Convert DOI URLs in related_publications to related resources
#1417
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #1417 +/- ##
==========================================
+ Coverage 88.54% 88.61% +0.06%
==========================================
Files 77 77
Lines 10537 10554 +17
==========================================
+ Hits 9330 9352 +22
+ Misses 1207 1202 -5
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
|
LGTM! |
|
@bendichter Then add an approving review for your review request. |
|
It's currently marked as WIP |
|
@bendichter How about now? |
|
@yarikoptic Ping. |
| related.append( | ||
| models.Resource( | ||
| identifier=v, | ||
| relation=models.RelationType.IsDescribedBy, |
There was a problem hiding this comment.
I think this one is is too specific to assume that any reference is the description of the data we find in the file.
What we can say only that the file references that publication, and hence I would better go with
| relation=models.RelationType.IsDescribedBy, | |
| relation=models.RelationType.References, |
as the default. WDYT @bendichter @satra ?
There was a problem hiding this comment.
I see your point that we cannot necessarily infer the relationship between the data and the paper.
The DataCite definition for "References" is "indicates B is used as a source of information for A." Here, this would mean that the paper is used as a source of information for the Dandiset. I don't know if this really fits here.
I have been wondering what the best way to associate papers with datasets using these relations. What are the different types of ways a paper and a dataset can be associated? So far, I have been using isDescribedBy (indicates that A describes B) for everything, assuming that the publication is the primary publication in which the data is introduced and described. Another type of relationship might be that a paper reuses a dataset. In this case, IsSupplementTo (indicates that A is a supplement to B ) might be a better choice, since it applies to both. Are there other relationship types that describe how a paper and dandiset might be related? What are the different use-cases here?
This is challenging in part because there is currently no aspect of the dandi schema that allows us to describe what the resource is (see issue here), so that needs to be inferred at least somewhat by the relation type.
There was a problem hiding this comment.
@bendichter Now that dandi/dandi-schema#231 has been resolved, how can we move this PR forwards?
There was a problem hiding this comment.
If I got it right we still have not made our minds on the default type of the relationship here... The most logical is that if we do not know relationship - do not fill it in. But ATM our model seems to not allow relation to not be present . Should we change that in the model? or you still feel comfortable @bendichter in using IsDescribedBy here ?
There was a problem hiding this comment.
Yes, exactly. My preference would be to change to schema to make this optional. Now that we have a resource type I am much less reliant on this particular field, though I do appreciate the caveat that optional fields are very rarely populated. I would also be happy with defaulting to IsDescribedBy or IsSupplementTo.
There was a problem hiding this comment.
ok, to expedite, let's proceed with IsDescribedBy and see where it takes us ;-)
|
🚀 PR was released in |
Closes #1308.