noodlepartners / hedex Goto Github PK

Hedex installs into Sakai and starts gathering data on user logins, assignment submissions, and course visits. This data is made available on a set of RESTful endpoints, authenticated with a particular user based on the requesting agent.

Copy this code into your Sakai installation, alongside the other Sakai modules such as announcement. Change to the hedex directory and compile with maven:

mvn clean install sakai:deploy

Once built, restart your Tomcat. The code will automatically create the tables it needs, HDX_SESSION_DURATION, HDX_COURSE_VISITS, and HDX_ASSIGNMENT_SUBMISSIONS, and then start to populate them as users use the system.

Note: There are both master and 12.x branches. The one we are using in production will be 12.x.

HEDEX returns JSON containing a tenant id. This is initially set to UNSPECIFIED and if you want it set to something other than that use:

hedex.tenantId=OurSakaiTenantId

If you need to, you can disable the digester altogether. This might be useful in a cluster.

hedex.digester.enabled = false (default: true)

HEDEX uses a site property, 'hedex-agent' to identify a site as being associated with a particular analytics consumer. For example, Noodle Partners sites will be marked with the site property 'hedex-agent=noodle'. The digester uses a scheduled executor to update maps of both the marked sites and their members. You can configure the interval for this site scan with:

hedex.site.update.interval = 20

The unit is minutes and it defaults to 30

When retrieving data, a specific Sakai user must be used as the login. This defaults to the reqesting agent, 'noodle' for instance, appended with '-hedex-user'. The agent should also be passed as the ReqestingAgent parameter. Without these two things matching, your requests will be rejected.

Firstly, have a look at scripts/pulldataexamples.py. To pull the HEDEX data, your script first needs to login and get a session string. Every call henceforth passes that session string. The calls will fail otherwise. Data is returned to the spec at: http://petstore.swagger.io/?url=http://employees.idatainc.com/~bparish/NoodleBus/swagger_NoodleBus_Retention_API_v4.yaml

NOTE: assignAvgAttmpt in the returned JSON for Get_Retention_Engagement_Assignments is calculated as the average of the grades applied to the submission. It's not based on the number of times the student submits. In other words, it's not really based on number of attempts, but on gradings by instructors.

To run the script, you'll need python3.x installed, and the SOAP client 'zeep' installed. You can use pip for that: pip install zeep.