Dev david issue 723 #771

davidszkiba · 2024-12-06T08:25:10Z

Closes #723

This is a stub for the webservice that accepts response data from a client.

For testing, I created a site at URL /local/catquiz/client/submit_responses.php where submitting data can be easily tested. There are now basic plugin settings: - central_host: the address of the central host - central_token: the token used to access the central host - central_scale: The scale that should be synced. For now, this is a basic input field that expects an integer, but this should be changed in the future. There are database tables to store information about send requests and mappings of questions. There is now also a question mapper that calculates a hash from question properties - this way we will be able to map questions based on their properties (text, answers, etc).

This changes the data to be returned and also creates a simple page fetch_parameters.php for testing during development.

Changes local_catquiz_question_hashmap to local_catquiz_qhashmap

This adds a "Sync" button to the scale detail view. It will sync all parameters associated with that scale or any of its subscales from the central instance.

- Only create a new contextid if parameters were synced - Carry-over old itemparameters to the new context and include them in the decision of which model should be the active one. E.g., if we get 3 new models for questionid 1 but already have a different model locally with a higher status, this one should still be set as the active one.

This sets the lowercase name as label. Since we already have a validation step that checks if the name is unique, it should be save to use the lowercase name as label.

Only set the scale label when it is first created.

In a future commit we could remove the rest of the code that refers to users. Instead, we can just refer to attempts instead.

We always import the whole set of item parameters from the central instance, as long as at least one item parameter has changed. Also, we duplicate the items in local_catquiz_items under certain conditions: - If the existing items (and context) are in use by at least one test, we create new items with the synced params. - Otherwise, we update the existing items and change their contextid to the new context of the synced params.

This can be removed later, it just makes testing easier.

The question hashmap alone can't be unique.

If there is no active main/root scale for a given contextid, skip calculations.

…ponses

This reverts commit 254362f.

Just pass the responses data directly. This way, we do not need to access the DB in the test

The matrix of responses is not correctly pruned to filter out items with all-correct or all-incorrect responses. The same is done to filter out persons or attempt with all-correct or all-incorrect answers. This means the expected item parameters in the characterization test had to be updated, because we previously didn't prune iteratively.

- Changed the string of the save button for remote settings config: its more specific AND behat checks for the default save string when doing another test and having two buttons with that string on the same page made it fail. - Remove php unit `resetAfterTest()` after converting a unit test to be a `basic_testcase` - Pass some person params as variables to fix pass-by-reference errors

eynimeni · 2024-12-09T09:50:10Z

actions succeed only for php 7.4 @davidszkiba

This adds checkboxes so that users can opt-in to either share data with a central hub or to make the instance act as a central hub. Also, it removes the obsolete `sync_scales` config that was used by just a page to test submission of items.

davidszkiba added 30 commits December 6, 2024 09:24

GH-723 Add stub for webservice to receive shared data

feb9843

This is a stub for the webservice that accepts response data from a client.

GH-723 Update logic to store responses in central instance

ffda991

GH-723 Update response_handler format

88118d9

GH-723 Tranfer mapping

f688aca

GH-723 Update service to return correct data

e15c6fb

This changes the data to be returned and also creates a simple page fetch_parameters.php for testing during development.

GH-723 Match scales via label

932ef72

GH-723 Rename too long DB table

2b95678

Changes local_catquiz_question_hashmap to local_catquiz_qhashmap

Fix access to missing var

3765583

GH-723 Add UI to fetch parameters

af79e0e

This adds a "Sync" button to the scale detail view. It will sync all parameters associated with that scale or any of its subscales from the central instance.

WIP: refactor model_strategy::get_params_from_db()

fbf36d2

Update handling of items table

1bacb9a

Update code for switching context

a0adf0f

Temporariliy prevent actions to run

42a20bf

Fix if no old context exists

743412b

Fix copying old item param

cedc06f

GH-723 UI: wider modal and fixed language string

edc4190

GH-723 Cleanup: remove class used just for testing

89af423

GH-723 Set a label when creating a new CAT scale

4767f06

This sets the lowercase name as label. Since we already have a validation step that checks if the name is unique, it should be save to use the lowercase name as label.

WIP

c0d1fe8

Fix setting scale label

0c096d4

Only set the scale label when it is first created.

GH-723 Fix errors due to model_responses change to use attempt IDs

6f666b5

In a future commit we could remove the rest of the code that refers to users. Instead, we can just refer to attempts instead.

GH-723 Fix actions errors (PHPUnit and linting)

c1723b9

Undo workflow file change

8a86bd6

Fix label validation for saving CAT scales

6832915

GH-723 Rename parameter for scale to sync

fd21c00

GH-723 Linting fixes

710088d

GH-723 Prepare response data to send to central instance

73c641b

davidszkiba added 23 commits December 6, 2024 09:24

GH-723 Add simple page to test param calculation

66d595d

This can be removed later, it just makes testing easier.

GH-723 Prune model_responses before doing param calculations

9330dcb

GH-723 Fix index for rresponses table

1b656ac

The question hashmap alone can't be unique.

GH-723 Save calculated responses to new context

49701e0

Fix: prevent error in recalculate params adhoc task

f02e1dd

If there is no active main/root scale for a given contextid, skip calculations.

GH-723 Add services to trigger param calculations based on remote res…

97e4d2f

…ponses

WIP

ef54370

GH-723 Combine webservices related to param calculation

31695bb

GH-723 Fix webservice config

3e57c52

GH-723 Update webservice UI

19e2dd4

GH-723 Render certain buttons only on root scales

ee87e96

GH-723 Update UI

7f5ad58

GH-723 Add remote settings config to Calculations tab

f000c34

Fix missing check if property exists

e70ed16

Fix inconsistent non-unique index for rresponses

4354175

Remove file used for testing

daeebfd

Linting fixes

16ed13c

Add JS files from older VM

5ee76ad

Revert "GH-723 Add mapping of attempts to users in model_responses"

cf3d15d

This reverts commit 254362f.

Fix model_responses unit test

bdff8af

Just pass the responses data directly. This way, we do not need to access the DB in the test

GH-723 Fix lint and test errors

69c0701

davidszkiba self-assigned this Dec 6, 2024

davidszkiba added 3 commits December 6, 2024 09:32

Codestyle

aa00a68

Update lang string

bda73ef

Codestyle

39b9fb9

davidszkiba requested a review from eynimeni December 6, 2024 08:47

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dev david issue 723 #771

Dev david issue 723 #771

davidszkiba commented Dec 6, 2024

eynimeni commented Dec 9, 2024

Dev david issue 723 #771

Are you sure you want to change the base?

Dev david issue 723 #771

Conversation

davidszkiba commented Dec 6, 2024

eynimeni commented Dec 9, 2024