Periodically in September and October 2023, admissions staff have found that applicants in PTCAS who are verified and should be complete do not have Slate records.
Upon investigation, the issue is that no “in progress” subscription file was sent via the API for these students, which is where the application record is created in Slate. It appears that the in progress files should have been sent at the beginning of the cycle (August 2023) but were not.
My current hypothesis is that when turning on the subscriptions for the cycle and executing replays somehow these files were missed--either due to an error when running the replay, due to an issue with Liaison and the CAS database, or possibly because when the replay was run these applicants had “deselected” our program and so were not “in progress” at the time of the replay, but they later “reselected” our program which did not send another file?
Description of Issue:
10/19/2023 the following list was provided by the DPT admissions team of records that they identified as verified in the DPT-BOS instance of PTCAS but which had no Slate application record.
Applicant Name | CAS ID | Designation Submitted Date |
|---|---|---|
Cooper Ernst | 6909309498 | 6/17/2023 |
Katharine Manning | 6544491980 | 8/29/2023 |
Nicole Marshall | 2895233819 | 6/19/2023 |
Chloe McGonagle | 6163363961 | 6/25/2023 |
Eleise Trias | 8811883669 | 6/16/2023 |
Isabelle Volney | 1641478338 | 6/15/2023 |
Troubleshooting:
I created the following query to identify person records which were in the CAS - Application source format but which did not have applications to any of the DPT programs. It generates a row for each person record and indicates what program they have applied to by checking the source file name for the CAS Program ID for each DPT program.
https://tusmgp.admissions.tufts.edu/manage/query/query?id=6dc671d3-c44c-437e-99f0-c0baf163feed
There were these six students from Boston (above), plus two additional, and some of them had applications for more than one program.
Person Reference ID | Person Name | Person CAS/Liaison Person ID | DPT-BOS | DPT-PHX | DPT-SEA |
894097973 | Manning, Katharine Marie | 6544491980 | Yes | Yes | |
847434425 | Ernst, Cooper | 6909309498 | Yes | ||
655854718 | Trias, Eleise | 8811883669 | Yes | Yes | Yes |
311366551 | Davydov, Daniel | 6290569725 | Yes | ||
668561708 | Xiong, Jose Vang | 9540655614 | Yes | ||
152404006 | Marshall, Nicole | 2895233819 | Yes | ||
296453684 | McGonagle, Chloe Lynne | 6163363961 | Yes | ||
664761843 | Volney, Isabelle Lynn | 1641478338 | Yes |
Short-Term Fix:
To catch when a student is missing an in progress application file but later submits their application in the CAS system, the DPT admissions team uses this query:
https://tusmgp.admissions.tufts.edu/manage/query/query?id=6dc671d3-c44c-437e-99f0-c0baf163feed
If a record is found, the following steps can be taken to resolve this issue for this student:
Go to the student record in Slate, Timeline > Interactions, then filter to “Source.” Review the source files to see what files have been sent by the API. Review the file name to determine which program(s) are sending data (349019 is BOS, 349022 is SEA, and 349027 is PHX):
2a. If only one program is sending data, then skip to step 3.
2b. If more than one program is sending data, or if there are applications in the 2024 round for DPT already, review the information in the Application Details (PTCAS) tab in the application record to ensure that the program ID and application ID use the correct code to correlate to the correct program.
If they do, go to step 3. If they don’t, then edit the application record to change the program ID and application ID to use the correct program code.
Use the instructions for obtaining a single application file from Postman to get a copy of the application you are missing from the correct CAS environment. This will be a file containing all the “live” application data, but we will treat it as if it was the file sent by the API. Save the file as a CSV using the following naming convention, saving one as the IP File and one as the App file, and replacing the information between <> with the correct ID from the record.
Program | IP File Name | App File Name |
|---|---|---|
BOS | PTCAS-app-started-adhoc-6861_12748_349019_<applicationID><CASapplicantid>.csv | PTCAS-app-s-adhoc-6861_12748_349019_<applicationID><CASapplicantid>.csv |
PHX | PTCAS-app-started-adhoc-6861_14347_349027_<applicationID><CASapplicantid>.csv | PTCAS-app-s-adhoc-6861_14347_349027_<applicationID><CASapplicantid>.csv |
SEA | PTCAS-app-started-adhoc-6861_12748_349022_<applicationID><CASapplicantid>.csv | PTCAS-app-s-adhoc-6861_12748_349022_<applicationID><CASapplicantid>.csv |
Once you have the files saved, go to Slate > Database > Source Formats > PTCAS In Progress (…) > Upload File(s) and upload the IP file through the source format. After uploading, go to Database > Force Process Import and run to import the data immediately.
Go back to Source Formats > PTCAS Application (…) > Upload Files and upload the App file through this source format. (This process is used instead of doing a retroactive refresh against the source that was sent by the API because when an application has more than one application record, the retroactive refresh has caused issues, it seems better to just do a direct upload through the full source format than to do retroactive refreshes).
Now rules will run against the record, and overnight any documents pending import that had been sent by the API will append to the record/application.
To fix the issue in the short term, I put all the application files through the API In Progress source format to create the applications, and then I refreshed the existing files to update them to their current status of verified. Overnight Slate should run the rules and match all the documents, so tomorrow morning these should be in good shape for you to finish them up and get them ready for review.
On 10/24/2023 I was notified that two students had been updated to Verified in Slate, but were otherwise not complete and ready for import. Upon investigation, the two students were the two from this fix who had more than one application in this round. It seems like what happened is that the “retroactive refresh” of their data files in the PTCAS Application (All CAS by Liaison - Applications) source format did not behave exactly as if the files had been freshly added. Instead of matching on the CAS/Application ID and updating the correct program application, doing the retroactive refresh reverted to the matching behavior where only the Rank 1 application was used to match, and so each subsequent refresh updated the DPT-BOS application for these two applicants and did not match on and update the DPT-PHX or DPT-SEA applications. In test, I downloaded the original file for DPT-PHX and DPT-SEA and put it through the Source Format as a new file (instead of a refresh) and this behaved as expected. So there is something about “retroactive refresh” that doesn’t match application records exactly (and it doesn’t seem to be just that I did all files at the same time, as I did test a retroactive refresh on only one file and it still matched incorrectly).
11/15/2023
✅ Benson, Bailey 2283360939 14400572 Looks like same issue with no IP file
✅ Nekrasova, Anastasia 6381531350 app ID: 14328303Looks like same issue with no IP file
✅ Moen, Kailey 988176475 5485667558 - has verified BOS app, is in Slate with only PHX and SEA app 14477459 Looks like had IP file for SEA and PHX, but not BOS. App file was sent for BOS on 11/8, but didn’t go anywhere.
Rocha-Heijenga, Sylvia Gia 539972316 - verified BOS app, in Slate as awaiting submission in BOS 14386948
Long-Term Fix:
I need to determine if we are not getting certain in progress files, and if so why. Or if the issue is that there were some in progress files missed during the API setup process, which is why we do not have them now. If it is the latter, then an audit of the in progress files in Slate vs. PTCAS should show in progress applicants in PTCAS that are not in Slate.
Export a list from Slate of all in progress applicants for the 2024 cycle, including CAS ID for matching.
Export a list from PTCAS of all in progress applicants for the 2024 cycle.
Compare the two lists to determine if there are applicant records in PTCAS for in progress that are not currently in Slate.
In Progress WebAdMIT vs. Slate
11/15/2023: I ran a spreadsheet out of WebAdMIT for each instance, and then compared the CAS IDs from the WA list of in progress to what applications existed in Slate. For each instance, these were the results:
BOS: 25 in progress apps in WA that were not in Slate. Dates of designation added ranged from 6/15 - 8/25, with the most occuring in between 6/15 - 6/28.
PHX: 14 in progress apps in WA that were not in Slate, dates ranged from 6/15 - 8/25.
SEA: 12 in progress apps in WA that were not in Slate. Dates ranged from 6/15 - 8/18.
Spreadsheet compiling all missing in progress apps across programs:
PTCAS-app-started_<instanceId><organizationId><programId><applicationId><casApplicantId>.csv
BOS: PTCAS-app-started_6861_12748_349019_
Application Matching Issues
[3:58 PM] Williams, Helen
So I have two source formats -- one for in progress files and one for application files. In progress consumes the initial application data file, and includes Round: Always Create. The application source format does not include Round: Always Create because it is supposed to just match and it will get at least 3 data files for every application and therefore it cannot be creating a new application every time.
[4:01 PM] Williams, Helen
This student had an application file come through the IP source format for BOS on 11/15. That created the BOS application, with a BOS app ID. On 11/26, the same student had an application data file come through when she submitted to SEA, so it went through the app source format. The app source format couldn't match on the ID, but also isn't set to round always create, so it overwrote the app ID with the SEA ID (but kept the program = BOS). Then, when I tried to upload the SEA app through the IP source format, it matched to the SEA ID (which should've been the BOS ID, but wasn't) and overwrote the data again.
[4:02 PM] Williams, Helen
I should've looked at the app ID and not just the program field, but also I think this part (having a submitted app to one program without an IP app, but an IP app for a different program) is the final mystery of WTF has been going on with these applications
Add Comment