Joss: Google Scholar crawling

[genomematt]

It does not at the moment look like JOSS is being indexed by google scholar.
I think this is one of the sites we want to ensure visibility for the journal on.

https://scholar.google.com.au/intl/en/scholar/inclusion.html#troubleshooting

[arfon]

JOSS papers have included the Google Scholar metadata tags since this update about 10 days ago.

Their docs suggest it can take 4-6 weeks to be indexed so I think we're still in that window. Let's leave this open for now to keep tracking this issue. I agree we must be indexed by Google Scholar, I'm just not sure there's anything we need to do other than wait at this point 🕐

[pjotrp]

Excellent!

[sherrillmix]

At least a couple papers show up now:
https://scholar.google.com/scholar?hl=en&q=GeneNetwork%3A+framework+for+web-based+genetics
https://scholar.google.com/scholar?hl=en&q=R3D2%3A+Relativistic+Reactive+Riemann+problem+solver+for+Deflagrations+and+Detonations

However other articles on the same page currently do not:
https://scholar.google.com/scholar?q=Xenomapper%3A+Mapping+reads+in+a+mixed+species+context
https://scholar.google.com/scholar?q=pyuca%3A+a+Python+implementation+of+the+Unicode+Collation+Algorithm

Pretty cool that some are showing anyway and maybe google's getting around to processing the others.

[arfon]

@sherrillmix, thanks for finding these. Let's keep monitoring this.

[FaustinCarter]

It looks like Google Scholar is not actually indexing anything from JOSS. The two examples listed by @sherrillmix only show up in Google Scholar because the first is also listed at researchgate.net and the second is listed at eprints.soton.ac.uk. Google Scholar is indexing them based off those domains, not because of anything going on at JOSS.

[cMadan]

Maybe Google Scholar isn't crawling through the Github links? My JOSS paper is indexed, but that's from being uploaded to my own website (https://scholar.google.com/citations?view_op=view_citation&citation_for_view=2QJwoAwAAAAJ:3BvdIg-l-ZAC), but I feel like that's still sufficiently different than being on researchgate or a university's own repository.

[pjotrp]

Maybe it is an idea to move the static pages off github anyway. I know people who do not want to publish with JOSS because of the tight github connection. I think it will be fine to use the issue tracker etc., but at least it won't look like JOSS being a github subsidiary.

[cMadan]

Github pages is more integrated now (no longer requires the gh-pages branch). It might be easier to switch to that now, rather than having the paper PDFs being loaded through the Github file preview.

[kyleniemeyer]

OK, so I actually just submitted JOSS to Google Scholar page for requesting indexing of a journal. This is the (automated) response I got:

Homepage: http://joss.theoj.org
Contact name: Kyle Niemeyer
Contact email: kyle.[email protected]
Inclusion type: Other journal website
Inclusion size: 51-100
Volume URLs: http://joss.theoj.org/papers/popular
Issue URLs: http://joss.theoj.org/papers/popular
TOC URLs: http://joss.theoj.org/papers/popular
Abstract URLs: http://joss.theoj.org/papers/10.21105/joss.00194
http://joss.theoj.org/papers/10.21105/joss.00011
http://joss.theoj.org/papers/10.21105/joss.00189
Article URLs: https://github.com/openjournals/joss-papers/blob/master/joss.00194/10.21105.joss.00194.pdf
https://github.com/openjournals/joss-papers/blob/master/joss.00011/10.21105.joss.00011.pdf
https://github.com/openjournals/joss-papers/blob/master/joss.00189/10.21105.joss.00189.pdf
https://github.com/openjournals/joss-papers/blob/master/joss.00012/10.21105.joss.00012.pdf
https://github.com/openjournals/joss-papers/blob/master/joss.00016/10.21105.joss.00016.pdf
If your content meets our guidelines, you can generally expect to find it included within the Google Scholar results within 4-6 weeks.

Please keep in mind that bibliographic data is extracted from your pages by automatic software. If you aren’t satisfied with the accuracy of your listings, please refer to our technical guidelines at http://scholar.google.com/intl/en/scholar/inclusion.html for ways to provide more accurate bibliographic data.

Regards,

The Google Scholar team

I agree that having the paper PDFs linked to directly rather than the GitHub file preview may be smart—not sure if that will matter for Google Scholar.

[arfon]

I agree that having the paper PDFs linked to directly rather than the GitHub file preview may be smart—not sure if that will matter for Google Scholar.

Just incase that's an issue I've updated the URLs on the site to link to the 'raw' GitHub URLs which means they don't display in the GitHub UI. An example of this is:

• Old: https://github.com/openjournals/joss-papers/blob/master/joss.00194/10.21105.joss.00194.pdf
• New: https://github.com/openjournals/joss-papers/raw/master/joss.00194/10.21105.joss.00194.pdf

[wojdyr]

I used to have the same problem a couple years ago when I put reprints of my papers into a github repository. I waited more than a year and it was still not in Google Scholar. Then I moved PDFs to a repository served through github pages -- and this helped.

I cannot be sure what the issue was, but perhaps it's because PDFs from github.com/.../raw/master/... are served as Content-Type "application/octet-stream" instead of "application/pdf".

Additional benefit of serving PDFs through github pages would be that the URL would look better, e.g.
https://openjournals.github.io/joss-articles/10.21105.joss.00194.pdf

[arfon]

I used to have the same problem a couple years ago when I put reprints of my papers into a github repository. I waited more than a year and it was still not in Google Scholar. Then I moved PDFs to a repository served through github pages -- and this helped.

👍 thanks @wojdyr, that's very helpful. Good point about the application/octet-stream content type possibly upsetting the Google bot.

[arfon]

👍 thanks @wojdyr, that's very helpful. Good point about the application/octet-stream content type possibly upsetting the Google bot.

OK in https://github.com/openjournals/whedon/pull/11 I've modified the URLs we're serving to e.g. http://www.theoj.org/joss-papers/joss.00411/10.21105.joss.00411.pdf . Fingers-crossed that helps!

[FaustinCarter]

My paper is now indexed on Google Scholar! Others should check as well. That last modification might have done the trick.

[arfon]

Which paper is that @FaustinCarter?

[FaustinCarter]

Actually, my excitement may have been premature. It looks like it may only be indexed because it was listed as a citation here (thanks to the kind soul who cited it in a more traditional publication): http://adsabs.harvard.edu//abs/2016JOSS.2016...46B.

The JOSS link is: http://joss.theoj.org/papers/10.21105/joss.00046

If you search for "pygtc" on http://scholar.google.com it shows up as the first link, but with a [CITATION] tag preceding the title. I think this means that it is grabbing it from the adsabs rather than indexing it directly. This is further motivated by the fact that on both Google Scholar and the Harvard Adsabs service the abstract is listed as "Not available".

Bummer.

[arfon]

I believe this is now fixed.

[arfon]

See https://scholar.google.com/scholar?hl=en&as_sdt=0%2C9&q=10.21105&btnG= for example.

[Benjamin-Lee]

Still a little confused as to why my paper isn't being listed. @arfon is there any reason you can think of?

[arfon]

@Benjamin-Lee - not sure. I'm following up with some folks about this.

[FaustinCarter]

Another paper that hasn't made it is: https://joss.theoj.org/papers/cf6f8ac309d6a18b6d6cf08b64aa3f62

[arfon]

@FaustinCarter - yes, it looks like something stopped working in early August this year.

[eendebakpt]

@arfon This https://joss.theoj.org/papers/0c6638f84a1a574913ed7c6dd1051847 paper was indexed, but the date was not (yet) extracted. The format of the JOSS papers does not meet the specifications that google scholar used to have. The specifications have changed a little but, currently, the date in joss papers is not as suggested at https://scholar.google.com/intl/en/scholar/inclusion.html#indexing section 2.a.C.
This issue was closed some time ago, should we reopen a new issue?

[arfon]

This issue was closed some time ago, should we reopen a new issue?

AFAIK, Google Scholar doesn't index us directly, rather, our papers are picked up via ADS. I'll follow up with the folks at ADS to see if there's something different we should be doing.

[leios]

I was going to create a new issue for this, but seeing as there seems to be recent discussion on this thread, I would like to also mention that our article does not seem to be picked up by Google Scholar.

I don't know if this information is helpful, but the last article that seems to be picked up by theoj.org is this one (50 days ago): https://www.theoj.org/joss-papers/joss.01102/10.21105.joss.01102.pdf.

The last one picked up by ads is this one (147 days ago): http://adsabs.harvard.edu/abs/2018JOSS....3..854M

It seems indexing via google scholar seems to have stopped roughly 50 days ago, but there doesn't seem to be any PR that should affect this around that time. I suppose a solution for now would be to upload my article to an institutional repository of my university?

[kgjerde]

@arfon

I wonder if this might be the problem:

1. Published JOSS paper pages now seem to leave the Google Scholar tag "citation_author" empty:
   <meta name="citation_author" content="">.
   Example: https://joss.theoj.org/papers/10.21105/joss.01342

2. Google Scholar (https://scholar.google.com/intl/en/scholar/inclusion.html#indexing) states that:

At least one author tag is required for inclusion in Google Scholar.

The time frame of the problem, as stated by you, supports this hypothesis:

@FaustinCarter - yes, it looks like something stopped working in early August this year.

Paper from 1 September 2018 with no author tag: https://joss.theoj.org/papers/efb9242db91adee8c8265f000f26ef5a

Paper from 29 June 2018 with author tag:
https://joss.theoj.org/papers/049f6d3dab9391e8353484028148dd0d

[mschubert]

The

<meta name="citation_author" content="">

tag (denoted by ^empty) is not enough to explain the missing indexing (). I checked some of the papers below, and it doesn't explain why some are listed from JOSS

However, there's definitely still issue with Google Scholar indexing. Maybe this issue should be reopened? @arfon

From 1 month ago:

• treex - indexed, but also multiple citations ^empty
• diwanalr - actually indexed from theoj.org ^empty
• sbpy - not indexed ^empty
• phylosmith - actually indexed from theoj.org ^empty
• Multi-attribute task builder - actually indexed from theoj.org ^empty

Also 2 months:

• QMRTools - not indexed ^empty
• stingray - indexed, but also in arXiv ^empty
• MPIFiles.jl - indexed, but also cited from the International Journal on Magnetic Particle Imaging ^empty
• SMACT - actually indexed from theoj.org ^empty
• pymcmcstat - only indexed as citation ^empty

Also 3 months ago:

• HRDS - indexed, but from eprints.whiterose.ac.uk ^empty
• anesthetic - indexed, but from arXiv ^empty
• AMReX - indexed, but from cloudfront.escholarship.org ^empty
• ggparliament - actually indexed from theoj.org ^empty
• rGUIDANCE - only indexed as citation ^empty
• uJVM - only indexed as citation ^empty

Also 6 months ago:

• Multiblock PLS - indexed, but also on orbit.dtu.dk ^empty
• fibergen - indexed, but also on researchgate ^empty
• BladeX - indexed, but also on researchgate and arXiv ^empty

[arfon]

Hi all, thanks for digging into this further. A couple of things:

• I've just pushed an update to the site which should mean the <meta name="citation_author" content=""> tag is now populated for all paper.
• I'm working with the folks at ADS to have JOSS indexed more regularly (daily) which should mean that papers start showing up on Google Scholar more rapidly.

I'm not sure what more to do at this point but am open to suggestions/improvements.

[pgrete]

Is there any update on this, e.g., is there a way to manually update the indexing? (Specifically asking for https://joss.theoj.org/papers/10.21105/joss.01636) As a side note I noticed that the "Copy bibtex button" doesn't seem to get updated automatically.

[labarba]

I looked up that paper on Google Scholar, using the title, and in this case it is being indexed through adsabs.harvard.edu.

The citation info in Harvard style is given as:

Brummel-Smith, C., Bryan, G., Butsky, I., Corlies, L., Emerick, A., Forbes, J., Fujimoto, Y., Goldbaum, N., Grete, P., Hummels, C. and Kim, J.H., 2019. ENZO: An Adaptive Mesh Refinement Code for Astrophysics (Version 2.6). The Journal of Open Source Software, 4.

and the BibTeX is given as:

@article{brummel2019enzo,
  title={ENZO: An Adaptive Mesh Refinement Code for Astrophysics (Version 2.6)},
  author={Brummel-Smith, Corey and Bryan, Greg and Butsky, Iryna and Corlies, Lauren and Emerick, Andrew and Forbes, John and Fujimoto, Yusuke and Goldbaum, Nathan and Grete, Philipp and Hummels, Cameron and others},
  journal={The Journal of Open Source Software},
  volume={4},
  year={2019}
}

... which all seems OK. What is your concern about the indexing of this article?

[pgrete]

I see. Looks like I got confused/missed it because neither authors nor references are parsed from ADS.
Regarding bibtex, when I click the button the content that ends up in my clipboard is "BibTex entry not available. Please check back later."

[labarba]

Oh, I was grabbing the BibTeX info from the Google Scholar "Cite" dialog, not the JOSS website.

[arfon]

Regarding bibtex, when I click the button the content that ends up in my clipboard is "BibTex entry not available. Please check back later."

Yeah, this is super-buggy and I've just removed it from the UI for now until we can find a long-term fix.