preVIEW: from a fast prototype towards a sustainable semantic search system for central access to COVID-19 preprints

Abstract

The current COVID-19 pandemic emphasizes the use of so-called preprints - a type of publication that is not subject to peer review. Due to its global relevance, there is an immense number of COVID-19-related preprints every day. To help researchers find relevant information, we have developed the semantic search engine preVIEW, it integrates preprints from currently seven different preprint servers. For semantic indexing, we implemented various text mining components to tag, for example, diseases or SARS-CoV-2 specific proteins. While the service initially served as a prototype developed together with users, we present a re-engineering towards a sustainable semantic search system, which was inevitable due to the continuously growing number of preprint publications. This enables easy reuse of the components and allows rapid adaptation of the service to further user needs.