Citations list differ for the same DOI when calling directly the resource and calling the listing the resources from other type.

[kjgarza]

Expected Behaviour

Current Behaviour

Citations list differ for the same DOI (e.g. 10.48321/d1h59r) when calling directly the resource and calling the listing the resources from other type.

e.g. 10.48321/d1h59r shows 3 citations

{
  dataManagementPlan(id: "10.48321/d1h59r") {
    partOf{
      nodes{
        id
      }
    }
        citations(resourceTypeId:"Dataset"){
      nodes{
        id
        type
      }
    }
  }
}

but when calling the organization with the same doi 10.48321/d1h59r one gets zero citations

{
  organization(id:"03yrm5c26") {
    name
    dataManagementPlans {
      nodes {
        id
        title: titles(first: 1) {
          title
        }
        citations {
          totalCount
          nodes {
            id: doi
            name: titles(first: 1) {
              title
            }
          }
        }
      }
    }
  }
}

Steps to Reproduce

1. step1
2. step2
3. step3
4. step4

Context (Environment)

Hypothesis

Detailed Description

Possible Implementation

[kjgarza]

counts are also wrong in the Index look at citationCount and citations.totalCount

citations.totalCount https://github.com/datacite/lupo/blob/52c6e7a/app/graphql/connections/hash_connection.rb#L58 https://github.com/datacite/lupo/blob/52c6e7a/app/graphql/types/doi_item.rb#L426

citationCount https://github.com/datacite/lupo/blob/52c6e7a/app/models/doi.rb#L1409

[kjgarza]

it seems to be all DOIs need to be updated after the citations events are created. It's also not happening automatically.

[kjgarza]

The difference in the counts is related to that there are differences between the DOI in the Database and the DOI in the Index. for example 10.1139/CJFR-2018-0211 was in the DB table but not in the index. So when one is querying for citations in graphQL the API queries the index. Event if the citations id are correctly collected the counts will be wrong if the citation DOI is not in the DOI index. This can happen when:

• (a) the citation DOI has not been imported to our database Or
• (b) the citation DOI has not been imported to our DB but indexing didn't happened correctly.

[kjgarza]

I think we should address each of the problems related to counts separately. I was thinking to create github issues but in a sense this issues already exist. For (B), there are multiple issues on indexing in which we are fixing this manually. See tickets from front. And w/r/t (A) Crossref DOIs there are multiple issues that seek to have a more complete list of DOIs in the index in https://github.com/datacite/datacite/issues/1082, https://github.com/datacite/datacite/issues/1090 and https://github.com/datacite/datacite/issues/1091.

I would suggest to keep this issue open to track progress on the other tasks. What do you think @mfenner ?

[mfenner]

I suggest to create a new issue to track these various issues and make it an epic. This issue is about a specific problem, differences in citations shown depending on how the dataManagementPlan is received.

Epic created at https://github.com/datacite/lupo/issues/649.

[mfenner]

This issue can no longer be reproduced.

[kjgarza]

This is still open, one can see the counts still do not match

{
  dataManagementPlan(id:"10.48321/d1h59r") {
    citationCount
        citations(resourceTypeId:"Dataset"){
      nodes{
        id
        type
      }
    }
  }
}

{
  works(query: "10.48321/d1h59r") {
    nodes{
      citationCount
      citations{
        totalCount
        nodes{
          id
        }
      }
    }
  }
}

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[digitaldogsbody]

Not stale