feat: Cross Product Recommendations Logic #32003
Conversation
| ### DEFAULT KEY DICTIONARY FOR CROSS_PRODUCT_RECOMMENDATIONS ### | ||
| CROSS_PRODUCT_RECOMMENDATIONS_KEYS = {} | ||
|
|
There was a problem hiding this comment.
The dictionary values for each environment will be stored in the edx-internal repository. This PR is for adding in the values for the dictionaries, but it is still a work in progress for getting keys for staging
lms/djangoapps/learner_recommendations/tests/test_serializers.py
Outdated
Show resolved
Hide resolved
lms/djangoapps/learner_recommendations/tests/test_serializers.py
Outdated
Show resolved
Hide resolved
| @ddt.data( | ||
| ("HKUx+FinTechT1x", ["ColumbiaX+BC24FNTC", "MITx+BLN"]), | ||
| ("RITx+CYBER501x", ["UniversityofUtah+BC24CYB", "HarvardX+CYB"]), | ||
| ('NoKeyAssociated', None) | ||
| ) |
There was a problem hiding this comment.
Can we update these course keys to some random test keys? We should avoid using actual course keys in the tests.
| course_data = [get_course_data(key, fields) for key in associated_course_keys] | ||
| filtered_courses = [course for course in course_data if course] |
There was a problem hiding this comment.
- Adding the query string
marketable_course_runs_only': 1will reduce the size of response and - return course runs for a course that are active. This will help recommend course runs that are active to users and not course runs that a user cannot enroll in.
| course_data = [get_course_data(key, fields) for key in associated_course_keys] | |
| filtered_courses = [course for course in course_data if course] | |
| course_data = [get_course_data(key, fields, querystring={'marketable_course_runs_only': 1}) for key in associated_course_keys] | |
| filtered_courses = [course for course in course_data if course and course_data.get("course_runs")] |
|
EdX Release Notice: This PR has been deployed to the staging environment in preparation for a release to production. |
|
EdX Release Notice: This PR has been deployed to the production environment. |
1 similar comment
|
EdX Release Notice: This PR has been deployed to the production environment. |
|
Is this feature available in open edX installations, or does it depend on a commercial service? None of the linked docs are public so I can't tell. |
|
It's not and has now been moved to private plugin. |
Description
This PR encompasses our initial logic for our Personalised Cross Product Recommendation Upsell project. It consists of adding a new endpoint to the learner recommendation app where associated course data is returned based on the key provided to the endpoint.
Key additions are a new view, utility method, and serializer that utilises fields from an existing serializer, but houses an extra field, "course_type". The experiment will be targeted towards anonymous user and will include course filtering based on their location.
Example of a successful response from the endpoint using dummy data:
Supporting information
Finishes PTECH-3169, PTECH-3170, PTECH-3171, PTECH-3172, PTECH-3190
PR for adding dictionary values
Testing instructions
We tested this endpoint by adding an entry to the
CROSS_PRODUCT_RECOMMENDATIONS_KEYSsettings variable of one course key associated to an array of two course keys, eg"TESTx+KEYx": ["TESTx+KEYy", "TESTx+KEYz]. These course keys will need to obviously be of courses that are present in your local dev environment. Then we query the endpoint/api/learner_recommendations/cross_product/<course_id>/and then expect back an array of two courses, or an empty array of courses.