Custom .pdf reference form grid display issue

https://tuftswork.atlassian.net/browse/SLAT-20

moved from Reference Custom pdfs.docx

Slate Instance

TUH

Requestor/Reporter

Elizabeth Storrs

Date

Sep 1, 2023

Status

complete

Bug Description

cannot get reference forms to display data correctly

SUMMARY/CURRENT STATUS:

  • prompt field values do not respect form rankings. This has been submitted to Technolutions, and maybe some day we will get a response.

  • Work-around in making sure all of the data is in free text fields has been accomplished for 2023 cycle

    • Manual Query->Import of newly completed recs

    • All previous recs from this cycle that used the old form have been imported

    • All NUTR defers have had their letters ported over

    • Working on report to find anyone in GSBS who is using last year’s letters.

  • And Technolutions gets back in touch 3 months later to tell us the answer was something we already tried and it didn’t work before

  • So, revert text fields/manual process back to the way it was displaying before.

  • Completed for 2024 cycle. See the troubleshooting and conversion notes below.

RESEARCH

Custom PDF

This is what we get without a Merge query—two forms with different data, like there should be, but not custom fields.

Things that DO NOT WORK

Brackets in the custom pdf

Things that DO work

 

+

 

=

Gets me A DIFFERENT ANSWER FOR EACH FORM!

For the top of the .pdf, I used the delivered Merge fields, as well as their “new” version without the NUM:

And this gets me

 Takeaways:

  • System > Merge queries must be application scoped.

  • The old NUM nomenclature, even if you choose a system field from the drop down, does not work any more

  • System fields, whether or not they are in the “available” list, do not need to be added to the Merge query.

  • The cache has to be broken in order to refresh the data. Have an easy thing to look at to determine if the .pdf you are reviewing has actually changed, and have a group of records you can cycle through.

  • Add Prefix, first, last and title, organization just as their own fields (rather than trying to get a concatenate to work) got me to:

Which just looks awful.   Next thing to try is signature to get potentially both the student and reference names:

Reference signature is perfectly reasonable…name of the student is not.

Name is one problem. The x’es are the other, bigger problem.

Reference_int_5, reference_int_10, etcetera on the query side, mapped to those exact same values on the pdf gets me the xes, but all of the values show on every pdf:

Int_5, int_10, mapped to reference_int_5, reference_int_10 in the .pdf get me nothing in the boxes:

Redux Forms, to make sure I have the correct group for testing:

https://gradhlthsci.admissions.tufts.edu/manage/query/run?id=4a6b0c50-b722-4279-8b91-21a75c16d617&run=60f858ab-9ddf-4620-8892-0717b80c29f1

  • Abbey Sykes (2)

  • Bobomatic (3)

  • Elle Pritchard (2)

  • Lynn Kaneko (2)

10/10/2022

 


 Now, to see what happens with another set of export prompts:

The brilliant idea of doing the existence in nested subquery  still feels like the correct way to go—but, it isn’t showing anything, since the export is buried.

So, let’s flip it:

Hrm. Check again in morning in test, then decide about porting back to production.

 

11/22/2022

I believe Technolutions has been working on this, as the behavior of the Merge query has changed.

Notes from right now:

Attempting a bottom join in CJ REF MERGE ALT goes nowhere quickly.

Even adding sorts and export values in the System query does not help.

None of the custom ref fields show at all, only the system field. Confirmed this by adding signature3 (which is the faculty certification), and it appeared.


If it is the prompt values that are the issue, what if we build new fields that contain the prompt value, and these are updated by rules?

Okay, so they cannot be updated by rules, but can be updated by file imports, I believe. Testing that now.

AND OMG THEY WORK.

  Create the 6 new fields, and add them to the .pdf….

ü  Update the existence fields—do they work? Try in Test.

ü  Either way, continue the build-out in Production.

o   If they work, add the x’es and then continue

o   If they do not work, change out the background and just include the text

ü  Swap over Nutrition so they can use these forms

ü  Starting with the Reference Redux form, query on form values, and build out to the applicant and person tables to have matching criteria and make it easier to see what/who I am working with. Shared/Admin/RefReduxUpload https://gradhlthsci.admissions.tufts.edu/manage/query/query?id=899703a0-4e8a-4aec-bdf4-b7f5f64ad387

o   Making it ‘only new records since last run’ makes it super-easy to see if it needs to be run and uploaded

ü  Import in production to new Source Format called Reference Text Fields

o   Note the matching criteria at the Person, Application, and Reference level. Initially I thought reference would be enough, but it didn’t work, so I matched on everything I had, EXCEPT the form GUID itself.

o   This upload will need to be done multiple times a day during the first week of December. It seem likely that once a day otherwise is fine, although it is SO HARD to see the forms continuing to come in and not doing anything about it. (clearly, I should not be staring at the page and refreshing constantly.)

ü  Figure out how to get the 29 early submissions from real applicants into the Reader

o   Turns out, the same Upload Dataset solves this problem, too! Those folks were imported on a custom mapping, and included pulling the English Proficiency and How Long Known fields as well, since they were apparently not mapped on v.1 of the new form (oops.)

 

Other takeaways:

  • Remember to Refresh Configurable Joins Libraries

  • Remember to Force Refresh cached field values

  • The existence of “Force Process Import” when things are too slow to deal with

Work around for showing old letters:

There are some nuances: only one student brought over all four of his previous references; most students have a mix, which is complicated. The plan of showing one type of rec form rather than the other only works if all are the same! 

For old references that insist on showing in the reader with the new form:

I am first trying to change the ‘type’ of the thing, to be folio and committee letters, which should not show in the Reader.  Change type, wait for rules to execute, check later today. This got rid of the letter, but not the form.

Testing in test: if I get rid of the ‘reference submitted’ activity, what happens? Ryan Ryan Ashiqueali—changed letter to “received Copy”. What does that do? Tried “hiding’ reference in checklist—any help?

The answer across the board appears to be ‘nothing useful’. Tried to use GUIDs or Ref ID’s to prevent particular references from showing, but only application fields are available for filtering in these legacy tools.

The work-around that seems best is waiting until all of the references are present, then pull the correct ones, bundling them up, and doing a single upload of them as “Committee Letters” to the individual applications or to the folio as necessary.

In order to have committee letters show, the Application ID (from the Edit Application Details tab):

must be added to both Committee Letters (to display those) and References (to hide the current form display) in Reader Tab Materials:

·        Ashiqueali, Ryan Ryan—has three letters, but another could come in from JCastellot. He also has not submitted as of 12/13, so this one may be moot. It is currently set up, correctly, if he were to submit—it is only the JC letter that might need to be added.

·        Okoronkwo, Clement – needs 2 more letters

I am using the “People with imported old ref” query to keep track of these folks, and see if any new ones need to be added.  

The \2022_reference_uploads\People with imported old ref STATUS TRACKING.xlsx document has the status updates for everyone who was checked; 12/13 added new refs for Bootwala; still only Ryan and Clement are outstanding.


Amanda Holder (Technolutions)

Jan 30, 2023, 7:08 PM EST

Hi Elizabeth, 
 
Thank you for submitting this request and for your patience as we work through a high volume during this time. In reviewing the /apply/GSBS/reference.pdf it appears that the original reference scoped prompt based fields could have been added without the initial tag of reference which would cause the behavior you saw with the field not populating the value. In the Custom PDF KB article, it discusses that when adding a reference scoped field to the custom PDF, the initial tag of reference_  that is placed in front of the field id is required. Currently the custom reference scoped text fields have this which is why the fields are pulling data. 
 
In the test environment, we tested this by adding the original prompt based reference scoped fields and added the reference_  tag to the beginning. With that tag, the fields started to display the value for that question. The cache on the application has to be broken to get this updated as well.
 


We hope this information is helpful. We will go ahead and set this request to solved for now but if you have additional questions please let us know. 
 
Best Regards, 
Amanda & the Technolutions Team!

 State of affairs at end of 2023 cycle for both GSBS and NUTR:

I have copied this over and renamed it to be reference2023, and will work on the original reference.pdf. I believe all I need to do is revert what shows here to the old fields (with the billion boxes). The test case shows that if I follow the instructions Technolutions gave (at the end of the 2023 document), things work.

This also seems to indicate that I do not have to make any changes in order for the 2023 references to be re-used in 2024….all I have done is remove the necessity to turn the values into text.

Confirmed that data is showing correctly in the non-text version of the forms; have cleaned these up, and reference.pdf are now available across the entire instance.