ucd-library / rp-ucd-harvest

This container will be responsible for harvesting UC Davis IAM, and CDL publication data for the research profiles
MIT License
0 stars 0 forks source link

Verify we can replicate the Grants feed for CDL #60

Open qjhart opened 2 years ago

qjhart commented 2 years ago

The next step in the Grants administration is to verify that we can get the Grants feed replicated for the CDL data. I think the idea is that we use the same methodology as with MIV, and hopefully be able to craft an appropriate feed directly from the DB.

CDL sent a note on 2021/12/13, with the following:

I hope this message finds you well. Please find in this folder several items related to the grants feed: Drive Folder

grants_metadata-example.csv: An example of two grants from RGPO. "id" is the internal identifier, while "funder-reference" is the grant identifier given by the funding agency. They do not necessarily need to match.

grants_metadata_template-FULL.csv: The full plant grants metadata template. (RGPO only uses a portion of the available fields.)

hr_feed-example.csv: Two users who appear in these examples. UCPathIDs are used as the proprietary ID.

grant_user_links-example.csv: Used to link the researcher to the grant. "id-1" is the grant's internal identifier and should match against an "id" in the grants metadata table/file. "id-2" should match against a "ProprietaryID" of an existing user in the system.

Symplectic Support - Grants module underlying field definitions.pdf: The Symplectic Support help document on the grants module.

qjhart commented 2 years ago

I've created the initial feed using code from this gist and saved the output to this spreadsheet

Waiting on Alainna now

qjhart commented 2 years ago

Here are some notes from Alainna's email:

_Regarding your updates to the grants_links sheet, I'm not sure about the categories there.  What I did for that, was to look at the table in QA [dbo].[Relationship Type]. Those link ids don't seem to be attached to a particular grant type, but to all grants. My suspicions is that that category roles might not even be used.

On Wed, Feb 2, 2022 at 3:25 PM UC Open Access Policies oapolicy-help@universityofcalifornia.edu wrote: Hello Quinn, Thanks so much for sending this. And good timing, as I'm working on the RGPO grants file as well right now.  I'll ping the Symplectic team about the process for adding this to QA for testing purposes and whether the data can be removed (and how). 

Some initial thoughts after looking at the file (my recommended edits: https://docs.google.com/spreadsheets/d/1V4Tiv7g0CvfAc5bWPrHjcXYDJ0q9xJtAlgMdBZ2FJ3o/edit?usp=sharing ) : 

  1. I wonder if I can grant access to see the grants module settings? I've tried adding you and Vessela as admins for grants module on QA. Are you able to access https://qa-oapolicy.universityofcalifornia.edu/objecttypes.html?cid=2 ?_ 

Yes, I can see this,and I can edit the settings as well.   I can also potentially edit the overall grants module?  Not sure, since I didn't try to save.  I see a setting there, which seems to be system/ grant-wide, and that's the ability for users to adjust individual privacy levels.  We are expecting c-davis grant users to be able to do that.

image

  1. type: This should be a "grant type" in the system. There are the current grant types:grant: Defaultc-fund: Used for LBL funding records c-grant et al (c-grant-cloned, c-grant-cloned-cloned): RGPO awards. We're planning to condense these to a single c-grant group in the latter part of this year.c-davis: Just created for the Davis feed (https://qa-oapolicy.universityofcalifornia.edu/objecttype.html?cid=2&tid=74 ) In the attached edit, I've updated the "type" to "c-davis".   To retain the data already in the column, I'd recommend instead that we instead use a different field within the c-davis type: funding_type. I've added the field to the c-davis type with all the specified types, and set it as a new column in the file. (https://qa-oapolicy.universityofcalifornia.edu/objecttypefieldusage.html?cid=2&tid=74&fid=57 ) This seems reasonable to me.

  2. grants_feed: id: There may be a requirement for character length on this; plus, there's apparently a possible URL field for grants! Therefore I'd recommend the following:       a. using "davis-[identifier]" for "id", e.g. "http://experts.ucdavis.edu/grant/115166" -> "davis-115166"       b. Adding a "url" column with e.g. "http://experts.ucdavis.edu/grant/115166

This also sounds fine.  It looks like the URL is only part of the discovery module, so we'll see if users get to see it anywhere. 

  1. grants_links: id_2: I'm not 100% certain whether the email field can be used as the linking identifier, but if it can, this would save us so much work. (We've been using proprietary IDs/UCPathIDs).  Yeah, it seems like we should  be able to use this, it's actually more strange that a ID not in the system could be used.  I'll wait and see what we hear about that.