Merge pull request #653 from TG1999/models/migrations

Reference: https://github.com/vulnerablecode/issue/650

Work around PosgreSQL index issue.

Author: TG1999

requirements-dev.txt

@@ -1,4 +1,4 @@
-black==22.1.0
+black==22.3.0
 freezegun==1.1.0
 ipython==8.0.1
 isort==5.10.1

vulnerabilities/migrations/0004_advisory_unique_content_id.py

@@ -0,0 +1,19 @@
+# Generated by Django 4.0.2 on 2022-03-29 09:38
+
+from django.db import migrations
+from django.db import models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("vulnerabilities", "0003_alter_advisory_created_by"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="advisory",
+            name="unique_content_id",
+            field=models.CharField(blank=True, max_length=32, null=True),
+        ),
+    ]

vulnerabilities/migrations/0005_auto_20220329_0938.py

@@ -0,0 +1,26 @@
+# Generated by Django 4.0.2 on 2022-03-29 09:38
+
+import hashlib
+import json
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+    def md5hash(apps, schema_editor):
+        Advisory = apps.get_model("vulnerabilities", "Advisory")
+        for advisory in Advisory.objects.all():
+            checksum = hashlib.md5()
+            for field in (advisory.summary, advisory.affected_packages, advisory.references):
+                value = json.dumps(field, separators=(",", ":")).encode("utf-8")
+                checksum.update(value)
+            advisory.unique_content_id = checksum.hexdigest()
+            advisory.save()
+
+    dependencies = [
+        ("vulnerabilities", "0004_advisory_unique_content_id"),
+    ]
+
+    operations = [
+        migrations.RunPython(md5hash),
+    ]

vulnerabilities/migrations/0006_alter_advisory_unique_together.py

@@ -0,0 +1,17 @@
+# Generated by Django 4.0.2 on 2022-03-29 09:41
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("vulnerabilities", "0005_auto_20220329_0938"),
+    ]
+
+    operations = [
+        migrations.AlterUniqueTogether(
+            name="advisory",
+            unique_together={("aliases", "unique_content_id", "date_published")},
+        ),
+    ]

vulnerabilities/models.py

@@ -21,6 +21,7 @@
 #  Visit https://github.com/nexB/vulnerablecode/ for support and download.
 
 import dataclasses
+import hashlib
 import importlib
 import json
 import logging
@@ -326,6 +327,15 @@ class Advisory(models.Model):
     into structured data
     """
 
+    def save(self, *args, **kwargs):
+        checksum = hashlib.md5()
+        for field in (self.summary, self.affected_packages, self.references):
+            value = json.dumps(field, separators=(",", ":")).encode("utf-8")
+            checksum.update(value)
+        self.unique_content_id = checksum.hexdigest()
+        super().save(*args, **kwargs)
+
+    unique_content_id = models.CharField(max_length=32, blank=True, null=True)
     aliases = models.JSONField(blank=True, default=list, help_text="A list of alias strings")
     summary = models.TextField(blank=True, null=True)
     # we use a JSON field here to avoid creating a complete relational model for data that
@@ -356,9 +366,7 @@ class Advisory(models.Model):
     class Meta:
         unique_together = (
             "aliases",
-            "summary",
-            "affected_packages",
-            "references",
+            "unique_content_id",
             "date_published",
         )