Merge pull request #1216 from LibraryOfCongress/more-database-optimiz…

…ations

WIP: Add more combined indexes for popular queries

concordia/migrations/0048_auto_20200324_1820.py

@@ -0,0 +1,43 @@
+# Generated by Django 2.2.11 on 2020-03-24 22:20
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("concordia", "0047_auto_20200324_1103"),
+    ]
+
+    operations = [
+        migrations.RemoveIndex(model_name="asset", name="concordia_a_id_0c37bf_idx",),
+        migrations.AlterField(
+            model_name="item",
+            name="published",
+            field=models.BooleanField(blank=True, default=False),
+        ),
+        migrations.AlterField(
+            model_name="project",
+            name="published",
+            field=models.BooleanField(blank=True, default=False),
+        ),
+        migrations.AddIndex(
+            model_name="asset",
+            index=models.Index(
+                fields=["id", "item", "published", "transcription_status"],
+                name="concordia_a_id_137ca8_idx",
+            ),
+        ),
+        migrations.AddIndex(
+            model_name="item",
+            index=models.Index(
+                fields=["project", "published"], name="concordia_i_project_d8caf0_idx"
+            ),
+        ),
+        migrations.AddIndex(
+            model_name="project",
+            index=models.Index(
+                fields=["id", "campaign", "published"], name="concordia_p_id_17c9c9_idx"
+            ),
+        ),
+    ]

concordia/models.py

@@ -158,7 +158,7 @@ class Project(MetricsModelMixin("project"), models.Model):
 
     campaign = models.ForeignKey(Campaign, on_delete=models.CASCADE)
 
-    published = models.BooleanField(default=False, blank=True, db_index=True)
+    published = models.BooleanField(default=False, blank=True)
 
     title = models.CharField(max_length=80)
     slug = models.SlugField(max_length=80, allow_unicode=True)
@@ -174,6 +174,7 @@ class Project(MetricsModelMixin("project"), models.Model):
     class Meta:
         unique_together = (("slug", "campaign"),)
         ordering = ["title"]
+        indexes = [models.Index(fields=["id", "campaign", "published"])]
 
     def __str__(self):
         return self.title
@@ -190,7 +191,7 @@ class Item(MetricsModelMixin("item"), models.Model):
 
     project = models.ForeignKey(Project, on_delete=models.CASCADE)
 
-    published = models.BooleanField(default=False, blank=True, db_index=True)
+    published = models.BooleanField(default=False, blank=True)
 
     title = models.CharField(max_length=600)
     item_url = models.URLField(max_length=255)
@@ -208,6 +209,7 @@ class Item(MetricsModelMixin("item"), models.Model):
 
     class Meta:
         unique_together = (("item_id", "project"),)
+        indexes = [models.Index(fields=["project", "published"])]
 
     def __str__(self):
         return f"{self.item_id}: {self.title}"
@@ -276,7 +278,7 @@ class Asset(MetricsModelMixin("asset"), models.Model):
     class Meta:
         unique_together = (("slug", "item"),)
         indexes = [
-            models.Index(fields=["id", "published", "transcription_status"]),
+            models.Index(fields=["id", "item", "published", "transcription_status"]),
             models.Index(fields=["published", "transcription_status"]),
         ]
 

concordia/views.py

@@ -1511,7 +1511,9 @@ def filter_and_order_transcribable_assets(
         | Q(transcription_status=TranscriptionStatus.IN_PROGRESS)
     )
 
-    potential_assets = potential_assets.filter(assettranscriptionreservation=None)
+    potential_assets = potential_assets.exclude(
+        pk__in=Subquery(AssetTranscriptionReservation.objects.values("asset_id"))
+    )
     potential_assets = potential_assets.select_related("item", "item__project")
 
     # We'll favor assets which are in the same item or project as the original:
@@ -1544,7 +1546,9 @@ def filter_and_order_reviewable_assets(
         transcription_status=TranscriptionStatus.SUBMITTED
     )
     potential_assets = potential_assets.exclude(transcription__user=user_pk)
-    potential_assets = potential_assets.filter(assettranscriptionreservation=None)
+    potential_assets = potential_assets.exclude(
+        pk__in=Subquery(AssetTranscriptionReservation.objects.values("asset_id"))
+    )
     potential_assets = potential_assets.select_related("item", "item__project")
 
     # We'll favor assets which are in the same item or project as the original: