This is release 5.2.0 of antiSMASH. Highlights are: * added detection for lanthidine clusters * improved detection of thioamide-NRP clusters * updates the default PFAM database to 32.0, with matching GO term update * updates the clusterblast database to the antiSMASH-DB version 3 * PKS candidate clusters with products determined by non-colinear gene ordering now record and show the order used * a new command line option `--html-start-compact`, which defaults the HTML overview to the compact view **NOTE:** As the antiSMASH-DB links have changed slightly to be more future proof, the clusterblast database format has also changed slightly to take advantage of that. It is necessary to update the clusterblast database at the same time as updating to 5.2 from previous versions. This can be done by running the command `download-databases` when antiSMASH is installed in your environment (or `antismash/download-databases.py` in the source directory). There is also a direct download of the database tarball available [here](https://dl.secondarymetabolites.org/releases/clusterblast/clusterblast_20201012.tar.xz). Detailed changes: Kai Blin (11): nrps_pks_domains: Also detect SMART PKS_PP domain as ACP main: Bump the antiSMASH version logging line to INFO level chore: Bump version to 5.2.0 common: Fix the java version detection when _JAVA_OPTIONS are set config: fix alternative fasttree executable spelling hmm_detection: Improve thioamide-NRP rules to pick up more clusters detection: Add detection rule for lanthidines chore: Pin Biopython to 1.76, as 1.77 refuses to work with still support Python 3.5 nrps_pks_domains: Fix off-by-one errors in featute/motif translation calculations nrps/pks: Don't crash in orderfinder if there's no complete module main: More control over output filenames Simon Shaw (40): secmet: fix region genbank outputs not adjusting locations in qualifiers nrps_pks_domains: allow AT and CP without KS to be complete module nrps_pks_domains: only allow condensationless modules to be complete if first in a CDS common.record_processing: include record id in error message html: remove duplicate mod-kr CSS rule secmet: handle genbank conversions introducing spaces into aSDomains *peptides: import joblib directly for future scikit-learn versions secmet: replace minimock usage with unittest lassopeptides: replace minimock usage with unittest active_site_finder: replace minimock usage with unittest t2pks: replace minimock usage with unittest nrps_pks: replace minimock usage with unittest clusterblast: replace minimock usage with unittest cassis: replace minimock usage with unittest antismash.test: remove unnecessary mocking pfam2go: replace minimock usage with unittest nrps_pks: replace minimock usage with unittest cluster_hmmer: replace minimock usage with unittest full_hmmer: replace minimock usage with unittest lanthipeptides: replace minimock usage with unittest thiopeptides: replace minimock usage with unittest antismash.test: replace minimock usage with unittest chore: remove minimock as a test dependency secmet: remove list inheritance from NRPSPKSQualifier secmet: improve error message for invalid translation table qualifiers common: only check relevant parts in RODEO's sklearn version check hmm_detection: find more transAT-PKS clusters pfam2go: update GO mapping download_databases: update to 32.0 for latest Pfam database common.record_processing: conditionally allow long record identifiers clusterfinder: fix subregions being added twice thiopeptides: fix results regeneration for extra ORFs t2pks: only use CDSs completely within the protocluster