Hi,
I’m routinely using taxize to match species across datasets (btw thanks to all people involved for providing these super useful tools!)
I am looking for a way to better reproduce results by looking up a previous version of the Gbif backbone. Is it possible to use taxize with local data or only via the web API? From what I understand taxadb and taxizedb work with local data, but I haven’t found how to request older versions of the databases either.
Thanks!
apologies, we’re migrating taxadb to a new data format and have just added the latest versions in the most recent release. We will be adding the old versions back in soon though! Do you know specifically which range of versions you would be interested in from GBIF?
Meanwhile, older versions of GBIF taxonomic backbone can always be accessed directly from GBIF here: GBIF Hosted Datasets
Thanks for the quick reply! For now I’m mainly interested by one of the last versions of GBIF (I think 2021-11-26) just so that I can reproduce results from last year, as the last update seems to have messed up with my analyses.
I already downloaded the archive from gbif, my question was more about how to get the packages to use it (e.g. in taxize functions such as gnr_resolve or get_gbif_id)
Hi @mney,
The way taxizedb works is you first download a database with db_download() and then look up your query within this database. All databases are stored in a single directory on your system, you can get its path using tdb_cache$cache_path_get(). In case of GBIF the downloaded file will be called gbif.sqlite. Once downloaded, it remains unchanged until you force update with db_download(overwrite = TRUE). If you want to keep different versions of your database to make your queries reproducible, you can copy files to a secure location or simply rename them before updating.
So, if by any change you or one of your colleagues downloaded the 2021-11-26 version of GBIF with taxizedb and still has it on their system, you should be able to recover it. If not, there is another repo about converting the download GBIF database into the format used by taxizedb: GitHub - sckott/gbif-backbone-sql: GBIF Darwin Core taxonomic backbone to SQLite.
Do you think these could help?