collections as agents #5327
Labels
Help wanted
I have a question on how to use Arctos
Priority - Wildfire Potential
ignore this at everyone's peril, may smolder for now ...
Milestone
Comments
dustymc
added
Function-ExternalLinks
Help wanted
|
I see a dedicated addresstype that only Arctos can access and use upon collection creation as a necessary part of this equation. I hesitate to mention (because it's not the answer to everything) but this is a case where an external resolver like DOI would be appropriate for each collection -- so you can have your I'm still a lot unclear how the GBIF GUM handles these collection agents if not just as Institution_cde? |
https://github.com/gbif/model-material/blob/master/schema.sql "collection_id TEXT PRIMARY KEY REFERENCES agent"==="these are agents with weird metadata for some reason" |
|
ahh, so is there anything we should be storing from GRSciColl? Maybe that's our key too? yea and it's too early for me to look at sql... (MOUs or bust for me) |
That still seems like me refusing to call you Michelle and insisting I mint and use my own identifier for you (hi Ugh. |
|
This sounds great. I guess I'm not sure why we can't just use
https://arctos.database.museum/collection/MVZ:Herp , or put both that and
https://arctos.database.museum/agent/21334675 in the agent page? We could
add these as alternate agent names or addresses?
…On Fri, Dec 2, 2022 at 9:37 AM dustymc ***@***.***> wrote:
* [EXTERNAL]*
GBIF's GUM model treats collections as agents (yay, I think...) and the
idea has come up a few times locally (eg for printing loan forms). There's
no decent connection between those things in Arctos, and there needs to be
if we're going to make hard links between them.
https://arctos.database.museum/agent/21334582 (and it alone) includes the
collection identifier, but as a url. I can PROBABLY make the connection
with that, but there's lots of potential for confusion (eg, people might
use the same identifier for their employer). I think we need a new focused address
type
<https://arctos.database.museum/info/ctDocumentation.cfm?table=ctaddress_type>.
For just my immediate needs that could be "collectionID" and constrained to
starting with https://arctos.database.museum/collection/, but that also
seems somewhat likely to lead to an unmanageable proliferation of
identifiers. Is there any reason to make that a bit more general, or
somehow shaped differently, or ?????
I was going to ask @lkvoong <https://github.com/lkvoong> to create these
as part of collection setup, but I think many (maybe most) will already
exist, but not in forms that will be easy to detect, so I think this, if we
adopt it, will need to be a part of come manual setup process. Eg
https://arctos.database.museum/agent/21334675 is probably
https://arctos.database.museum/collection/MVZ:Herp but I don't see
anything at all that would let machines make that connection.
Help!
—
Reply to this email directly, view it on GitHub
<#5327>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADQ7JBESDOPXUY3DOF4NJATWLIQWFANCNFSM6AAAAAASSC6FBQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
And yes, I agree with adding whatever GRSciColl assigns as an alternate id or address. |
It is currently part of the new collection creation process, but I will check to make sure it is formal and includes the addition of whatever identifier we decide upon. I think using the GRSciColl ID makes sense - MVZ:herp = https://registry.gbif.org/collection/dc806805-f05c-4a09-9d58-78743a9e8cd6 |
|
Also, currently, iDigBio gets their collection info from GRSciColl. Baby steps |
I just need to get from Arctos Agents to Arctos Collections..... |
|
There are some Arctos collections that are NOT in GRSciColl - that is part of what I plan to remedy with them and this reconciliation. |
|
I probably should have brought this up at the agents meeting, but I didn't. I need these data. I don't want our intro to the GUM to involve a big gap on the thing that EVERYTHING else touches, so I don't really want to fake it. I don't want to create a few hundred duplicates, at at least some of these exist. I still don't have the slightest inkling of a clue how this is related to GRSciColl (and I'm hoping to be pleasantly surprised?). I guess I'll create address_type |
dustymc
added
the
Priority - Wildfire Potential
|
I just started working on this (creating agents for all collections) so I can work it up, wipe out dupes, and add the new identifier. Can I have tomorrow to get it done? |
|
That would be fabulous and way faster than I had any hope of doing anything. I made collectionID and will try to get it trigger-protected before anything interesting happens. |
Because there can be multiple kinds of IDs, should we make this one specific (GRSciColl ID), or is identifying it in the remark good enough? Eventually, this is going to get caught up in #5172. While I am at this, I could very easily add GBIF Publisher ID and Arctos ID |
|
Sounds like a good idea.
…On Thu, Dec 15, 2022, 8:10 AM Teresa Mayfield-Meyer < ***@***.***> wrote:
* [EXTERNAL]*
I made collectionID
Because there can be multiple kinds of IDs, should we make this one
specific (GRSciColl ID), or is identifying it in the remark good enough?
Eventually, this is going to get caught up in #5172
<#5172>.
While I am at this, I could very easily add GBIF Publisher ID and Arctos ID
—
Reply to this email directly, view it on GitHub
<#5327 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADQ7JBFRW7QCGGOZBKUGVADWNMYF3ANCNFSM6AAAAAASSC6FBQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
Yes.
Now I'm lost again. Can you type real slow, use small words, and explain to me what that has to do with anything and why it seems to be effectively blocking this? If I'm somehow being unclear - well, see the documentation, I need things like https://arctos.database.museum/collection/MVZ:Herp, because I need to join to that and pull some data into an agent-object.
Address should work for that.
And we're lost again.... |
That's the Arctos ID
Every collection has lots of identifiers. Some are aka (like MVZ:Mamm) and now we are adding "collection ID" - which is really the Arctos collection ID as an address. But this particular collection has a couple of other known identifiers: Museum of Vertebrate Zoology Mammal Collection My point is, I can add what you need, but I can also add the other while I'm in there. Should these bottom two also be collectionID with the appropriate remark or should I have new address types for them. I see from the definition that you are expecting an Arctos url, so I guess that answers my question - BUT I feel like this could use a little more thought and discussion because as we are constantly putting out "wildfires", I fear we just create future messes. @ArctosDB/arctos-code-table-administrators |
I dropped what I was doing and added the promised constraint so they can't be.
As far as I know those don't do anything so
Not sure how, there are a bunch of reasons (this, pulling addresses to print loans, etc.) to make the connection and a fairly-obvious way to do so. If the mechanism is wrong then it's still functional and we can use it to do whatever needs done. I'll be fairly surprised if this turns out to be some sort of problem. If you're talking about the other stuff, I'm still not sure why that got drug in here. |
|
I think I have added agents with a collectionID for all collections. BUT I am pretty sure some are missing because they have 0 records. @dustymc please write some SQL that will get me a list of all agents with a collectionID, then I can see if any are missing. |
|
|
PSU:Mamm see https://github.com/ArctosDB/internal/issues/209 ALMNH:ES all redirected to ALMNH:Paleo how do you recommend we handle this? Add both collectionIDs to one agent? Create the second agent and relate them somehow? DGR - I didn't bother creating these because they need to go? Otherwise, I think it's done. |
That would break my scripts/expectations. I think ignoring the orphans and GP weirdness is fine, at least for now. Thanks! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
GBIF's GUM model treats collections as agents (yay, I think...) and the idea has come up a few times locally (eg for printing loan forms). There's no decent connection between those things in Arctos, and there needs to be if we're going to make hard links between them.
https://arctos.database.museum/agent/21334582 (and it alone) includes the collection identifier, but as a url. I can PROBABLY make the connection with that, but there's lots of potential for confusion (eg, people might use the same identifier for their employer). I think we need a new focused address type. For just my immediate needs that could be "collectionID" and constrained to starting with
https://arctos.database.museum/collection/, but that also seems somewhat likely to lead to an unmanageable proliferation of identifiers. Is there any reason to make that a bit more general, or somehow shaped differently, or ?????I was going to ask @lkvoong to create these as part of collection setup, but I think many (maybe most) will already exist, but not in forms that will be easy to detect, so I think this, if we adopt it, will need to be a part of come manual setup process. Eg https://arctos.database.museum/agent/21334675 is probably https://arctos.database.museum/collection/MVZ:Herp but I don't see anything at all that would let machines make that connection.
Help!
The text was updated successfully, but these errors were encountered: