Within the Shiny app (#4), once an interaction in the catalog is selected through the dynamic table (#6), an action button should display a detailed view of the entry featuring the attributes of the interaction detailed in the framework
Upon discussions with Alan Pacheco & Daniel Segrè, the catalog could be exported as a JSON towards the database of correlative interactions (MIND).
But NCBI TaxID are needed beforehand (#16)
The - could be translated as NULL as the R language would consider this entry empty.
The unknown string could be translated as a proper NA because knowledge is non available.
Using a set of boxes to ticks, entries to select, the new entry could be properly encoded following the framework.
The interaction name can be deduced from the two strains names and suggested with updateTextBox. The interaction number should be deduced from the existing table.
In the end, the builder will provide first a csv line to be pasted either online on Github or in any editor, into the csv file of the catalog, to be in the end submitted as a pull request.
The detailed entry in #7 has only the doi whereas a bibliography formatting should be preferred. The package rcrossref could be suitable.
Perhaps a correspondence list of doi towards biblio. format could be built to prevent needless queries to crossref for doi already known. This of course should be re-run for each new contribution
A serious improvement from the builder of new entry (#9) would be to trigger a PR directly from the Shiny app. I am not sure if it is possible but worth a try!
Consider the possibility to use doi not from peer-reviewed articles but from experiments registered with zenodo for new catalog entry. This could add flexibility for ongoing work without sacrificing traceability with ORCID and affiliations.
The API of rcrossref (#8) does also work with zenodo doi.