TDM click-through service

Hi there,

Caveat: Forgive me if I have missed the documentation about this on the Crossref site. Can you share how the click-through service works for TDM, and what would need to be done to enable this for publishers looking to leverage Crossref’s TDM feature?

Thank you!

1 Like

Hi @dplavin, welcome to the community forum! We are deprecating the click-through service because it was just not supported or used by the community.

This API query -,has-full-text:true,license.version:tdm - will give you all the DOIs with a text and data mining license registered with us.

Please let me know if you have any additional questions.

Kind regards,

Is there any further information on this? I do use the click through service in this R package GitHub - ropensci-archive/crminer: ⛔ ARCHIVED Fetch 'Scholary' Full Text from 'Crossref' At least with respect to pointing users to to get authenticated . In crminer, we then use the user’s authentication details from the click through page to fetch pdf/xml/etc. Is there anything to replace this aspect, allowing users to get full text with appropriate authentication?

Hi @sckott. Thanks for your message. We haven’t defined a timeline for that deprecation yet, but, based on my understanding, click-through was only ever meant to be an interim solution until publishers included tdm in their regular licenses. Many are doing that now.

what does included tdm in their regular licenses that mean exactly?

Will there still be a way for users to request full text (pdf/xml) with some kind of Crossref API key? Or will that be on a publisher by publisher basis?

1 Like

Hi @sckott, once the click-through service is deprecated there will not be a Crossref API key to request full text. We won’t be retaining that role as an intermediary.

So, the way that will work for researchers is:

To use Crossref’s metadata services to perform TDM, researchers should have a list of DOIs for the content they want to download, and a whitelist of licenses that they accept. Researchers can get a list of DOIs from citations, our metadata search, our metadata API, or another source.

For each DOI, researchers should:

  1. Use content negotiation to get the metadata for the DOI
  2. Check to see if the DOI has license and full-text details in its metadata
  3. Check the license against your whitelist of acceptable licenses
  4. If you agree to the license, follow the link and download the full-text of the content item.

Does that answer your question?

My best,

Sure, that makes sense. Does deprecated here mean gone, or in the software sense, where it’s no longer meant to be used, but is still useable for some time period? And when will it be completely gone/unavailable?

I believe that means that the software is no longer meant to be used - that is we are no longer supporting it - but it will still be useable for a period of time. I don’t have a timetable, but let me see if I can get an estimate for you.

@Ginny do you know any more specifics about deprecating click-through?

1 Like

Hi @sckott - the desire to deprecate it depends on a review scheduled for December. As part of that we should definitely chat to understand better how the rCrossref (or crminer?) package is using it. Only three publishers have ever set up this click-through so it doesn’t cover our full metadata anyway and in total over many years only 80 users have actually clicked through (last time we checked). It may be that it’s just not a necessary step for your tool and you can bypass this now. But we won’t make a decision until we review properly how it’s being used (for example those 80 ‘users’ could each be representing thousands of others). So this is very good to know. I’ll make sure the product team (Bryan Vickery is our new product director) is aware and we will reach out to you soon. Tagging @KirstyM in the meantime.

Hope you are well too!

1 Like

Hi @Ginny Thanks for the detailed response. The click-through service is used in both rcrossref and crminer R packages; I don’t know what usage is like for TDM through these packages, but I imagine user agent strings could be examined to get a sense; we send user agents strings from those two packages that include the package name and the string ropensci.

A Ruby library (textminer) i maintain also uses the service, but usage of this library is probably low to non-existent.

All I’m looking for is 1) timing on when things will change, and 2) clear steps on how to allow users to fetch full text assuming they have access. For 2), @ifarley gave some steps above; however, links to full text in my experience are not always up to date, so that could be an issue