Optional ReadonlybioconductorMirror?: stringthe Bioconductor host (default https://bioconductor.org)
Optional ReadonlybioconductorRelease?: stringthe Bioconductor release to fetch from (default release)
Optional ReadonlycacheDir?: stringwhere to cache what was fetched (default the on-demand directory of the signature cache)
Optional ReadonlycranMirror?: stringthe CRAN src/contrib/ to fetch from (default https://cloud.r-project.org/src/contrib/)
Readonlyenabled: booleanOptional ReadonlygithubMirrors?: booleanwhether to use the GitHub mirrors of CRAN and Bioconductor (cran/<pkg>, bioc/<pkg>, default true)
Optional ReadonlymaxArchiveMb?: numberthe most MiB a package archive may have (default 32), it may unpack to three times that
Optional ReadonlymaxPackages?: numberthe most packages to fetch for one analysis (default 25)
Optional Readonlyoffline?: booleanonly use what was fetched before, never the network (default false)
Optional Readonlypackages?: string[]only recover packages whose name matches one of these regular expressions, empty means any
Optional Readonlyrepositories?: Record<string, string>GitHub repositories (owner/repo[/subdir][@ref]) to fetch packages from, by package name
Optional ReadonlyretryAfterHours?: numberhow many hours to wait before asking again for a package no source had (default 24)
Optional Readonlysources?: string[]where to look, in order (default github, cran, bioconductor), a plugin may add sources of its own
The package as fetched before, undefined if it never was or no source had it.
Fetches the packages of requests the cache does not answer yet, at most maxPackages of them, and returns
those it recovered. A failing source is logged and the next one tried, so this never throws.
Recovers packages no signature database knows from their sources: asks each OnDemandSource in turn for archives, reads
DESCRIPTION,NAMESPACE, andR/from the first one that holds the package, and caches the result (or that no source had it) as one JSON file per package, so later analyses need no network.