improved README and fixed many things

README.rst

@@ -1,46 +1,101 @@
-Tutor Cairn: scalable, real-time analytics for Open edX
-========================================================
+Cairn: scalable, real-time analytics for Open edX
+==================================================
 
-TODO: Sweet readme
+Analytics are an essential component of an online learning platform: you need to know whether your courses are effective and which parts need some improvement. You need to know if your students are falling by, and if they do, you need to detect the early warning signs. When your courses are successful, you want to get periodical engagement reports.
+
+We created a tool to help you answer all these questions. Cairn is a Tutor plugin that you install on top of an Open edX platform and that gives you access to a powerful, full-blown analytics stack. Cairn comes with the following features out of the box:
+
+🖴 **Unified datalake of learner events and stateful data**: both learner-triggered events, coming from the Open edX tracking logs, and stateful data, coming from the existing databases, are available for querying in a single unified interface. This means that you can, for instance, query the grades of the students that visited your platform in the past 24 hours, or collect the email addresses of the students who did not yet complete the latest assignment.
+
+⚡ **Real-time:** new events are visible immediately in your analytics interface. No more waiting for slow batch jobs to complete!
+
+🔑 **Course- and org-based data access rights:** your course staff is granted access only to the data rows that concern them. Cairn makes it easy to create new users with granular access permissions.
+
+🎁 **Working dashboards out of the box:** Cairn comes with a fully functional dashboard that you can start playing with right away.
+
+🛠️ **Fully customizable data and dashboards:** your data scientists, business intelligence team and other tinkerers can freely explore your course data, create and share their own queries, datasets and dashboards. All it takes is a little bit of SQL.
+
+🚀 **Scalable:** Cairn scales as much as its backend, which was designed for Internet scale.
+
+Cairn vs alternatives
+---------------------
+
+========================================== =====  ===================================================================================  ===================================================
+List of features                           Cairn  `Open edX Insights <https://edx.readthedocs.io/projects/edx-insights/en/latest/>`__  `Figures <https://github.com/appsembler/figures>`__
+========================================== =====  ===================================================================================  ===================================================
+Event aggregation                            ✅      ✅                                                                                    ❌
+Real-time data                               ✅      ❌                                                                                    ✅
+Easy to install                              ✅      ❌                                                                                    ✅
+Custom queries and dashboards                ✅      ❌                                                                                    ❌
+Works with the latest Open edX versions      ✅      ✅                                                                                    ❌
+========================================== =====  ===================================================================================  ===================================================
+
+
+How does Cairn work?
+--------------------
+
+Cairn uses the same collect/store/expose paradigm made popular by other frameworks such as the `ELK Stack <https://www.elastic.co/fr/elastic-stack>`__ -- excepts that all the components are different and better suited to Open edX:
+
+- On the server side, tracking logs are collected by `Vector <https://vector.dev/>`__, an efficient, cloud-native log collector.
+- Tracking log events are then stored in a `Clickhouse <https://clickhouse.tech/>`__ table, which is the cornerstone of Cairn. Clickhouse also exposes MySQL data via live and materialized views. This is the magic piece of the puzzle which allows us to join event and MySQL data.
+- The data inside Clickhouse is made visible to the end-users in a `Superset <https://superset.apache.org/>`__ frontend.
 
 Installation
 ------------
 
-::
+Cairn requires a `Tutor Wizard Edition license <https://overhang.io/tutor/wizardedition>`__. Once you have enabled your license, installing the plugin is as simple as running::
 
     tutor license install tutor-cairn
 
 Usage
 -----
 
-::
+Getting started
+~~~~~~~~~~~~~~~
+
+Enable the plugin with::
 
     tutor plugins enable cairn
+
+Then, restart your platform and run the initialization scripts::
+
     tutor local quickstart
 
+Create credentials to access the Clickhouse database::
+
+    tutor local run cairn-clickhouse cairn createuser YOURUSERNAME
+
 Create an admin user to access the frontend::
 
     # You will be prompted for a new password
-    tutor local run cairn-superset superset fab create-admin --username yourusername --email [email protected]
+    tutor local run cairn-superset cairn createuser --admin YOURUSERNAME [email protected]
 
 You can then access the frontend with the user credentials you just created. Open http(s)://data.<YOUR_LMS_HOST> in your browser. When running locally, this will be http://data.local.overhang.io. The admin user will automatically be granted access to the "openedx" database in Superset and will be able to query all tables.
 
-Management
-----------
+To import the "Course overview" dashboard that comes with Cairn, run::
 
-Most of your users should probably not have access to all data from all courses. To restrict a given user to one or more courses or organizations, select the course IDs and/or organization IDS to which the user should have access and create a user with limited access to the datalake::
+    tutor local run cairn-superset cairn bootstrap-dashboards YOURUSERNAME /app/bootstrap/courseoverview.json
+
+Some event data will be missing from your dashboards: just start using your LMS and refresh your dashboard. The new events should appear immediately.
+
+.. image:: https://overhang.io/static/catalog/screenshots/cairn.png
+    :alt: Alpine cairn
 
-    tutor local run cairn-clickhouse cairn createuser --course-id='course-v1:edX+DemoX+Demo_Course' --org-id='edX' yourusername
+Data-based access control
+~~~~~~~~~~~~~~~~~~~~~~~~~
 
-Then, create the corresponding user on the frontend::
+Most of your users should probably not have access to all data from all courses. To restrict a given user to one or more courses or organizations, select the course IDs and/or organization IDS to which the user should have access and create a user with limited access to the datalake::
+
+    tutor local run cairn-clickhouse cairn createuser --course-id='course-v1:edX+DemoX+Demo_Course' --org-id='edX' YOURUSERNAME
 
-    tutor local run cairn-superset cairn createuser yourusername [email protected]
+Then, create the corresponding user on the frontend with the same command as above (but without the ``--admin`` option)::
 
-Your frontend user will automatically be associated to the datalake database you created, provided they share the same name.
+    tutor local run cairn-superset cairn createuser YOURUSERNAME [email protected]
 
-Cairn comes with a convenient pre-built dashboard that you can add to any user account::
+Your frontend user will automatically be associated to the datalake database you created.
 
-    tutor local run cairn-superset cairn bootstrap-dashboards yourusername /app/bootstrap/courseoverview.json
+Refreshing course block data
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Course block IDs and names are loaded from the Open edX modulestore into the datalake. After making changes to your course, you might want to refresh the course structure stored in the datalake. To do so, run::
 
@@ -53,18 +108,20 @@ Or, if you want to avoid running the full plugin initialization::
         -v $(tutor config printroot)/env/plugins/cairn/apps/clickhouse/auth.json:/openedx/clickhouse-auth.json \
         lms python /openedx/scripts/importcoursedata.py
 
+Running on Kubernetes
+~~~~~~~~~~~~~~~~~~~~~
+
 When running on Kubernetes instead of locally, most commands above can be re-written with `tutor k8s exec service "command"` instead of `tutor local run service command`. For instance::
 
     # Privileved user creation
-    tutor k8s exec cairn-superset "superset fab create-admin --username yourusername --email [email protected]"
+    tutor k8s exec cairn-superset "superset fab create-admin --username YOURUSERNAME --email [email protected]"
     # Unprivileged user creation
-    tutor k8s exec cairn-clickhouse "cairn createuser --course-id='course-v1:edX+DemoX+Demo_Course' --org-id='edX' yourusername"
-    tutor k8s exec cairn-superset "cairn createuser yourusername [email protected]"
+    tutor k8s exec cairn-clickhouse "cairn createuser --course-id='course-v1:edX+DemoX+Demo_Course' --org-id='edX' YOURUSERNAME"
+    tutor k8s exec cairn-superset "cairn createuser YOURUSERNAME [email protected]"
 
 Development
 -----------
 
-
 To reload Vector configuration after changes to vector.toml, run::
 
     tutor config save && tutor local exec cairn-vector sh -c "kill -s HUP 1"
@@ -77,6 +134,13 @@ To launch a Python shell in Superset, run::
 
     tutor local run cairn-superset superset shell
 
+.. image:: https://overhang.io/static/catalog/img/cairn.png
+    :alt: Alpine cairn
+
+Support
+-------
+
+Are you having trouble with Cairn? Do you have questions about this plugin? Please get in touch with us at [email protected]. Community support is also available on the official Tutor forums: https://discuss.overhang.io
 
 License
 -------

setup.py

@@ -39,7 +39,7 @@ def load_about():
     packages=find_packages(exclude=["tests*"]),
     include_package_data=True,
     python_requires=">=3.5",
-    install_requires=["tutor-openedx"],
+    install_requires=["tutor-openedx>=11.0.0,<12.0.0"],
     entry_points={"tutor.plugin.v0": ["cairn = tutorcairn.plugin"]},
     classifiers=[
         "Development Status :: 3 - Alpha",

tutorcairn/patches/k8s-deployments

@@ -85,14 +85,19 @@ spec:
             - name: SYSFS_ROOT
               value: /host/sys
           volumeMounts:
+            - name: data
+              mountPath: /var/lib/vector
             - name: var-log
               mountPath: /var/log/
               readOnly: true
             - mountPath: /etc/vector/vector.toml
               name: config
-              subPath: vector.toml
+              subPath: k8s.toml
               readOnly: true
       volumes:
+        - name: data
+          persistentVolumeClaim:
+            claimName: cairn-vector
         - name: var-log
           hostPath:
             path: /var/log/

tutorcairn/patches/k8s-volumes

@@ -1,3 +1,17 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: cairn-vector
+  labels:
+    app.kubernetes.io/component: volume
+    app.kubernetes.io/name: cairn-vector
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 2Gi
 {% if CAIRN_RUN_CLICKHOUSE %}
 ---
 apiVersion: v1

tutorcairn/patches/kustomization-configmapgenerator

@@ -1,6 +1,6 @@
 - name: cairn-vector-config
   files:
-    - plugins/cairn/apps/vector/vector.toml
+    - plugins/cairn/apps/vector/k8s.toml
 - name: cairn-clickhouse-user-config
   files:
     - plugins/cairn/apps/clickhouse/users.d/cairn.xml

tutorcairn/patches/local-docker-compose-services

@@ -4,7 +4,8 @@
 cairn-vector:
     image: docker.io/timberio/vector:0.13.X-alpine
     volumes:
-        - ../plugins/cairn/apps/vector/vector.toml:/etc/vector/vector.toml:ro
+        - ../../data/cairn/vector:/var/lib/vector
+        - ../plugins/cairn/apps/vector/local.toml:/etc/vector/vector.toml:ro
         {% if CAIRN_DOCKER_HOST_SOCK_PATH %}- {{ CAIRN_DOCKER_HOST_SOCK_PATH }}:/var/run/docker.sock:ro{% endif %}
     environment:
         - DOCKER_HOST=/var/run/docker.sock
@@ -52,7 +53,7 @@ cairn-superset-worker-beat:
         - cairn-redis
         - cairn-postgresql
 cairn-redis:
-    image: docker.io/redis:5.0-alpine
+    image: docker.io/redis:6.2.4-alpine
     restart: unless-stopped
 {% if CAIRN_RUN_POSTGRESQL %}
 cairn-postgresql:

tutorcairn/templates/cairn/apps/openedx/scripts/importcoursedata.py

@@ -38,7 +38,7 @@ def import_course(course_key):
     course = get_course(course_key, depth=None)
     print("======================", course_id, course.display_name)
     values = [
-        sql_query(
+        sql_format(
             "('{}', '{}', '{}', '{}', '{}', '{}')",
             course_id,
             str(child.location),
@@ -54,12 +54,12 @@ def import_course(course_key):
             f"Inserting {len(values)} items in course_blocks for course '{course_id}'..."
         )
         make_query(
-            sql_query(
+            sql_format(
                 "ALTER TABLE course_blocks DELETE WHERE course_id = '{}';",
                 course_id,
             ),
         )
-        insert_query = sql_query(
+        insert_query = sql_format(
             "INSERT INTO course_blocks (course_id, block_key, block_id, position, display_name, full_name) VALUES "
         )
         insert_query += ", ".join(values)
@@ -74,7 +74,7 @@ def iter_course_blocks(item, prefix=""):
         yield from iter_course_blocks(child, prefix=prefix)
 
 
-def sql_query(template, *args, **kwargs):
+def sql_format(template, *args, **kwargs):
     args = [sql_escape_string(arg).decode() for arg in args]
     kwargs = {key: sql_escape_string(value).decode() for key, value in kwargs.items()}
     return template.format(*args, **kwargs)