mirror of
https://github.com/msberends/AMR.git
synced 2025-07-22 14:53:20 +02:00
(v1.5.0.9008) Internal data sets to pkg, speed for auto col determination
This commit is contained in:
8
NEWS.md
8
NEWS.md
@ -1,5 +1,5 @@
|
||||
# AMR 1.5.0.9007
|
||||
## <small>Last updated: 18 January 2021</small>
|
||||
# AMR 1.5.0.9008
|
||||
## <small>Last updated: 22 January 2021</small>
|
||||
|
||||
### New
|
||||
* Support for EUCAST Clinical Breakpoints v11.0 (2021), effective in the `eucast_rules()` function and in `as.rsi()` to interpret MIC and disk diffusion values. This is now the default guideline in this package.
|
||||
@ -32,10 +32,12 @@
|
||||
* Updated the data set `microorganisms.codes` (which contains popular LIS and WHONET codes for microorganisms) for some species of *Mycobacterium* that previously incorrectly returned *M. africanum*
|
||||
* Added Pretomanid (PMD, J04AK08) to the `antibiotics` data set
|
||||
* WHONET code `"PNV"` will now correctly be interpreted as `PHN`, the antibiotic code for phenoxymethylpenicillin ('peni V')
|
||||
* Fix for verbose output of `mdro(..., verbose = TRUE)` for German guideline (3MGRN and 4MGRN) and *P. aeruginosa* in Dutch guideline (BRMO)
|
||||
* Fix for verbose output of `mdro(..., verbose = TRUE)` for German guideline (3MGRN and 4MGRN) and Dutch guideline (BRMO, only *P. aeruginosa*)
|
||||
* `is.rsi.eligible()` now returns `FALSE` immediately if the input does not contain any of the values "R", "S" or "I". This drastically improves speed, also for a lot of other functions that rely on automatic determination of antibiotic columns.
|
||||
|
||||
### Other
|
||||
* Big documentation updates
|
||||
* Loading the package (i.e., `library(AMR)`) now is ~50 times faster than before
|
||||
|
||||
|
||||
# AMR 1.5.0
|
||||
|
Reference in New Issue
Block a user