bvreede / oamonitoring Goto Github PK

I went through the README and saw that is focused on the Utrecht situation with precise data from when files have been exported, etc. I would suggest to move this specific information to the Wiki and let the README be more general and really about the code itself and mock input and output. I can then also add information about the Leiden situation to the Wiki. What do you think?

It is hard to predict what elements are present in a CRIS that can be used to resolve green OA.

In 2018, Utrecht University used a combination of columns indicating what kind of document was submitted in their CRIS (Author Accepted Manuscript or Published Version), and the presence of an embargo period. This is hard to replicate and generalize. For one, the notation in another CRIS may be different. Secondly, the presence alone of an embargo period is not a good indicator for something being openly available (though if embargo dates are present, they could be read as dates and confirmed to exist in the current year?).

Unpaywall tends to overestimate green, but has a more sophisticated ability to look into this (I believe they screen the pdfs themselves for copyright information?). Unpaywall also provides a link to documents, so they also confirm that the document is hosted on a repository.

Unpaywall is the best generalizable option we have at the moment to confirm green... Any other ideas?

At the moment, the APIs for DOAJ and Unpaywall are used. Mining this api takes a while, and the data is saved, along with a date stamp of the record. It would be good if saved data can be used in future iterations of the project, but it should be optional: a user should be able to get current data from an API or refer back to saved data, depending on their wish.

How can we do this?

• With a simple option in the config file
• With a user question when the API is accessed

Which route is preferable? And how do we execute it?

It would be helpful to include essential data, namely the VSNU list of publications. Not sure if this is possible though so... @boer0107 ? What do you think?

We should include a license. At the Leiden University Libraries we usually use GPL 3.0 for our developed software. Do you agree with that? I am also fine with MIT, Apache 2 or another permissive license.

Many universities used Taverne to make output available via the university repository last year. It would be interesting to be able to score these papers separately. Additionally, other universities may have other projects that they want to label independently.

We could envision having the option of either providing a list of DOIs (or system IDs) for papers that a university wants to classify separately, as well as the classification (GOLD/HYBRID/GREEN -- or even an additional option?) and the source.