From 46cda5df5d2d735f7d7fced7526aa1c9c79f0435 Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Thu, 4 Jun 2026 14:50:23 -0400
Subject: [PATCH 1/6] Basic domain filter using regex

---
 documentcloud/addons/tests/test_views.py | 57 ++++++++++++++++++++++++
 documentcloud/addons/views.py            | 46 +++++++++++++++++++
 2 files changed, 103 insertions(+)

diff --git a/documentcloud/addons/tests/test_views.py b/documentcloud/addons/tests/test_views.py
index 21a285f4..da6cf4a9 100644
--- a/documentcloud/addons/tests/test_views.py
+++ b/documentcloud/addons/tests/test_views.py
@@ -309,6 +309,44 @@ def test_filter_site_absent_is_noop(self, client):
         assert response.status_code == status.HTTP_200_OK
         assert len(response.json()["results"]) == 3
 
+    def test_filter_domain(self, client):
+        """Filter runs by the host of the event's parameters.site"""
+        user = UserFactory()
+        matching_event = AddOnEventFactory(
+            user=user,
+            parameters={
+                "site": "https://www.nifc.gov/fire-information/statistics/wildfires",
+                "selector": "*",
+            },
+        )
+        other_event = AddOnEventFactory(
+            user=user, parameters={"site": "https://www.other.com/path"}
+        )
+        no_site_event = AddOnEventFactory(user=user, parameters={"selector": "*"})
+        matching_run = AddOnRunFactory(user=user, event=matching_event)
+        AddOnRunFactory(user=user, event=other_event)
+        AddOnRunFactory(user=user, event=no_site_event)
+        AddOnRunFactory(user=user, event=None)
+        client.force_authenticate(user=user)
+        # bare host and full origin both match, regardless of path
+        for domain in ("www.nifc.gov", "https://www.nifc.gov"):
+            response = client.get("/api/addon_runs/", {"domain": domain})
+            assert response.status_code == status.HTTP_200_OK
+            uuids = [r["uuid"] for r in response.json()["results"]]
+            assert uuids == [str(matching_run.uuid)], domain
+
+    def test_filter_domain_no_partial_host_match(self, client):
+        """The domain filter matches whole hosts, not substrings"""
+        user = UserFactory()
+        event = AddOnEventFactory(
+            user=user, parameters={"site": "https://www.nifc.gov.evil.com/path"}
+        )
+        AddOnRunFactory(user=user, event=event)
+        client.force_authenticate(user=user)
+        response = client.get("/api/addon_runs/", {"domain": "www.nifc.gov"})
+        assert response.status_code == status.HTTP_200_OK
+        assert response.json()["results"] == []
+
 
 @pytest.mark.django_db()
 class TestAddOnEventAPI:
@@ -351,6 +389,25 @@ def test_filter_site_no_match(self, client):
         assert response.status_code == status.HTTP_200_OK
         assert response.json()["results"] == []
 
+    def test_filter_domain(self, client):
+        """Filter events by the host of their parameters.site"""
+        user = UserFactory()
+        matching = AddOnEventFactory(
+            user=user,
+            parameters={
+                "site": "https://www.nifc.gov/fire-information/statistics/wildfires",
+                "selector": "*",
+            },
+        )
+        AddOnEventFactory(user=user, parameters={"site": "https://www.other.com/path"})
+        AddOnEventFactory(user=user, parameters={"selector": "*"})
+        client.force_authenticate(user=user)
+        for domain in ("www.nifc.gov", "https://www.nifc.gov"):
+            response = client.get("/api/addon_events/", {"domain": domain})
+            assert response.status_code == status.HTTP_200_OK
+            ids = [r["id"] for r in response.json()["results"]]
+            assert ids == [matching.pk], domain
+
     def test_filter_message(self, client):
         """Filter runs by message"""
         user = UserFactory()
diff --git a/documentcloud/addons/views.py b/documentcloud/addons/views.py
index acd03fd0..3e643b4f 100644
--- a/documentcloud/addons/views.py
+++ b/documentcloud/addons/views.py
@@ -30,6 +30,7 @@
 import hmac
 import json
 import logging
+import re
 from collections import defaultdict
 from datetime import timedelta
 from functools import lru_cache
@@ -65,6 +66,21 @@
 logger = logging.getLogger(__name__)
 
 
+def domain_site_regex(value):
+    """Build a regex matching `site` URLs whose host equals the given domain.
+
+    Accepts either a bare host (`www.nifc.gov`) or a full URL
+    (`https://www.nifc.gov/path`) and matches stored site values regardless of
+    scheme, port, path or query string. Returns None if no host can be parsed.
+    """
+    value = value.strip()
+    # furl only parses the host when an authority is present, so ensure one
+    host = furl(value if "//" in value else "//" + value).host
+    if not host:
+        return None
+    return r"^(https?://)?" + re.escape(host) + r"(:\d+)?($|[/?#])"
+
+
 class AddOnViewSet(viewsets.ModelViewSet):
     serializer_class = AddOnSerializer
     queryset = AddOn.objects.none()
@@ -746,6 +762,14 @@ class Filter(django_filters.FilterSet):
             label="Site",
             help_text="Filter runs by the `site` value in the event's parameters.",
         )
+        domain = django_filters.CharFilter(
+            method="domain_filter",
+            label="Domain",
+            help_text=(
+                "Filter runs by the host of the event's `site` parameter, e.g. "
+                "`www.nifc.gov` or `https://www.nifc.gov`."
+            ),
+        )
         message = django_filters.CharFilter(
             field_name="message",
             lookup_expr="exact",
@@ -753,6 +777,13 @@ class Filter(django_filters.FilterSet):
             help_text="Filter runs by their progress message.",
         )
 
+        def domain_filter(self, queryset, name, value):
+            # pylint: disable=unused-argument
+            pattern = domain_site_regex(value)
+            if pattern is None:
+                return queryset.none()
+            return queryset.filter(event__parameters__site__iregex=pattern)
+
         class Meta:
             model = AddOnRun
             fields = {
@@ -988,6 +1019,21 @@ class Filter(django_filters.FilterSet):
             label="Site",
             help_text="Filter events by the `site` value in their parameters.",
         )
+        domain = django_filters.CharFilter(
+            method="domain_filter",
+            label="Domain",
+            help_text=(
+                "Filter events by the host of their `site` parameter, e.g. "
+                "`www.nifc.gov` or `https://www.nifc.gov`."
+            ),
+        )
+
+        def domain_filter(self, queryset, name, value):
+            # pylint: disable=unused-argument
+            pattern = domain_site_regex(value)
+            if pattern is None:
+                return queryset.none()
+            return queryset.filter(parameters__site__iregex=pattern)
 
         class Meta:
             model = AddOnEvent

From 3bc91baf4fd9883b43dbe37a1f46fbad66cc0946 Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Mon, 8 Jun 2026 12:38:26 -0400
Subject: [PATCH 2/6] Filter on domain using an index

---
 ...030_addonevent_addonevent_site_host_idx.py | 22 +++++++++++
 documentcloud/addons/models.py                | 25 +++++++++++-
 documentcloud/addons/views.py                 | 38 ++++++++++++-------
 3 files changed, 70 insertions(+), 15 deletions(-)
 create mode 100644 documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py

diff --git a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py
new file mode 100644
index 00000000..b18c2381
--- /dev/null
+++ b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py
@@ -0,0 +1,22 @@
+# Generated by Django 4.2.2 on 2026-06-08 16:21
+
+from django.db import migrations, models
+import documentcloud.addons.models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("addons", "0029_addonrun_data"),
+    ]
+
+    operations = [
+        migrations.AddIndex(
+            model_name="addonevent",
+            index=models.Index(
+                documentcloud.addons.models.SiteHost("parameters"),
+                condition=models.Q(("parameters__has_key", "site")),
+                name="addonevent_site_host_idx",
+            ),
+        ),
+    ]
diff --git a/documentcloud/addons/models.py b/documentcloud/addons/models.py
index 71884a8e..9e4e8cc2 100644
--- a/documentcloud/addons/models.py
+++ b/documentcloud/addons/models.py
@@ -2,7 +2,7 @@
 from django.conf import settings
 from django.core.cache import cache
 from django.db import models, transaction
-from django.db.models import F, Q
+from django.db.models import F, Func, Q
 from django.utils.translation import gettext_lazy as _
 
 # Standard Library
@@ -34,6 +34,24 @@
 logger = logging.getLogger(__name__)
 
 
+class SiteHost(Func):
+    # pylint: disable=abstract-method
+    """Extract the lowercased host from the `site` URL of a JSON parameters field.
+
+    Captures the authority (host) of the stored URL, stopping at a port, path,
+    query or fragment, with or without a leading scheme. Used both to build the
+    expression index on `AddOnEvent` and to filter by it, so the index and the
+    query share one definition and cannot drift apart. All component functions
+    are IMMUTABLE, so this is safe to index.
+    """
+
+    template = (
+        "LOWER(SUBSTRING(%(expressions)s ->> 'site' "
+        "FROM '^(?:https?://)?([^/:?#]+)'))"
+    )
+    output_field = models.TextField()
+
+
 class AddOn(models.Model):
     objects = AddOnQuerySet.as_manager()
 
@@ -583,6 +601,11 @@ class Meta:
                 name="addonevent_param_site_idx",
                 condition=Q(parameters__has_key="site"),
             ),
+            models.Index(
+                SiteHost("parameters"),
+                name="addonevent_site_host_idx",
+                condition=Q(parameters__has_key="site"),
+            ),
         ]
 
     def __str__(self):
diff --git a/documentcloud/addons/views.py b/documentcloud/addons/views.py
index 3e643b4f..25955816 100644
--- a/documentcloud/addons/views.py
+++ b/documentcloud/addons/views.py
@@ -30,7 +30,6 @@
 import hmac
 import json
 import logging
-import re
 from collections import defaultdict
 from datetime import timedelta
 from functools import lru_cache
@@ -52,6 +51,7 @@
     AddOnRun,
     GitHubAccount,
     GitHubInstallation,
+    SiteHost,
     VisualAddOn,
 )
 from documentcloud.addons.serializers import (
@@ -66,19 +66,17 @@
 logger = logging.getLogger(__name__)
 
 
-def domain_site_regex(value):
-    """Build a regex matching `site` URLs whose host equals the given domain.
+def domain_to_host(value):
+    """Normalize a domain filter value to a bare, lowercased host.
 
     Accepts either a bare host (`www.nifc.gov`) or a full URL
-    (`https://www.nifc.gov/path`) and matches stored site values regardless of
-    scheme, port, path or query string. Returns None if no host can be parsed.
+    (`https://www.nifc.gov/path`) and returns just the host, lowercased to match
+    the `SiteHost` expression index. Returns None if no host can be parsed.
     """
     value = value.strip()
     # furl only parses the host when an authority is present, so ensure one
     host = furl(value if "//" in value else "//" + value).host
-    if not host:
-        return None
-    return r"^(https?://)?" + re.escape(host) + r"(:\d+)?($|[/?#])"
+    return host.lower() if host else None
 
 
 class AddOnViewSet(viewsets.ModelViewSet):
@@ -779,10 +777,16 @@ class Filter(django_filters.FilterSet):
 
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
-            pattern = domain_site_regex(value)
-            if pattern is None:
+            host = domain_to_host(value)
+            if host is None:
                 return queryset.none()
-            return queryset.filter(event__parameters__site__iregex=pattern)
+            # filter on has_key + the SiteHost expression so the partial
+            # `addonevent_site_host_idx` index on AddOnEvent is usable
+            return (
+                queryset.filter(event__parameters__has_key="site")
+                .annotate(_site_host=SiteHost("event__parameters"))
+                .filter(_site_host=host)
+            )
 
         class Meta:
             model = AddOnRun
@@ -1030,10 +1034,16 @@ class Filter(django_filters.FilterSet):
 
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
-            pattern = domain_site_regex(value)
-            if pattern is None:
+            host = domain_to_host(value)
+            if host is None:
                 return queryset.none()
-            return queryset.filter(parameters__site__iregex=pattern)
+            # filter on has_key + the SiteHost expression so the partial
+            # `addonevent_site_host_idx` index is usable
+            return (
+                queryset.filter(parameters__has_key="site")
+                .annotate(_site_host=SiteHost("parameters"))
+                .filter(_site_host=host)
+            )
 
         class Meta:
             model = AddOnEvent

From 66bec189e6697ab974724ac62278b9947a5864cf Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Mon, 8 Jun 2026 15:21:29 -0400
Subject: [PATCH 3/6] Match on domain, not host

---
 ..._addonevent_addonevent_site_origin_idx.py} |  6 +--
 documentcloud/addons/models.py                | 23 ++++----
 documentcloud/addons/tests/test_views.py      | 30 +++++++----
 documentcloud/addons/views.py                 | 53 ++++++++++---------
 4 files changed, 59 insertions(+), 53 deletions(-)
 rename documentcloud/addons/migrations/{0030_addonevent_addonevent_site_host_idx.py => 0030_addonevent_addonevent_site_origin_idx.py} (70%)

diff --git a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
similarity index 70%
rename from documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py
rename to documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
index b18c2381..08af6c0f 100644
--- a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_host_idx.py
+++ b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
@@ -1,4 +1,4 @@
-# Generated by Django 4.2.2 on 2026-06-08 16:21
+# Generated by Django 4.2.2 on 2026-06-08 19:10
 
 from django.db import migrations, models
 import documentcloud.addons.models
@@ -14,9 +14,9 @@ class Migration(migrations.Migration):
         migrations.AddIndex(
             model_name="addonevent",
             index=models.Index(
-                documentcloud.addons.models.SiteHost("parameters"),
+                documentcloud.addons.models.SiteOrigin("parameters"),
                 condition=models.Q(("parameters__has_key", "site")),
-                name="addonevent_site_host_idx",
+                name="addonevent_site_origin_idx",
             ),
         ),
     ]
diff --git a/documentcloud/addons/models.py b/documentcloud/addons/models.py
index 9e4e8cc2..1a2b5187 100644
--- a/documentcloud/addons/models.py
+++ b/documentcloud/addons/models.py
@@ -34,21 +34,18 @@
 logger = logging.getLogger(__name__)
 
 
-class SiteHost(Func):
+class SiteOrigin(Func):
     # pylint: disable=abstract-method
-    """Extract the lowercased host from the `site` URL of a JSON parameters field.
+    """Extract the lowercased origin (scheme + host) from a JSON `site` URL.
 
-    Captures the authority (host) of the stored URL, stopping at a port, path,
-    query or fragment, with or without a leading scheme. Used both to build the
-    expression index on `AddOnEvent` and to filter by it, so the index and the
-    query share one definition and cannot drift apart. All component functions
-    are IMMUTABLE, so this is safe to index.
+    Captures the scheme and authority of the stored URL, stopping at the path,
+    query or fragment, so `https://www.nifc.gov/foo` yields `https://www.nifc.gov`.
+    Used both to build the expression index on `AddOnEvent` and to filter by it,
+    so the index and the query share one definition and cannot drift apart. All
+    component functions are IMMUTABLE, so this is safe to index.
     """
 
-    template = (
-        "LOWER(SUBSTRING(%(expressions)s ->> 'site' "
-        "FROM '^(?:https?://)?([^/:?#]+)'))"
-    )
+    template = "LOWER(SUBSTRING(%(expressions)s ->> 'site' FROM '^(https?://[^/?#]+)'))"
     output_field = models.TextField()
 
 
@@ -602,8 +599,8 @@ class Meta:
                 condition=Q(parameters__has_key="site"),
             ),
             models.Index(
-                SiteHost("parameters"),
-                name="addonevent_site_host_idx",
+                SiteOrigin("parameters"),
+                name="addonevent_site_origin_idx",
                 condition=Q(parameters__has_key="site"),
             ),
         ]
diff --git a/documentcloud/addons/tests/test_views.py b/documentcloud/addons/tests/test_views.py
index da6cf4a9..54d66894 100644
--- a/documentcloud/addons/tests/test_views.py
+++ b/documentcloud/addons/tests/test_views.py
@@ -310,7 +310,7 @@ def test_filter_site_absent_is_noop(self, client):
         assert len(response.json()["results"]) == 3
 
     def test_filter_domain(self, client):
-        """Filter runs by the host of the event's parameters.site"""
+        """Filter runs by the origin of the event's parameters.site"""
         user = UserFactory()
         matching_event = AddOnEventFactory(
             user=user,
@@ -328,22 +328,26 @@ def test_filter_domain(self, client):
         AddOnRunFactory(user=user, event=no_site_event)
         AddOnRunFactory(user=user, event=None)
         client.force_authenticate(user=user)
-        # bare host and full origin both match, regardless of path
-        for domain in ("www.nifc.gov", "https://www.nifc.gov"):
+        # the origin matches regardless of path
+        response = client.get("/api/addon_runs/", {"domain": "https://www.nifc.gov"})
+        assert response.status_code == status.HTTP_200_OK
+        uuids = [r["uuid"] for r in response.json()["results"]]
+        assert uuids == [str(matching_run.uuid)]
+        # a bare host and a mismatched scheme do not match
+        for domain in ("www.nifc.gov", "http://www.nifc.gov"):
             response = client.get("/api/addon_runs/", {"domain": domain})
             assert response.status_code == status.HTTP_200_OK
-            uuids = [r["uuid"] for r in response.json()["results"]]
-            assert uuids == [str(matching_run.uuid)], domain
+            assert response.json()["results"] == [], domain
 
     def test_filter_domain_no_partial_host_match(self, client):
-        """The domain filter matches whole hosts, not substrings"""
+        """The domain filter matches whole origins, not substrings"""
         user = UserFactory()
         event = AddOnEventFactory(
             user=user, parameters={"site": "https://www.nifc.gov.evil.com/path"}
         )
         AddOnRunFactory(user=user, event=event)
         client.force_authenticate(user=user)
-        response = client.get("/api/addon_runs/", {"domain": "www.nifc.gov"})
+        response = client.get("/api/addon_runs/", {"domain": "https://www.nifc.gov"})
         assert response.status_code == status.HTTP_200_OK
         assert response.json()["results"] == []
 
@@ -390,7 +394,7 @@ def test_filter_site_no_match(self, client):
         assert response.json()["results"] == []
 
     def test_filter_domain(self, client):
-        """Filter events by the host of their parameters.site"""
+        """Filter events by the origin of their parameters.site"""
         user = UserFactory()
         matching = AddOnEventFactory(
             user=user,
@@ -402,11 +406,15 @@ def test_filter_domain(self, client):
         AddOnEventFactory(user=user, parameters={"site": "https://www.other.com/path"})
         AddOnEventFactory(user=user, parameters={"selector": "*"})
         client.force_authenticate(user=user)
-        for domain in ("www.nifc.gov", "https://www.nifc.gov"):
+        response = client.get("/api/addon_events/", {"domain": "https://www.nifc.gov"})
+        assert response.status_code == status.HTTP_200_OK
+        ids = [r["id"] for r in response.json()["results"]]
+        assert ids == [matching.pk]
+        # a bare host and a mismatched scheme do not match
+        for domain in ("www.nifc.gov", "http://www.nifc.gov"):
             response = client.get("/api/addon_events/", {"domain": domain})
             assert response.status_code == status.HTTP_200_OK
-            ids = [r["id"] for r in response.json()["results"]]
-            assert ids == [matching.pk], domain
+            assert response.json()["results"] == [], domain
 
     def test_filter_message(self, client):
         """Filter runs by message"""
diff --git a/documentcloud/addons/views.py b/documentcloud/addons/views.py
index 25955816..cb5134f2 100644
--- a/documentcloud/addons/views.py
+++ b/documentcloud/addons/views.py
@@ -51,7 +51,7 @@
     AddOnRun,
     GitHubAccount,
     GitHubInstallation,
-    SiteHost,
+    SiteOrigin,
     VisualAddOn,
 )
 from documentcloud.addons.serializers import (
@@ -66,17 +66,18 @@
 logger = logging.getLogger(__name__)
 
 
-def domain_to_host(value):
-    """Normalize a domain filter value to a bare, lowercased host.
+def domain_to_origin(value):
+    """Normalize a domain filter value to a lowercased origin (scheme + host).
 
-    Accepts either a bare host (`www.nifc.gov`) or a full URL
-    (`https://www.nifc.gov/path`) and returns just the host, lowercased to match
-    the `SiteHost` expression index. Returns None if no host can be parsed.
+    Expects an origin such as `https://www.nifc.gov`; a value without a scheme
+    will not match. A full URL is accepted and its path is discarded. Returns
+    None if no scheme + host can be parsed, lowercased to match the `SiteOrigin`
+    expression index.
     """
-    value = value.strip()
-    # furl only parses the host when an authority is present, so ensure one
-    host = furl(value if "//" in value else "//" + value).host
-    return host.lower() if host else None
+    parsed = furl(value.strip())
+    if not parsed.scheme or not parsed.host:
+        return None
+    return f"{parsed.scheme}://{parsed.netloc}".lower()
 
 
 class AddOnViewSet(viewsets.ModelViewSet):
@@ -764,8 +765,8 @@ class Filter(django_filters.FilterSet):
             method="domain_filter",
             label="Domain",
             help_text=(
-                "Filter runs by the host of the event's `site` parameter, e.g. "
-                "`www.nifc.gov` or `https://www.nifc.gov`."
+                "Filter runs by the origin of the event's `site` parameter, e.g. "
+                "`https://www.nifc.gov`."
             ),
         )
         message = django_filters.CharFilter(
@@ -777,15 +778,15 @@ class Filter(django_filters.FilterSet):
 
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
-            host = domain_to_host(value)
-            if host is None:
+            origin = domain_to_origin(value)
+            if origin is None:
                 return queryset.none()
-            # filter on has_key + the SiteHost expression so the partial
-            # `addonevent_site_host_idx` index on AddOnEvent is usable
+            # filter on has_key + the SiteOrigin expression so the partial
+            # `addonevent_site_origin_idx` index on AddOnEvent is usable
             return (
                 queryset.filter(event__parameters__has_key="site")
-                .annotate(_site_host=SiteHost("event__parameters"))
-                .filter(_site_host=host)
+                .annotate(_site_origin=SiteOrigin("event__parameters"))
+                .filter(_site_origin=origin)
             )
 
         class Meta:
@@ -1027,22 +1028,22 @@ class Filter(django_filters.FilterSet):
             method="domain_filter",
             label="Domain",
             help_text=(
-                "Filter events by the host of their `site` parameter, e.g. "
-                "`www.nifc.gov` or `https://www.nifc.gov`."
+                "Filter events by the origin of their `site` parameter, e.g. "
+                "`https://www.nifc.gov`."
             ),
         )
 
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
-            host = domain_to_host(value)
-            if host is None:
+            origin = domain_to_origin(value)
+            if origin is None:
                 return queryset.none()
-            # filter on has_key + the SiteHost expression so the partial
-            # `addonevent_site_host_idx` index is usable
+            # filter on has_key + the SiteOrigin expression so the partial
+            # `addonevent_site_origin_idx` index is usable
             return (
                 queryset.filter(parameters__has_key="site")
-                .annotate(_site_host=SiteHost("parameters"))
-                .filter(_site_host=host)
+                .annotate(_site_origin=SiteOrigin("parameters"))
+                .filter(_site_origin=origin)
             )
 
         class Meta:

From 7153fa0b83a24a9128f69fde392f429934fe1a9b Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Mon, 8 Jun 2026 15:42:38 -0400
Subject: [PATCH 4/6] Fix indexing on site lookup

---
 ...vent_addonevent_param_site_idx_and_more.py | 29 +++++++++++++++++++
 documentcloud/addons/models.py                |  8 +++--
 documentcloud/addons/views.py                 | 24 ++++++++++++---
 3 files changed, 55 insertions(+), 6 deletions(-)
 create mode 100644 documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py

diff --git a/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py b/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
new file mode 100644
index 00000000..aab19125
--- /dev/null
+++ b/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
@@ -0,0 +1,29 @@
+# Generated by Django 4.2.2 on 2026-06-08 19:30
+
+from django.db import migrations, models
+import django.db.models.fields.json
+import django.db.models.functions.text
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("addons", "0030_addonevent_addonevent_site_origin_idx"),
+    ]
+
+    operations = [
+        migrations.RemoveIndex(
+            model_name="addonevent",
+            name="addonevent_param_site_idx",
+        ),
+        migrations.AddIndex(
+            model_name="addonevent",
+            index=models.Index(
+                django.db.models.functions.text.Upper(
+                    django.db.models.fields.json.KeyTextTransform("site", "parameters")
+                ),
+                condition=models.Q(("parameters__has_key", "site")),
+                name="addonevent_param_site_idx",
+            ),
+        ),
+    ]
diff --git a/documentcloud/addons/models.py b/documentcloud/addons/models.py
index 1a2b5187..0af4ca61 100644
--- a/documentcloud/addons/models.py
+++ b/documentcloud/addons/models.py
@@ -2,7 +2,9 @@
 from django.conf import settings
 from django.core.cache import cache
 from django.db import models, transaction
-from django.db.models import F, Func, Q
+from django.db.models import Func, Q
+from django.db.models.fields.json import KeyTextTransform
+from django.db.models.functions import Upper
 from django.utils.translation import gettext_lazy as _
 
 # Standard Library
@@ -594,7 +596,9 @@ class AddOnEvent(models.Model):
     class Meta:
         indexes = [
             models.Index(
-                F("parameters__site"),
+                # matches the `UPPER(parameters ->> 'site')` the `site` filter's
+                # iexact lookup emits, so a case-insensitive match can use it
+                Upper(KeyTextTransform("site", "parameters")),
                 name="addonevent_param_site_idx",
                 condition=Q(parameters__has_key="site"),
             ),
diff --git a/documentcloud/addons/views.py b/documentcloud/addons/views.py
index cb5134f2..375dd03b 100644
--- a/documentcloud/addons/views.py
+++ b/documentcloud/addons/views.py
@@ -756,8 +756,7 @@ class Filter(django_filters.FilterSet):
         )
         dismissed = django_filters.BooleanFilter(help_text="Was this run dismissed?")
         site = django_filters.CharFilter(
-            field_name="event__parameters__site",
-            lookup_expr="iexact",
+            method="site_filter",
             label="Site",
             help_text="Filter runs by the `site` value in the event's parameters.",
         )
@@ -776,6 +775,15 @@ class Filter(django_filters.FilterSet):
             help_text="Filter runs by their progress message.",
         )
 
+        def site_filter(self, queryset, name, value):
+            # pylint: disable=unused-argument
+            # the has_key clause lets the partial `addonevent_param_site_idx`
+            # index on AddOnEvent serve the case-insensitive match
+            return queryset.filter(
+                event__parameters__has_key="site",
+                event__parameters__site__iexact=value,
+            )
+
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
             origin = domain_to_origin(value)
@@ -1019,8 +1027,7 @@ class Filter(django_filters.FilterSet):
             help_text="Filter events by a specific add-on ID.",
         )
         site = django_filters.CharFilter(
-            field_name="parameters__site",
-            lookup_expr="iexact",
+            method="site_filter",
             label="Site",
             help_text="Filter events by the `site` value in their parameters.",
         )
@@ -1033,6 +1040,15 @@ class Filter(django_filters.FilterSet):
             ),
         )
 
+        def site_filter(self, queryset, name, value):
+            # pylint: disable=unused-argument
+            # the has_key clause lets the partial `addonevent_param_site_idx`
+            # index serve the case-insensitive match
+            return queryset.filter(
+                parameters__has_key="site",
+                parameters__site__iexact=value,
+            )
+
         def domain_filter(self, queryset, name, value):
             # pylint: disable=unused-argument
             origin = domain_to_origin(value)

From 88a1c0c9d82ba067cba17c63d3c1db96caa7ea84 Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Mon, 8 Jun 2026 15:48:58 -0400
Subject: [PATCH 5/6] Add indexes concurrently

---
 .../0030_addonevent_addonevent_site_origin_idx.py      |  5 ++++-
 ...ve_addonevent_addonevent_param_site_idx_and_more.py | 10 ++++++++--
 2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
index 08af6c0f..77af7ed4 100644
--- a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
+++ b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
@@ -1,17 +1,20 @@
 # Generated by Django 4.2.2 on 2026-06-08 19:10
 
+from django.contrib.postgres.operations import AddIndexConcurrently
 from django.db import migrations, models
 import documentcloud.addons.models
 
 
 class Migration(migrations.Migration):
 
+    atomic = False
+
     dependencies = [
         ("addons", "0029_addonrun_data"),
     ]
 
     operations = [
-        migrations.AddIndex(
+        AddIndexConcurrently(
             model_name="addonevent",
             index=models.Index(
                 documentcloud.addons.models.SiteOrigin("parameters"),
diff --git a/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py b/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
index aab19125..82139ac9 100644
--- a/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
+++ b/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
@@ -1,5 +1,9 @@
 # Generated by Django 4.2.2 on 2026-06-08 19:30
 
+from django.contrib.postgres.operations import (
+    AddIndexConcurrently,
+    RemoveIndexConcurrently,
+)
 from django.db import migrations, models
 import django.db.models.fields.json
 import django.db.models.functions.text
@@ -7,16 +11,18 @@
 
 class Migration(migrations.Migration):
 
+    atomic = False
+
     dependencies = [
         ("addons", "0030_addonevent_addonevent_site_origin_idx"),
     ]
 
     operations = [
-        migrations.RemoveIndex(
+        RemoveIndexConcurrently(
             model_name="addonevent",
             name="addonevent_param_site_idx",
         ),
-        migrations.AddIndex(
+        AddIndexConcurrently(
             model_name="addonevent",
             index=models.Index(
                 django.db.models.functions.text.Upper(

From 81c76b3cd13ab9c76ea08c021de50acf9c5b1ef8 Mon Sep 17 00:00:00 2001
From: Chris Amico <eyeseast@gmail.com>
Date: Mon, 8 Jun 2026 16:02:27 -0400
Subject: [PATCH 6/6] one migration not two

---
 ...0_addonevent_addonevent_site_origin_idx.py | 25 -------------------
 ...ore.py => 0030_addonevent_site_indexes.py} | 16 ++++++++++--
 2 files changed, 14 insertions(+), 27 deletions(-)
 delete mode 100644 documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
 rename documentcloud/addons/migrations/{0031_remove_addonevent_addonevent_param_site_idx_and_more.py => 0030_addonevent_site_indexes.py} (58%)

diff --git a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py b/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
deleted file mode 100644
index 77af7ed4..00000000
--- a/documentcloud/addons/migrations/0030_addonevent_addonevent_site_origin_idx.py
+++ /dev/null
@@ -1,25 +0,0 @@
-# Generated by Django 4.2.2 on 2026-06-08 19:10
-
-from django.contrib.postgres.operations import AddIndexConcurrently
-from django.db import migrations, models
-import documentcloud.addons.models
-
-
-class Migration(migrations.Migration):
-
-    atomic = False
-
-    dependencies = [
-        ("addons", "0029_addonrun_data"),
-    ]
-
-    operations = [
-        AddIndexConcurrently(
-            model_name="addonevent",
-            index=models.Index(
-                documentcloud.addons.models.SiteOrigin("parameters"),
-                condition=models.Q(("parameters__has_key", "site")),
-                name="addonevent_site_origin_idx",
-            ),
-        ),
-    ]
diff --git a/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py b/documentcloud/addons/migrations/0030_addonevent_site_indexes.py
similarity index 58%
rename from documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
rename to documentcloud/addons/migrations/0030_addonevent_site_indexes.py
index 82139ac9..3cd8725a 100644
--- a/documentcloud/addons/migrations/0031_remove_addonevent_addonevent_param_site_idx_and_more.py
+++ b/documentcloud/addons/migrations/0030_addonevent_site_indexes.py
@@ -1,4 +1,4 @@
-# Generated by Django 4.2.2 on 2026-06-08 19:30
+# Generated by Django 4.2.2 on 2026-06-08
 
 from django.contrib.postgres.operations import (
     AddIndexConcurrently,
@@ -7,6 +7,7 @@
 from django.db import migrations, models
 import django.db.models.fields.json
 import django.db.models.functions.text
+import documentcloud.addons.models
 
 
 class Migration(migrations.Migration):
@@ -14,10 +15,21 @@ class Migration(migrations.Migration):
     atomic = False
 
     dependencies = [
-        ("addons", "0030_addonevent_addonevent_site_origin_idx"),
+        ("addons", "0029_addonrun_data"),
     ]
 
     operations = [
+        # index the `site` origin (scheme + host) for the `domain` filter
+        AddIndexConcurrently(
+            model_name="addonevent",
+            index=models.Index(
+                documentcloud.addons.models.SiteOrigin("parameters"),
+                condition=models.Q(("parameters__has_key", "site")),
+                name="addonevent_site_origin_idx",
+            ),
+        ),
+        # rebuild the `site` index as UPPER(...) so the case-insensitive
+        # (iexact) `site` filter can use it
         RemoveIndexConcurrently(
             model_name="addonevent",
             name="addonevent_param_site_idx",