.core.cache import cache
 import pprint
 from catalog.models import *
-from journal.models import ShelfMember, query_item_category, ItemCategory
+from journal.models import ShelfMember, query_item_category, ItemCategory, Comment
 from datetime import timedelta
 from django.utils import timezone
 from django.db.models import Count
 
 
-MAX_GALLERY_ITEMS = 64
-MIN_MARKS = 3
+MAX_ITEMS_PER_PERIOD = 12
+MIN_MARKS = 2
+MAX_DAYS_FOR_PERIOD = 64
+MIN_DAYS_FOR_PERIOD = 4
 
 
 class Command(BaseCommand):
@@ -22,16 +24,29 @@ class Command(BaseCommand):
             help="generate discover data",
         )
 
-    def get_popular_item_ids(self, category, days):
-        # self.stdout.write(f"Generating popular {category} items for {days} days...")
+    def get_popular_marked_item_ids(self, category, days, exisiting_ids):
         item_ids = [
             m["item_id"]
             for m in ShelfMember.objects.filter(query_item_category(category))
             .filter(created_time__gt=timezone.now() - timedelta(days=days))
+            .exclude(item_id__in=exisiting_ids)
             .values("item_id")
             .annotate(num=Count("item_id"))
             .filter(num__gte=MIN_MARKS)
-            .order_by("-num")[:MAX_GALLERY_ITEMS]
+            .order_by("-num")[:MAX_ITEMS_PER_PERIOD]
+        ]
+        return item_ids
+
+    def get_popular_commented_item_ids(self, category, days, exisiting_ids):
+        item_ids = [
+            m["item_id"]
+            for m in Comment.objects.filter(query_item_category(category))
+            .filter(created_time__gt=timezone.now() - timedelta(days=days))
+            .exclude(item_id__in=exisiting_ids)
+            .values("item_id")
+            .annotate(num=Count("item_id"))
+            .filter(num__gte=MIN_MARKS)
+            .order_by("-num")[:MAX_ITEMS_PER_PERIOD]
         ]
         return item_ids
 
@@ -55,20 +70,24 @@ class Command(BaseCommand):
             ]
             gallery_list = []
             for category in gallery_categories:
-                days = 128
+                days = MAX_DAYS_FOR_PERIOD
                 item_ids = []
-                while days > 4:
-                    ids = [
-                        i
-                        for i in self.get_popular_item_ids(category, days)
-                        if i not in item_ids
-                    ]
-                    if len(ids) > MAX_GALLERY_ITEMS // 5:
-                        ids = ids[: MAX_GALLERY_ITEMS // 5]
-                    self.stdout.write(f"{category} for last {days} days: {len(ids)}")
+                while days >= MIN_DAYS_FOR_PERIOD:
+                    ids = self.get_popular_marked_item_ids(category, days, item_ids)
+                    self.stdout.write(
+                        f"Marked {category} for last {days} days: {len(ids)}"
+                    )
                     item_ids = ids + item_ids
                     days //= 2
-                items = list(Item.objects.filter(id__in=item_ids))
+                if category == ItemCategory.Podcast:
+                    extra_ids = self.get_popular_commented_item_ids(
+                        ItemCategory.Podcast, MAX_DAYS_FOR_PERIOD, item_ids
+                    )
+                    self.stdout.write(
+                        f"Commented podcast for last {MAX_DAYS_FOR_PERIOD} days: {len(extra_ids)}"
+                    )
+                    item_ids = extra_ids + item_ids
+                items = [Item.objects.get(pk=i) for i in item_ids]
                 if category == ItemCategory.TV:
                     items = self.cleanup_shows(items)
                 gallery_list.append(