Add ruleset template #302
Add ruleset template #302
Conversation
| feedback_form=Form([question], submit_label='Continue', is_profile=True)).action() | ||
| for question in questions | ||
| ] | ||
|
|
There was a problem hiding this comment.
To my understanding, a questionnaire cannot be part of first_round. We don't have a session before the StartSession view, so we wouldn't be able to process the results. So better to move this to next_round instead.
There was a problem hiding this comment.
As things are now, we save profile data with Participant, so asking Profile questions in the first_round() is common in many experiments. If we decide to link Profile answers to Session, then many other experiments will also have to be modified.
There was a problem hiding this comment.
I haven't been able to find any experiment rules in which there are Profile questions in the first round. For the participant, the difference should also not be noticeable, as StartSession is a "silent" view.
There was a problem hiding this comment.
For example, the categorization experiment has a questionnaire in the first_round. Consent is also a profile question that is in every experiment's first_round.
It should not be noticeable to the participant, but the implementation depends on what our choice for saving Profile questions will be.
There was a problem hiding this comment.
I'd say the Categorization experiment is an atypical case, so perhaps better to follow the majority of the experiments here, which don't ask questions before StartSession?
| session.save() | ||
|
|
||
| # Calculate average score | ||
| score_avg = session.result_set.all().aggregate(Avg('score'))['score__avg'] |
There was a problem hiding this comment.
Perhaps this should be also moved to a class method get_final_view or something along these lines? This makes this easier to override.
There was a problem hiding this comment.
Nice to have this updated, thanks! Before we merge though -- I've been wondering whether we can make the base.py class the template, as all other rules files derive from that. It would be neat if the template is in itself a working experiment, then subclassing / overriding functions would make the first steps towards your own experiment easiest. Perhaps something to look into tomorrow?
| @@ -200,6 +201,6 @@ def get_feedback(cls, session): | |||
| instruction=instruction, | |||
| steps=[], | |||
| button_label=button_label | |||
| ).action() | |||
| ) | |||
|
|
|||
| return feedback | |||
|
Should we close this PR now that we have a ruleset template in the |
|
We should incorporate this work as a starting place for the 2AFC paradigm, I think. I'm thinking |
Okay, in that case we might want to update the name of this PR and its corresponding issue to avoid confusion. |
|
I have updated this PR to reflect new changes in our infrastructure. I am not sure if there is any need to use a command to "generate" this template, because it is already "generated" and sitting in the rules directory. I think we should merge it as it is, close this PR and reopen if we come up with new ideas. |
There was a problem hiding this comment.
We don't need an empty consent form. Consent is now set through admin or otherwise will be lorem ipsum by default.
|
You are perfectly right that we should merge this PR. I do think that instead of "ruleset_template", the experiment could be renamed to "2afc". And let's refer in the wiki to this experiment as a nice guideline of how to set up an experiment. Perhaps even ship a mini playlist in this repo to use it with? (Wiki / playlist: leave for later.) Calling it "ruleset_template" might suggest to users that this one is not an experiment in its own right, while I think it can be nicely tweaked for multiple purposes. |
albertas-jn
requested review from
BeritJanssen
and removed request for
drikusroor
Closes #99