Some of your publications are already available in open access on the publisher’s website and you wish to deposit them? What if HAL collected them for you? This will soon be possible with a new service planned for the end of 2023, which will offer researchers a deposit suggestion enabling them to automatically import their publications if they are identified with a DOI and distributed under a Creative Commons licence.
A new step towards simplifying the process of making submissions to HAL is to collect and import publications that are already open access. This is one of the measures of the second National Plan for Open Science to make it easier for researchers to use HAL, and thus enables them to comply with the incentives or obligations to deposit, whether institutional or from funders.
The development and deployment of such a service is a major focus of the Equipex+ HALiance project (work package 6).
This automatic collection of full-text publications service is currently in the final stages of development. The project leader, Yannick Barborini, presented it on the 10th of July. To watch the replay of his presentation and demonstration (in french), see below:
HAL makes suggestions, researchers make decision
This new service dedicated to researchers is built upon suggestions of publications to be imported into HAL. The service uses DOI to identify publications that are already open access on other platforms, such as publishers’, and suggests them in an user interface: if the user validates the suggestion, the deposit is automatically imported.
The feature cover both the creation of a new record and the addition of a file to an already existing record.
The feature has two parts:
- a screen which displays a list of suggestions, and the possible actions for each of them: to see the file which can be imported, to import it into HAL, to report that it is not the correct file or to remove the suggestion;
- another screen to configure the service and to refine the search criteria, for example by adding the ORCID (very useful in case of homonym). Users can also activate an alert parameter to be notified of any new suggestions.
How are suggestions produced?
HAL uses a database created by Inist, which is updated once a month.
The first step is to identify the French scientific production: this is carried out by the data aggregator OpenAlex. Then, several data sources (CrossRef, HAL and Pubmed) are used to retrieve bibliographic metadata of this french production in order to refine and build a database of unified records.
Metadata is then enriched by adding a HAL scientific domain and, as much as it is possible, the RNSR identifier for authors’ affiliations (RNSR is a national repertory for research structures). Then, the process carries out an alignment with the Unpaywall data to select the file that is appropriate to import into HAL.
Finally, the status in HAL is checked to know if the publication is not already in HAL, or if it is only a bibliographic record.
The publication must fulfil at least two conditions to be included in the Suggestions list:
- the file is not under embargo on the source platform,
- and metadata contains a Creative Commons licence: this guarantees that HAL can legally distribute the file.
A quantitative analysis of the current database gives an estimate of 72,000 potential imports covering the period 2014-2023: 5,000 new publications and 67,000 fulltext completing existing bibliographic records*. These are mainly journal articles and book chapters, of all scientific fields.
The service is currently under development. A phase of interviews with researchers has already helped to refine and improve the user screens. Once users’ feedback has been integrated, a more massive test phase will follow. The service is scheduled to be available by the end of 2023.
Yannick Barborini will present it at Open Science FAIR on the next 26th of September.
*the calculation method having been refined, note that these figures are slightly different from those announced during the webinar.
Federico
Excellent! Will user suggestions/confirmations be required only for the 5k new records or also for the 67k existing records without files? For existing records, making available the full text increases the chances of identifying and correcting any metadata mistakes. There’s less need to archive the full text if it’s already available in another open repository (“has_repository_copy” field in Unpaywall).
Can’t wait for this to happen automatically for all CC papers before 2014 and eventually all the millions of “French” papers which are not yet preserved, or are only preserved in the Internet Archive (https://blog.archive.org/2021/03/09/search-scholarly-materials-preserved-in-the-internet-archive/ ).
Agnès Magron
Thanks for your comment.
Confirmation of import is required for both a new deposit (creation) and the addition of full text to an existing bibliographic record. It is a new way for researchers to deposit their publications, making it easier for them to use HAL. They can also take the opportunity to correct or improve the metadata of their deposit.