feat: Use openedx_catalog app, backfill it with all known courses

Author: bradenmacdonald

cms/envs/common.py

@@ -898,6 +898,9 @@ def make_lms_template_path(settings):
 
     'openedx_events',
 
+    # Core models to represent courses
+    "openedx_catalog",
+
     # Core apps that power libraries
     "openedx_content",
     *openedx_content_backcompat_apps_to_install(),

lms/envs/common.py

@@ -2020,6 +2020,9 @@
 
     'openedx_events',
 
+    # Core models to represent courses
+    "openedx_catalog",
+
     # Core apps that power libraries
     "openedx_content",
     *openedx_content_backcompat_apps_to_install(),

openedx/core/djangoapps/content/course_overviews/migrations/0030_backfill_new_catalog_courseruns.py

@@ -0,0 +1,126 @@
+"""
+Data migration to populate the new CourseRun and CatalogCourse models.
+"""
+
+# Generated by Django 5.2.11 on 2026-02-13 21:47
+import logging
+
+from django.conf import settings
+from django.db import migrations
+from organizations.api import ensure_organization, exceptions as org_exceptions
+
+log = logging.getLogger(__name__)
+
+
+def backfill_openedx_catalog(apps, schema_editor):
+    """
+    Populate the new CourseRun and CatalogCourse models.
+    """
+    # CourseOverview is a cache model derived from modulestore; modulestore is the source of truth for courses, so we'll
+    # use it to get the list of "all courses on the system" to populate the new CourseRun and CatalogCourse models.
+    CourseIndex = apps.get_model("split_modulestore_django", "SplitModulestoreCourseIndex")
+    CourseOverview = apps.get_model("course_overviews", "CourseOverview")
+    CatalogCourse = apps.get_model("openedx_catalog", "CatalogCourse")
+    CourseRun = apps.get_model("openedx_catalog", "CourseRun")
+
+    created_catalog_course_ids: set[int] = set()
+    all_course_runs = CourseIndex.objects.filter(base_store="mongodb", library_version="").order_by("course_id")
+    for course_run in all_course_runs:
+        org_code: str = course_run.course_id.org
+        course_code: str = course_run.course_id.course
+        run_code: str = course_run.course_id.run
+
+        # Ensure that the Organization exists.
+        try:
+            org_data = ensure_organization(org_code)
+        except org_exceptions.InvalidOrganizationException as exc:
+            # Note: IFF the org exists among the modulestore courses but not in the Organizations database table,
+            # and if auto-create is disabled (it's enabled by default), this will raise InvalidOrganizationException. It
+            # would be up to the operator to decide how they want to resolve that.
+            raise ValueError(
+                f'The organization short code "{org_code}" exists in modulestore ({course_run.course_id}) but '
+                "not the Organizations table, and auto-creating organizations is disabled. You can resolve this by "
+                "creating the Organization manually (e.g. from the Django admin) or turning on auto-creation. "
+                "You can set active=False to prevent this Organization from being used other than for historical data. "
+            )
+        if org_data["short_name"] != org_code:
+            # On most installations, the 'short_code' database column is case insensitive (unfortunately)
+            log.warning(
+                'The course with ID "%s" does not match its Organization.short_code "%s"',
+                course_run.course_id,
+                org_data["short_name"],
+            )
+
+        # Fetch the CourseOverview if it exists
+        try:
+            course_overview = CourseOverview.objects.get(id=course_run.course_id)
+        except CourseOverview.DoesNotExist:
+            course_overview = None  # Course exists in modulestore but details aren't cached into CourseOverview yet
+        display_name: str = (course_overview.display_name if course_overview else None) or course_code
+
+        # Determine the course language.
+        language = settings.LANGUAGE_CODE
+        if course_overview and course_overview.language:
+            language = course_overview.language.lower()
+            if len(language) > 2 and language[2] == "_":
+                language[2] = "-"  # Ensure we use hyphens for consistency (`en-us` not `en_us`)
+            if len(language) > 2 and language[2] not in ("-", "@"):
+                # This seems like an invalid value; revert to the default:
+                log.warning(
+                    'The course with ID "%s" has invalid language "%s" - using default language "%s" instead.',
+                    course_run.course_id,
+                    language,
+                    settings.LANGUAGE_CODE,
+                )
+                language = settings.LANGUAGE_CODE
+
+        # Ensure that the CatalogCourse exists.
+        cc, created = CatalogCourse.objects.get_or_create(
+            org_id=org_data["id"],
+            course_code=course_code,
+            defaults={
+                "display_name": display_name,
+                "language": language,
+            },
+        )
+        if created:
+            created_catalog_course_ids.add(cc.pk)
+        elif cc.pk in created_catalog_course_ids:
+            # This CatalogCourse was previously created during this same migration
+            # Check if all the runs have the same display_name:
+            if (
+                course_overview
+                and course_overview.display_name
+                and course_overview.display_name != cc.display_name
+                and cc.display_name != course_code
+            ):
+                # The runs have different names, so just use the course code as the common catalog course name.
+                cc.display_name = course_code
+                cc.save(update_fields=["display_name"])
+
+        if cc.course_code != course_code:
+            raise ValueError(
+                f"The course {course_run.course_id} exists in modulestore with a different capitalization of its "
+                f'course code compared to other instances of the same run ("{course_code}" vs "{cc.course_code}"). '
+                'This really should not happen. To fix it, delete the inconsistent course runs (!). '
+            )
+
+        # Create the CourseRun
+        CourseRun.objects.get_or_create(
+            catalog_course=cc,
+            run=run_code,
+            course_id=course_run.course_id,
+            defaults={"display_name": display_name},
+        )
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("openedx_catalog", "0001_initial"),
+        ("course_overviews", "0029_alter_historicalcourseoverview_options"),
+        ("split_modulestore_django", "0003_alter_historicalsplitmodulestorecourseindex_options"),
+    ]
+
+    operations = [
+        migrations.RunPython(backfill_openedx_catalog, reverse_code=migrations.RunPython.noop),
+    ]