Studies rewriting

changed milestone to %0.4: Tests and studies improvements

added Studies & tests To do labels

assigned to @bridubois

marked this issue as related to #165 (closed)

marked this issue as related to #164 (closed)

marked this issue as related to #154 (closed)

marked this issue as related to #163

marked this issue as related to #139 (closed)

added Meta label and removed To do label

@dvanrossum for info

marked this issue as related to #162

marked this issue as related to #158

marked this issue as related to #157 (closed)

marked this issue as related to #153

changed the description

marked this issue as related to #84

changed the description

marked this issue as related to #80

marked this issue as related to #149

marked this issue as related to #97

marked the checklist item Does all studies allow to create sessions? Should this be configurable? as completed

marked the checklist item Is it possible to do the tests (typing & "surveys") without a study? (So, should we keep the current workflow?) #139 (closed) answer a part of it, but is it still how we should handle it? as completed

marked the checklist item Does studies define a code to join in, or do you require a direct link/registration (or is this configurable)? If so, can we have multiple code for one study? as completed

marked the checklist item Should the typing tests be reusable? The WIP workflow in #165 (closed) manage them at study's level, to prevent having to jump to too many pages as they are relatively short to create and adapt. But should we create a band new creation and management workflow for them? as completed

marked the checklist item Some of the issues linked before are quite old (#97, #84, #80, ...). Are they all still relevant? as completed

Great consolidation work! I appreciate it a lot.

Here are my responses to the open questions:

Do all studies allow to create sessions? Should this be configurable

Some studies will only involve tests and no use of the chat sessions, nor tutor selection, etc. You can handle this as the impossibility to create sessions or as the fact that the participant will never land on the sessions page. It's not really that they are forbidden to see the chat interface, it's just not part of that study. I hope it answers the question.

Is it possible to do the tests (typing & "surveys") without a study?

We could have a default "catch-all" study (study_id=0 or 1) that would show to the user as "No specific study" or "Just testing". So if someone is not involved in a study, they still have access to the test and the data is not lost.
BUT if it's simpler to always require a study (either in the URL or in the first select), I'm OK with that solution as well.

Do studies define a code to join in, or do you require a direct link/registration (or is this configurable)? If so, can we have multiple codes for one study?

I'm not sure I understand what "code" you're referring to. I'd say just study_id in URL or in selector. I don't think certain studies should be hidden behind a passcode or something like that. Am I answering your question?

Should the typing tests be reusable? The WIP workflow in #165 (closed) manage them at study's level, to prevent having to jump to too many pages as they are relatively short to create and adapt. But should we create a band new creation and management workflow for them?

Do whatever is simpler and cleaner in terms of code. Many studies won't have a typing test, so it's OK if we have to input the test manually for each study (if that's what you mean). Yet, there are also relatively few possible options for the test (for now: one version in English, one version in French, one version for English and French together... maybe an extra short version for EN/FR, that's it.), so it would also make sense to define them somewhere and just pick one among a list ("English, French, French-English, short English, short French").

Some of the issues linked before are quite old (#97, #84, #80, ...). Are they all still relevant?

Good point. They are not completely irrelevant, but they are certainly not a priority. They are "nice to have", but we could also close them and tag them as a "possible future development" if necessary to clean the repo.

I hope this answers everything! Thanks a lot!

marked the checklist item Developement Server - @sbibauw & @bridubois as completed

marked the checklist item Studies frontend for creation and edition - @dvanrossum as completed

changed the description

marked this issue as related to #169

Here are screenshots of the current look of a study flow.

marked the checklist item (Standalone) tests as a part of study -> part of test URL as completed

marked the checklist item tests: compute score for gapfills as completed

mentioned in issue #160

@dvanrossum @sbibauw @LucieBocken @DavePikopPokam

First of all,thank you all for your work on all those new features, it really starts to look great.

I've been able to merge the new studies & tests flow as well as the registration. Everything is now live on dev.languagelab.be and ready for extensive testings. Don't hesitate to open a new issue for everything you have a problem with, or if you see some improvements that we can do.

You can assign them depending on their nature:

Registration -> @DavePikopPokam
Studies & tests (Frontend) -> @dvanrossum
Studies & tests (Backend) -> @bridubois

If you are hesitant, you can always open it without assigning anyone and I will assign them accordingly.

Here are the different new features that you can try to hack:

/register -> Register as a student or a tutor.
/studies -> Take part in a study.
/studies/[ID] -> Take part in study ID.
/admin/studies -> Create, edit and download studies data.

@sbibauw As the way we store entries has changed a lot, I "simply" dropped the previous column and data. I guess you will need them, so I've of course wrote a backup, and anyway it's not yet deleted on the production server. Can we keep them that way, or do we have to write a conversion script to encode the "old" entries into the new database? The studies themself are preserved, only the entries are affected.

@dvanrossum @DavePikopPokam I will do some branch cleaning, but be careful that as we had a lot of changes on dev, you have to rebase any work-in-progress branch into the latest version of dev, otherwise we would have many merge conflict. I also voluntarily deleted all the merge requests branches, because it's better to create a new branch for bug fixing that continue on previous branches, as we had many conflicts during the merges.

Great work, everyone! Thanks!

@bridubois Re your question, not sure which column/data you're referring to. survey_responses? Or something else? In any case, I don't think we need a conversion script, but yes a clean export of the important data that would be scraped out by the new structure.

Studies rewriting

Prerequisites

Front-End

Back-End

Open questions (mainly for @sbibauw, but everybody's inputs are welcome!)

Designs

Éléments enfants ...

Activité