Update api and structure to support json schema #1178
Conversation
| text = serializers.SerializerMethodField() | ||
| # text = serializers.SerializerMethodField() | ||
| user = ShortUserSerializer() | ||
|
|
||
| def get_text(self, obj): | ||
| is_filtered = getattr(obj, "group_applications_filtered", False) | ||
| if is_filtered: | ||
| return None | ||
| return obj.text | ||
| # def get_text(self, obj): | ||
| # is_filtered = getattr(obj, "group_applications_filtered", False) | ||
| # if is_filtered: | ||
| # return None | ||
| # return obj.text |
There was a problem hiding this comment.
Whopsies... Will delete this before merging
There was a problem hiding this comment.
Skimmed through on my phone. Looks good :)
Although this seems like it would be a lot easier to do using pydantic (schema definition, types, validation etc) but this works so I won't force you to do anything. But worth keeping in mind for similar usecases later.
| print(responses) | ||
| print(question) |
Wow thanks a lot, that looks like a much better option! Will take a look at it |
|
Converted to draft to ensure it is not merged before 2023 admissions, will merge some time after that. |
norbye
added
approved
danielyanghansen
force-pushed
the
staging
branch
from
6a9cec2
to
09047bd
Compare
|
This was quite annoying to rebase at this point, so I'm closing it off for now - but keeping the branch so I can look into it at a later time, possibly changing the backend to use pydantic(: |
tl;dr: these PRs enable admins to customize info text, what questions are asked and how many questions are asked (with no need of Webkom supervision - hopefully)
Chunky changes, so I tried to spread it out a little to see if it makes it simpler to review. In case that is not the case, the complete code is available in branch https://github.com/webkom/admissions/tree/json-schema-complete. It might be necessary to run the code from this branch if you want to test it locally, as the codesplitting between the three PRs might be faulty.
This PR is the main bulk of functionality changes, and is required to make the other PRs run and make sense.
There is still plenty to be desired, so this is in no way perfect - but this should be a decent MVP with all necessary functionality intact, and enough to merge it into staging and continue working on it.
Testing is also non-existent, but I'll get to that later..
Changes
Previously, admissions were hardcoded so every group had a
response_labelstring - which is the text prompt that applying users see over the textarea. Each application and group application then had a storedtextstring, which was whatever the user wrote into the application.Now, I have restructured it so that each group contains a
questionsJSON object, and each application and group application contains aresponsesJSON object.The questions-object consists of what the group admins choose for it to be, configured through the admin panel (PR #1180).
So far this can either be a text (normal paragraph), textarea (long text input), textinput (short text) or a phoneinput (norwegian phone number). Each question contains an
id,nameandlabelthat is used to display the question.The responses-object consists of whatever was inserted into the application, in a pair of
idandvalue- where the id matches the id from the corresponding question.Example of the JSON bodies
Questions;
[{ "type": "textarea", "id": "550e8400-e29b-41d4-a716-446655440000", "name": "Søknadstekst", "label": "Hvorfor har du lyst til å søke denne gruppen? Skriv litt om din motivasjon." }]Responses;
{"550e8400-e29b-41d4-a716-446655440000": "Jeg har kjempelyst ..."}Resolves ABA-513