search

MDU-PHL / meningotype

In silico serotyping, finetyping and Bexsero antigen sequence typing of Neisseria meningitidis
GNU General Public License v3.0
11 stars 6 forks source link

Updating DB outside of tool-installation directory should be possible #41

Open pvanheus opened 1 year ago

pvanheus commented 1 year ago

Currently the tool can update DB by downloading updated files from PubMLST, and, with the specification of a --db location, this update can run even if the tool's install location is read-only (as in the case of e.g. a system-wide install on a multi-user system or in the case of the tool encapsulated in a Docker container). The resulting updated DB will, however, be incomplete.

A work-around is possible where the install-time DB is combined with the updated files downloaded from PubMLST, but preferably the tool should allow the DB to be updated to an arbitrary DB location without resulting in an incomplete DB.

Additionally, the tool should only download files if a newer file is available, i.e. it should use the 'If-Modified-Since' header for HTTP/HTTPS requests. While this is not super-easy with urlretrieve it is possible with (https://stackoverflow.com/a/59602931).

pvanheus commented 1 year ago

On the point of having checking if the PubMLST alleles files have been modified - currently they don't provide the right headers but there is an open issue: https://github.com/kjolley/BIGSdb/issues/913

pvanheus commented 1 year ago

On the point of having checking if the PubMLST alleles files have been modified - currently they don't provide the right headers but there is an open issue: kjolley/BIGSdb#913

Following on from this, Keith Jolley pointed out to me that alleles, schemes and isolates all have a last_updated or last_added attribute that can be used to get info on updates - e.g. http://rest.pubmlst.org/db/pubmlst_neisseria_seqdef/loci/PorA_VR1/alleles.