feat(Analytics): add an analytics module (#773)

* feat(Analytics): add an analytics module

* chore(analytics): Use a middleware instead of signals when setting user properties, better error handling, removing useless condition

* fix tests

---------

Co-authored-by: Quentin Gérôme <[email protected]>

config/settings/base.py

@@ -54,6 +54,7 @@
     "ariadne_django",
     "dpq",
     "hexa.user_management",
+    "hexa.analytics",
     "hexa.core",
     "hexa.catalog",
     "hexa.countries",
@@ -92,6 +93,7 @@
     "django.contrib.messages.middleware.MessageMiddleware",
     "django.middleware.clickjacking.XFrameOptionsMiddleware",
     "hexa.user_management.middlewares.login_required_middleware",
+    "hexa.analytics.middlewares.set_analytics_middleware",
 ]
 
 ROOT_URLCONF = "config.urls"
@@ -196,7 +198,7 @@
 RAW_CORS_ALLOWED_ORIGINS = os.environ.get("CORS_ALLOWED_ORIGINS")
 if RAW_CORS_ALLOWED_ORIGINS is not None:
     CORS_ALLOWED_ORIGINS = RAW_CORS_ALLOWED_ORIGINS.split(",")
-    CORS_URLS_REGEX = r"^/graphql/(\w+\/)?$"
+    CORS_URLS_REGEX = r"^[/graphql/(\w+\/)?|/analytics/track]$"
     CORS_ALLOW_CREDENTIALS = True
 
 

config/urls.py

@@ -37,6 +37,7 @@
     path("notebooks/", include("hexa.notebooks.urls", namespace="notebooks")),
     path("pipelines/", include("hexa.pipelines.urls", namespace="pipelines")),
     path("workspaces/", include("hexa.workspaces.urls", namespace="workspaces")),
+    path("analytics/", include("hexa.analytics.urls", namespace="analytics")),
     # Order matters, we override the default logout view defined later
     # We do this to logout the user from jupyterhub at the end of the openhexa
     # session. the jupyterhub will redirect to the openhexa login after it

hexa/core/analytics.py → hexa/analytics/api.py

@@ -13,7 +13,7 @@
 
 
 def track(
-    request: HttpRequest,
+    request: HttpRequest | None,
     event: str,
     properties: dict = {},
     user: User = None,
@@ -27,20 +27,20 @@ def track(
         An identifier for the event to track.
     properties : dict
        A dictionary holding the event properties
-    user: User |
+    user: User
        User entity to track
     """
     if mixpanel is None:
         return
 
     people = user if user else getattr(request, "user", None)
     can_track = (
-        people is None or isinstance(user, AnonymousUser) or people.analytics_enabled
+        people is None or isinstance(people, AnonymousUser) or people.analytics_enabled
     )
     if can_track is False:
         return
 
-    if request:
+    if request and "User-Agent" in request.headers:
         # Add request related properties
         parsed = user_agent_parser.Parse(request.headers["User-Agent"])
         properties.update(
@@ -66,7 +66,7 @@ def set_user_properties(user: User):
         return
 
     try:
-        mixpanel.people_set_once(
+        mixpanel.people_set(
             distinct_id=str(user.id),
             properties={
                 "$email": user.email,

hexa/analytics/apps.py

@@ -0,0 +1,8 @@
+from hexa.app import CoreAppConfig
+
+
+class AnalyticsConfig(CoreAppConfig):
+    name = "hexa.analytics"
+    label = "analytics"
+
+    ANONYMOUS_URLS = ["analytics:track"]

hexa/analytics/middlewares.py

@@ -0,0 +1,19 @@
+from typing import Callable
+
+from django.http import HttpRequest, HttpResponse
+
+from hexa.analytics.api import set_user_properties
+
+
+def set_analytics_middleware(
+    get_response: Callable[[HttpRequest], HttpResponse],
+) -> Callable[[HttpRequest], HttpResponse]:
+    """Send the user properties to the analytics service."""
+
+    def middleware(request: HttpRequest) -> HttpResponse:
+        response = get_response(request)
+        if getattr(request, "user") and request.user.is_authenticated:
+            set_user_properties(request.user)
+        return response
+
+    return middleware

hexa/core/tests/test_analytics.py → hexa/analytics/tests/test_analytics.py

@@ -3,7 +3,7 @@
 
 from django.test import RequestFactory
 
-from hexa.core.analytics import set_user_properties, track
+from hexa.analytics.api import set_user_properties, track
 from hexa.core.test import TestCase
 from hexa.files.tests.mocks.mockgcp import mock_gcp_storage
 from hexa.pipelines.models import Pipeline, PipelineRunTrigger
@@ -36,7 +36,7 @@ def setUpTestData(cls):
         )
         cls.factory = RequestFactory()
 
-    @mock.patch("hexa.core.analytics.Mixpanel")
+    @mock.patch("hexa.analytics.api.Mixpanel")
     def test_track_event_user_no_token(self, mock_mixpanel):
         with self.settings(
             MIXPANEL_TOKEN=None,
@@ -54,7 +54,7 @@ def test_track_event_user_no_token(self, mock_mixpanel):
             track(request, event, properties, user=self.USER)
             mock_mixpanel.assert_not_called()
 
-    @mock.patch("hexa.core.analytics.Mixpanel")
+    @mock.patch("hexa.analytics.api.Mixpanel")
     def test_track_user_analytics_not_enabled(
         self,
         mock_mixpanel,
@@ -85,7 +85,7 @@ def test_track_user_analytics_not_enabled(
             track(request, event, properties, user=self.USER)
             mock_mixpanel_instance.assert_not_called()
 
-    @mock.patch("hexa.core.analytics.mixpanel")
+    @mock.patch("hexa.analytics.api.mixpanel")
     def test_track_user_analytics_enabled(
         self,
         mock_mixpanel,
@@ -113,7 +113,7 @@ def test_track_user_analytics_enabled(
                 },
             )
 
-    @mock.patch("hexa.core.analytics.mixpanel")
+    @mock.patch("hexa.analytics.api.mixpanel")
     def test_track_pipelinerun_no_user(
         self,
         mock_mixpanel,
@@ -141,7 +141,7 @@ def test_track_pipelinerun_no_user(
                 distinct_id=None, event_name=event, properties=properties
             )
 
-    @mock.patch("hexa.core.analytics.mixpanel")
+    @mock.patch("hexa.analytics.api.mixpanel")
     def test_create_user_profile(
         self,
         mock_mixpanel,
@@ -151,7 +151,7 @@ def test_create_user_profile(
         with self.settings(MIXPANEL_TOKEN=mixpanel_token):
             set_user_properties(self.USER)
 
-            mock_mixpanel.people_set_once.assert_called_once_with(
+            mock_mixpanel.people_set.assert_called_once_with(
                 distinct_id=str(self.USER.id),
                 properties={
                     "$email": self.USER.email,

hexa/analytics/tests/test_views.py

@@ -0,0 +1,54 @@
+from unittest import mock
+
+from django.urls import reverse
+
+from hexa.core.test import TestCase
+from hexa.user_management.models import User
+
+
+class AnalyticsTests(TestCase):
+    @classmethod
+    def setUpTestData(cls):
+        cls.USER = User.objects.create_user(
+            "[email protected]",
+            "user_password",
+        )
+
+    def test_track_event_bad_request(self):
+        self.client.force_login(self.USER)
+        r = self.client.post(
+            reverse(
+                "analytics:track",
+            ),
+            {"properties": {}},
+            content_type="application/json",
+        )
+        self.assertEqual(r.status_code, 400)
+        self.assertEqual(r.json(), {"Bad request": "event name is required."})
+
+    def test_track_event(self):
+        self.client.force_login(self.USER)
+        r = self.client.post(
+            reverse(
+                "analytics:track",
+            ),
+            {"event": "page_viewed", "properties": {"page": "database"}},
+            content_type="application/json",
+        )
+        self.assertEqual(r.status_code, 200)
+
+    @mock.patch("hexa.analytics.api.mixpanel")
+    def test_track_event_analytics_not_enabled(self, mixpanel_mock):
+        self.USER.analytics_enabled = False
+        self.USER.save()
+        self.client.force_login(self.USER)
+        with self.settings(MIXPANEL_TOKEN="123"):
+            r = self.client.post(
+                reverse(
+                    "analytics:track",
+                ),
+                {"event": "page_viewed", "properties": {"page": "database"}},
+                content_type="application/json",
+            )
+            self.assertEqual(r.status_code, 200)
+            mixpanel_mock.track.assert_not_called()

hexa/analytics/urls.py

@@ -0,0 +1,9 @@
+from django.urls import path
+
+from . import views
+
+app_name = "analytics"
+
+urlpatterns = [
+    path("track", views.track_event, name="track"),
+]

hexa/analytics/views.py

@@ -0,0 +1,27 @@
+import json
+
+from django.http import HttpRequest, HttpResponse, JsonResponse
+from django.views.decorators.csrf import csrf_exempt
+from django.views.decorators.http import require_POST
+
+from .api import track
+
+
+@csrf_exempt
+@require_POST
+def track_event(request: HttpRequest) -> HttpResponse:
+    """This API endpoint is called by the frontend to track events."""
+    try:
+        payload = json.loads(request.body.decode("utf-8"))
+    except json.JSONDecodeError:
+        return JsonResponse({"Bad request": "Invalid JSON payload."}, status=400)
+
+    if "event" not in payload:
+        return JsonResponse({"Bad request": "event name is required."}, status=400)
+
+    track(
+        request,
+        payload.get("event"),
+        payload.get("properties", {}),
+    )
+    return JsonResponse({}, status=200)

hexa/datasets/schema/mutations.py

@@ -2,7 +2,7 @@
 from django.core.exceptions import PermissionDenied, ValidationError
 from django.db import IntegrityError, transaction
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.pipelines.authentication import PipelineRunUser
 from hexa.workspaces.models import Workspace
 

hexa/files/schema/mutations.py

@@ -1,6 +1,6 @@
 from ariadne import MutationType
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.files.api import get_storage
 from hexa.workspaces.models import Workspace
 

hexa/notebooks/schema.py

@@ -5,7 +5,7 @@
 from django.conf import settings
 from django.http import HttpRequest
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.workspaces.models import Workspace
 
 from .api import create_server, create_user, get_user, server_ready

hexa/pipelines/management/commands/pipelines_scheduler.py

@@ -6,7 +6,7 @@
 from django.core.management.base import BaseCommand
 from django.utils import timezone
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.pipelines.models import Pipeline, PipelineRunTrigger
 
 logger = getLogger(__name__)

hexa/pipelines/schema/mutations.py

@@ -6,7 +6,7 @@
 from django.db import IntegrityError
 from django.http import HttpRequest
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.databases.utils import get_table_definition
 from hexa.files.api import NotFound, get_storage
 from hexa.pipelines.authentication import PipelineRunUser

hexa/pipelines/views.py

@@ -16,8 +16,8 @@
 from django.views.decorators.csrf import csrf_exempt
 from django.views.decorators.http import require_POST
 
+from hexa.analytics.api import track
 from hexa.app import get_hexa_app_configs
-from hexa.core.analytics import track
 from hexa.pipelines.models import Environment
 
 from .credentials import PipelinesCredentials

hexa/user_management/apps.py

@@ -14,6 +14,3 @@ class UserManagementConfig(CoreAppConfig):
         "password_reset_done",
         "password_reset_complete",
     ]
-
-    def ready(self):
-        from . import signals  # noqa

hexa/user_management/schema.py

@@ -25,7 +25,7 @@
 from django_otp.plugins.otp_email.models import EmailDevice
 from graphql import default_field_resolver
 
-from hexa.core.analytics import track
+from hexa.analytics.api import track
 from hexa.core.graphql import result_page
 from hexa.core.templatetags.colors import hash_color
 from hexa.user_management.models import (