From 141dd29883bc0756545e0310a76bc86b1cb1cc16 Mon Sep 17 00:00:00 2001
From: David Fischer <david@readthedocs.org>
Date: Mon, 2 Dec 2024 15:54:20 -0800
Subject: [PATCH] Collect the domain on offer creation

---
 adserver/migrations/0100_add_offer_domain.py | 28 ++++++++++++++++++++
 adserver/models.py                           |  6 +++++
 adserver/tests/test_models.py                |  1 +
 3 files changed, 35 insertions(+)
 create mode 100644 adserver/migrations/0100_add_offer_domain.py

diff --git a/adserver/migrations/0100_add_offer_domain.py b/adserver/migrations/0100_add_offer_domain.py
new file mode 100644
index 00000000..0a8e675d
--- /dev/null
+++ b/adserver/migrations/0100_add_offer_domain.py
@@ -0,0 +1,28 @@
+# Generated by Django 5.0.9 on 2024-12-02 23:42
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('adserver', '0099_link_advertiser_guide'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='click',
+            name='domain',
+            field=models.CharField(blank=True, max_length=10000, null=True, verbose_name='Domain'),
+        ),
+        migrations.AddField(
+            model_name='offer',
+            name='domain',
+            field=models.CharField(blank=True, max_length=10000, null=True, verbose_name='Domain'),
+        ),
+        migrations.AddField(
+            model_name='view',
+            name='domain',
+            field=models.CharField(blank=True, max_length=10000, null=True, verbose_name='Domain'),
+        ),
+    ]
diff --git a/adserver/models.py b/adserver/models.py
index fe69f013..6009e53a 100644
--- a/adserver/models.py
+++ b/adserver/models.py
@@ -1775,6 +1775,7 @@ def _record_base(
         parsed_ua = parse(user_agent)
         country = get_client_country(request)
         url = url or request.headers.get("referer")
+        domain = get_domain_from_url(url)
 
         if (
             model != Click
@@ -1799,6 +1800,7 @@ def _record_base(
             client_id=client_id,
             country=country,
             url=url,
+            domain=domain,
             paid_eligible=paid_eligible,
             rotations=rotations,
             # Derived user agent data
@@ -2589,6 +2591,10 @@ class AdBase(TimeStampedModel, IndestructibleModel):
     country = CountryField(null=True)
     url = models.CharField(_("Page URL"), max_length=10000, blank=True, null=True)
 
+    # Domain of the URL or None if the URL is not available
+    # Should not include http or any characters after the TLD
+    domain = models.CharField(_("Domain"), max_length=10000, blank=True, null=True)
+
     # Fields derived from the user agent - these should not be user identifiable
     browser_family = models.CharField(
         _("Browser Family"), max_length=1000, blank=True, null=True, default=None
diff --git a/adserver/tests/test_models.py b/adserver/tests/test_models.py
index 8e9d036b..11bf59ec 100644
--- a/adserver/tests/test_models.py
+++ b/adserver/tests/test_models.py
@@ -633,6 +633,7 @@ def test_offer_ad(self):
         self.assertEqual(offer.advertisement, self.ad1)
         self.assertEqual(offer.div_id, div_id)
         self.assertEqual(offer.url, url)
+        self.assertEqual(offer.domain, "example.com")
         self.assertEqual(offer.ip, "1.1.0.0")  # anonymized
         self.assertEqual(offer.os_family, "Linux")
         self.assertEqual(offer.browser_family, "Chrome")