google scholar and twitter metadata not appearing

[davidschober]

Google Scholar/ Highwire style citation_ metadata fields in the header are not visible in arch. This should be on by default in a sufia instance. It's possible one of our views bonked it.

To Recreate:

• Go to a work https://arch.library.northwestern.edu/concern/generic_works/q811kj631
• View source note the twitter card and gs metadata is not available.

Example:

<!-- Google Scholar metadata -->


<title>Gwendolyb Brooks</title>

This looks like should be populated by:

https://github.com/projecthydra/sufia/blob/818cb4ca53d81fdb37835a0cd7c753d66eacd679/app/views/layouts/_head_tag_content.html.erb#L8-L11

https://github.com/projecthydra/sufia/blob/818cb4ca53d81fdb37835a0cd7c753d66eacd679/app/views/shared/_citations.html.erb

Note, since we're upgrading to 7.3, can we grab this #160

[kdid]

@cpd3149 - question:

So, we notice the same behavior in 7.2 as 7.3 where the Google scholar metadata does show up on the File, for example: https://arch.library.northwestern.edu/concern/parent/sj1391960/file_sets/6682x3948

But does not on the Work, for example: https://arch.library.northwestern.edu/concern/generic_works/sj1391960

We just wanted to clarify that this is not expected behavior.

Thanks.

[chrisdaaz]

@kdid -- good catch.

i'll have to investigate some more, but it seems strange that the citation_title is pulling from the file name rather than the name of the work.

[davidschober]

That's odd I thought I looked at the files themselves as well.

A couple of things:

I think it the pattern may have gotten a bit messy and may take some upstream conversations. Google Scholar expects a link to the full pdf, but a work can have multiple pdfs associated with it. It seems that if we're going to assume a journal article is a work (as opposed to say a journal issue being a work) that the work should have the citation information.

A few other things:

citation_title should be the title of the journal article, not the PDF. (Note A in the official guidelines)

citation_author should be the authors of the paper and allow for multiples. (Note B. in the official guidelines)

official guidelines

I thought Highwire themselves had docs for this stuff, but I'm not finding it.

Does someone want to kick off an upstream conversation? I'm thinking this may be unfinished work.

[kdid]

@davidschober -

After looking into this a little further, it seems like it's working as intended. (We didn't override something and break it)

(Note that the title of a file defaults to the filename but that title of the file is editable. If you do, that is what is used for citation_title.)

Since we're not modifying the defaults and we're using a Generic Work for any type of content anyone uploads we should have some further discussions about this.

[chrisdaaz]

EDIT: not sure the path matters. This Deep Blue Data example shows GS metadata, but pulls the filename for citation_title and the depositor name for citation_author instead of Creator.

i'd like to discuss the /concerns/generic_works/ structure vs. /files/

/files/ is used at PSU and DigitalHub. This PSU journal article is correctly pulling 'citation_title' from the user-entered Title field: https://scholarsphere.psu.edu/files/6682x4006

this is also the case for data: https://scholarsphere.psu.edu/files/cc08hf80f
and dissertations: https://scholarsphere.psu.edu/files/sf268g73h

i think most deposits will be 1 file at a time (e.g. journal articles, dissertations, data), and the edge cases will be multiple files (e.g books).

[davidschober]

Thanks @kdid. Let's mark this for discussion as it seems like there's no work to be done at this point.

Could you tell if citation_author is in fact pulling from depositor? That seems off.

[kdid]

Yes, holding a discussion about this would be a good idea.

@cpd3149 - just want to note that ScholarSphere has not yet migrated to Fedora4/Sufia7/PCDM and when they do their data structure will look more like ours. Peeking at their GitHub issues gives an idea. i.e. "Files" will become "GenericWorks" + "FileSets".

https://github.com/psu-stewardship/scholarsphere/wiki/Fedora-4-to-Sufia-7-%28PCDM%29-Migration-Information

@davidschober - it does look like the creator in the metadata on the fileSet (but not the solr doc) is set to the user_key. This appears to be the same in Hyrax. Let me do a little more digging on why that might be so.

[kdid]

Just adding because these issues are both relevant to track how this plays out for Hyrax

samvera/hyrax#110

samvera/hyrax#109

[davidschober]

Nice catch! Anyone on the team can feel free to jump in and push changes upstream. Volunteers?

…

On Apr 3, 2017, at 12:12 PM, kdid ***@***.***> wrote: Just adding because these issues are both relevant to track how this plays out for Hyrax samvera/hyrax#110 <samvera/hyrax#110> samvera/hyrax#109 <samvera/hyrax#109> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#161 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AElKnu-EaWgISngEXoJhpNvOXTtvhHF0ks5rsShkgaJpZM4MoyHk>.

[chrisdaaz]

Looks like this was resolved upstream: samvera/hyrax#1324

[davidschober]

Cool. When they cut it let's grab it.

…

Sent from my iPhone

On Jul 10, 2017, at 8:34 AM, Chris Diaz ***@***.***> wrote: Looks like this was resolved upstream: samvera/hyrax#1324 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[davidschober]

@csyversen This was resolved upstream, and we've deployed the latest version of hyrax, but it doesn't seem to be showing twitter/google scholar info on works. Can you check it out.

[davidschober]

Investigate we can back port.

[chrisdaaz]

fixed in hyrax 2.x