Simplifying OA Policy Compliance for Authors Through a Publisher-Repository Partnership

S In April of 2015

University of Toronto Libraries with an idea of automated transfer and deposit of author manuscripts into the U of T's research repository, TSpace 1 .This free service is available to authors publishing their papers in 21 CSP's NRC Research Press journals 2 and only takes authors one click to opt into.The partnership was launched in April 2015 and makes over 1,200 manuscripts openly available annually.

Workflow
The manuscript transfer and deposit workflow developed by the UTL's Information Technology Services team in close consultation with CSP is outlined in Figure 1 (Maistrovskaya, Xiao Zhao, 2017) and the script is shared on Github under the Apache License 2.0 (University of Toronto Libraries, 2018).
The process is initiated by the author in the CSP's peer-review submission system, ScholarOne, opting into their manuscript being archived in TSpace.A zip file containing the accepted manuscript, supplementary files, and metadata in PubMed XML is transferred to the library's SCP server.UTL's Python script scans the directory on an hourly basis and extracts each package.Beautiful Soup 3 is used to parse the received metadata into Dublin Core.The script then prepares DSpace Simple Archive for each manuscript and uses the DSpace import API to upload content into a respective journal's collection.An ingestion report is generated at this point and emailed to the CSP and TSpace staff.Transferred zip files as well as prepared DSpace Simple Archive directories are backed up for quick re-ingestion if needed.
Upon ingest, the manuscript is immediately released for public access in the repository and assigned a permanent URL (handle) in addition to the publisher supplied DOI that points to the final version on the journal website.

Challenges and opportunities 8
Automated repository workflows have been around for a while and can be implemented using a variety of tools and protocols.A common challenge they share is that even a welldeveloped and robust workflow cannot do entirely without human intervention which affects its overall sustainability.In the SCP-UTL case, while the script takes care of package processing, metadata crosswalk, and ingest, human assistance is still needed on both ends of the pipeline with CSP staff creating and depositing the packages and UTL staff maintaining repository collections and troubleshooting technical issues.Such issues are uncommon, but while the vast majority of deposits go smoothly and entirely automatically, the script depends on the consistent structure of the transferred packages and metadata, and occasionally fails when encountering unexpected elements.One promising development that may help address this through a standardized framework is the Manuscript Exchange Common Approach (MECA) recently accepted to be developed as a NISO standard (NISO, 2018).
Apart from technical challenges, establishing a publisher-repository partnership demands commitment, trust, and mutual investment of time and effort-something that can only be initiated when both parties share common values and goals in public sharing and preservation of research.While funder mandates, such as Tri-Agency Open Access Policy (Government of Canada, 2015) or NIH Public Access Policy (U.S.Department of Health & Human Services, 2016) have been a significant drive for these initiatives, our hope in sharing this workflow is to offer a fairly straightforward solution that can be implemented by small scale publishers and university presses looking to include green OA in their suite of services.

Conclusion
Making research openly accessible to the widest possible audience, and at the earliest possible opportunity has multiple benefits.Funded by Canadian federal grants, it makes the findings available to the public beyond journal subscriptions that are often limited to institutions and research organizations.Open availability of research benefits the authors and "science by accelerating dissemination and uptake of research findings" (Eysenbach, 2006, e157).It also facilitates societal advancement globally by "enabling researchers, scholars, clinicians, policymakers, private sector and not-for-profit organizations and the public to use and build on this knowledge."(Government of Canada, 2015).
Nonetheless compliance with funders' open access mandates remains challenging and complicated for many researchers, despite assistance available from their institutions in the form of consultations, institutional repository infrastructure, copyright support, and even mediated deposit services.As agreed by many Canadian institutional repository managers, one of the major stumbling blocks in streamlining compliance via the "green" route is getting hold of accepted manuscripts where publisher policy does not permit the deposit of a final article in a repository.
The automated manuscript deposit workflow removes the dependence on researchers keeping the ephemeral manuscript version.By transferring both the manuscript and accompanying metadata into a repository directly from the publisher's system, it makes compliance seamless and hassle free for grant recipients, and opens doors for further publisher-repository collaboration.