This replaces the hardcoded interface to GitHub releases on boettiger-lab/taxadb-cache with a content-based identifier approach. This lets taxadb be agnostic to where the source data lives, it instead merely resolves the identifier. (Well, actually it first looks up the identifier from the provenance record).
This shouldn't really change much for the user-facing interface, but it does mean we can update and relocate the data files behind the scenes without needing to update taxadb. Also, caching is automatic and we have a more reliable way to check precisely what version of the data a user is using.
Users wanting to interact directly with the data file without using an SQL database can access the tables directly through the tl_connect() function, which also powers the download step in tl_create() now.
once this looks good I think I'll push an update to CRAN.
This replaces the hardcoded interface to GitHub releases on boettiger-lab/taxadb-cache with a content-based identifier approach. This lets taxadb be agnostic to where the source data lives, it instead merely resolves the identifier. (Well, actually it first looks up the identifier from the provenance record).
This shouldn't really change much for the user-facing interface, but it does mean we can update and relocate the data files behind the scenes without needing to update
taxadb
. Also, caching is automatic and we have a more reliable way to check precisely what version of the data a user is using.Users wanting to interact directly with the data file without using an SQL database can access the tables directly through the
tl_connect()
function, which also powers the download step intl_create()
now.once this looks good I think I'll push an update to CRAN.