Added

• Added experimental support for running the classifier as a daemon. The classifier daemon will persist any database that it has loaded allowing users to save time when running multiple samples on the same large database. This feature is currently only available with the k2 wrapper by specifying the --use-daemon option when running k2 classify.

  Example: k2 classify --db pluspfp --use-daemon --threads 12 sample.fa --output sample.out On first invocation the pluspfp database will be loaded into memory. On subsequent runs of this command the database load step will be skipped. If a new database is provided that database will persisted in memory as well.

• Added --stop-daemon option to k2 clean. This option will stop a running classifier daemon.

• --resume option has been added to k2 download-library. k2 download-library has been updated so that files downloaded from NCBI will be saved with a .tmp extension. This extension will only be removed after k2 has verified that the MD5 sum of the downloaded file matches that of the server. The option builds off of that foundation by skipping files that do not have the extension. If a user has a library that was downloaded with a version of Kraken 2 prior to version 2.1.5, the resume option will fail if partially downloaded files exist. Run without the option first to ensure that all files are fully downloaded after which resume can be used successfully. When using 2.1.5 to download a new library, should work without issue on a failed download.