Builds: add data_hash field and update get_or_create method

readthedocs/builds/managers.py

@@ -1,5 +1,8 @@
 """Build and Version class model Managers."""
 
+import hashlib
+import json
+
 import structlog
 from django.core.exceptions import ObjectDoesNotExist
 from django.db import models
@@ -141,3 +144,19 @@ def register_match(self, rule, version, max_registers=15):
         for match in self.filter(rule__project=rule.project)[max_registers:]:
             match.delete()
         return created
+
+
+class BuildConfigManager(models.Manager):
+    """Manager for BuildConfig model."""
+
+    def get_or_create(self, **kwargs):
+        data = kwargs.get("data")
+        if data:
+            dump = json.dumps(data, sort_keys=False)
+            data_hash = hashlib.sha256(dump.encode("utf-8")).hexdigest()
+
+            return super().get_or_create(
+                data_hash=data_hash,
+                defaults={"data": data},
+            )
+        return super().get_or_create(**kwargs)

readthedocs/builds/migrations/0067_buildconfig_data_hash.py

@@ -0,0 +1,29 @@
+# Generated by Django 5.2.9 on 2026-01-21 15:23
+
+from django.db import migrations
+from django.db import models
+from django_safemigrate import Safe
+
+
+class Migration(migrations.Migration):
+    safe = Safe.before_deploy()
+
+    dependencies = [
+        ("builds", "0066_add_buildconfig_model"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="buildconfig",
+            name="data_hash",
+            field=models.CharField(editable=False, max_length=64, null=True, unique=True),
+        ),
+        migrations.AlterField(
+            model_name="buildconfig",
+            name="data",
+            field=models.JSONField(
+                help_text="The rendered YAML configuration used in the build",
+                verbose_name="Configuration data",
+            ),
+        ),
+    ]

readthedocs/builds/models.py

@@ -33,6 +33,7 @@
 from readthedocs.builds.constants import STABLE
 from readthedocs.builds.constants import VERSION_TYPES
 from readthedocs.builds.managers import AutomationRuleMatchManager
+from readthedocs.builds.managers import BuildConfigManager
 from readthedocs.builds.managers import ExternalBuildManager
 from readthedocs.builds.managers import ExternalVersionManager
 from readthedocs.builds.managers import InternalBuildManager
@@ -617,10 +618,22 @@ class BuildConfig(TimeStampedModel):
 
     data = models.JSONField(
         _("Configuration data"),
-        unique=True,
         help_text=_("The rendered YAML configuration used in the build"),
     )
 
+    # Used to quickly identify identical configurations.
+    # Using `unique=True` on a JSONField can be problematic due to limitations on its size,
+    # so we use a hash instead.
+    data_hash = models.CharField(
+        unique=True,
+        max_length=64,
+        null=True,
+        editable=False,
+    )
+
+    # Manager to handle creation based on data hash uniqueness.
+    objects = BuildConfigManager()
+
     class Meta:
         verbose_name = _("Build configuration")
         verbose_name_plural = _("Build configurations")