Extend metrics

It should be quite simple to extend metrics implementing your own.

If you do, please share your metrics with the community : )

Basic metric

Metrics are implemented as Python classes in the file models/metrics.py and must respect the following properties:

• name must start with Metric
• must have the following properties:
  • description: a brief text on what the metrics does
  • target_type: a choice specifying the type of the target (one among TARGET_USERS, TARGET_TWEETS, TARGET_ANY, TARGET_BOTHS)
• must overwrite the method computation(), which returns True when the computation ends without errors

Each Metric class, by default, depending on its target_type, has the attributes tweets and/or twitter_users, that are Django Querysets objects of type Tweet and/or TwitterUser. By default, they contain the selection target on which the metric was executed (alteratively, if executed campaign-wide, they contain all tweets/users linked to a campaign).

After you have implemented your Metric, the database need to be updated. To do so, run:

python manage.py makemigrations
python manage.py migrate

If you want to be able to manege your metric from the Django default admin application, remember to register it in the admin.py file.

More options

Metrics can be highly customised. Have a look at the code to learn how; if you need help, get in contact

Some common customisation might be:

• make a custom result page:
  • add a template file in templates/metrics
  • override the template_file attribute, pointing to your custom template
• add custom parameters:
  • create a form adding a template in templates/metrics/forms
  • override the template_form attribute, pointing to your form template file
  • override the method set_params_from_req to check and set your custom parameters
  • add a template file in templates/custom_fields to show the custom field values in the metric output page
  • override the template_form attribute, pointing to your custom code

If your metric need to perform some potentially long background operation that should not keep other metrics to run, you can add an Operation class to the models/operations.py file. See the code of MetricGraphCommunityNetwork for an example.