ENG-11734 | feat(osf): add DownloadEvent model and migration with dashb…

osf/migrations/0045_downloadevent.py

@@ -0,0 +1,74 @@
+import django.db.models.deletion
+from django.db import migrations, models
+
+
+DASHBOARD_GROUP_NAME = 'download_telemetry'
+
+DASHBOARD_USERS = [
+    'sheredko.andriy@gmail.com',
+    'bodintsov@exoft.net',
+    'isokhan@exoft.net',
+    'ykopka@exoft.net',
+    'bgeiger@cos.io',
+    'osmand@cos.io',
+    'ramya@cos.io',
+]
+
+
+def create_dashboard_group(apps, schema_editor):
+    """Create the allow-list group the dashboard loads against and seed it.
+
+    The group carries no permissions of its own — membership is the only gate.
+    """
+    Group = apps.get_model('auth', 'Group')
+    OSFUser = apps.get_model('osf', 'OSFUser')
+    group, _ = Group.objects.get_or_create(name=DASHBOARD_GROUP_NAME)
+    for username in DASHBOARD_USERS:
+        user = OSFUser.objects.filter(username=username).first()
+        if user:
+            group.user_set.add(user)
+
+
+def remove_dashboard_group(apps, schema_editor):
+    Group = apps.get_model('auth', 'Group')
+    Group.objects.filter(name=DASHBOARD_GROUP_NAME).delete()
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('osf', '0044_notification_scheduled'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='DownloadEvent',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('created', models.DateTimeField(auto_now_add=True, db_index=True)),
+                ('resource_guid', models.CharField(blank=True, db_index=True, default='', max_length=255)),
+                ('path', models.TextField(blank=True, default='')),
+                ('download_type', models.CharField(choices=[('file', 'Single file'), ('folder_zip', 'Folder zip'), ('project', 'Whole-project zip')], max_length=16)),
+                ('zip_completed', models.BooleanField(blank=True, null=True)),
+                ('size_bytes', models.BigIntegerField(blank=True, null=True)),
+                ('storage_region', models.CharField(blank=True, default='', max_length=64)),
+                ('user_region', models.CharField(blank=True, default='', max_length=64)),
+                ('ip', models.GenericIPAddressField(blank=True, null=True)),
+                ('source_area', models.CharField(blank=True, default='', max_length=128)),
+                ('user', models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.SET_NULL, related_name='download_events', to='osf.osfuser')),
+            ],
+        ),
+        migrations.AddIndex(
+            model_name='downloadevent',
+            index=models.Index(fields=['created', 'download_type'], name='download_event_crt_type'),
+        ),
+        migrations.AddIndex(
+            model_name='downloadevent',
+            index=models.Index(fields=['created', 'storage_region'], name='download_event_crt_regn'),
+        ),
+        migrations.AddIndex(
+            model_name='downloadevent',
+            index=models.Index(fields=['created', 'user_region'], name='download_event_crt_user'),
+        ),
+        migrations.RunPython(create_dashboard_group, remove_dashboard_group),
+    ]

osf/models/__init__.py

@@ -12,6 +12,7 @@
 from .admin_log_entry import AdminLogEntry
 from .admin_profile import AdminProfile
 from .analytics import UserActivityCounter, PageCounter
+from .download_event import DownloadEvent
 from .archive import ArchiveJob, ArchiveTarget
 from .banner import ScheduledBanner
 from .base import (

osf/models/download_event.py

@@ -0,0 +1,64 @@
+from django.db import models
+
+
+class DownloadEvent(models.Model):
+    """One metadata row per download — never the file contents.
+
+    Foundation for the download-telemetry capture and dashboard. Append-only:
+    rows are written from the download flow (single files at the osf.io redirect
+    view, folder/project zips from the WaterButler callback) and read, always
+    scoped to a time range, by the dashboard.
+    """
+
+    FILE = 'file'
+    FOLDER_ZIP = 'folder_zip'
+    PROJECT = 'project'
+    DOWNLOAD_TYPES = (
+        (FILE, 'Single file'),
+        (FOLDER_ZIP, 'Folder zip'),
+        (PROJECT, 'Whole-project zip'),
+    )
+
+    created = models.DateTimeField(auto_now_add=True, db_index=True)
+
+    # what was downloaded
+    resource_guid = models.CharField(max_length=255, blank=True, default='', db_index=True)
+    path = models.TextField(blank=True, default='')
+    download_type = models.CharField(max_length=16, choices=DOWNLOAD_TYPES)
+    # null for single files (only zips stream through WB, which reports completion)
+    zip_completed = models.BooleanField(null=True, blank=True)
+    size_bytes = models.BigIntegerField(null=True, blank=True)
+
+    # storage_region = where the bytes were served from (capacity);
+    # user_region = roughly where the user is. Kept separate on purpose.
+    storage_region = models.CharField(max_length=64, blank=True, default='')
+    user_region = models.CharField(max_length=64, blank=True, default='')
+    ip = models.GenericIPAddressField(null=True, blank=True)
+    source_area = models.CharField(max_length=128, blank=True, default='')
+
+    # nullable: anonymous downloads of public files
+    user = models.ForeignKey(
+        'osf.OSFUser',
+        null=True,
+        blank=True,
+        on_delete=models.SET_NULL,
+        related_name='download_events',
+    )
+
+    class Meta:
+        # `created` is indexed on the field; these cover the dashboard's
+        # time-range group-bys.
+        indexes = [
+            models.Index(fields=['created', 'download_type'], name='download_event_crt_type'),
+            models.Index(fields=['created', 'storage_region'], name='download_event_crt_regn'),
+            models.Index(fields=['created', 'user_region'], name='download_event_crt_user'),
+        ]
+
+    def __repr__(self):
+        return (
+            f'<DownloadEvent(id={self.id}, user={self.user_id}, '
+            f'type={self.download_type}, size_bytes={self.size_bytes})>'
+        )
+
+    def __str__(self):
+        return self.__repr__()