Release 0.2.7

CITATION.cff

@@ -1,8 +1,8 @@
 cff-version: "1.1.0"
 message: "If you use this software, please cite it using these metadata."
 title: ElasticBLAST
-version: "0.2.6"
-date-released: 2022-06-08
+version: "0.2.7"
+date-released: 2022-08-11
 license: "NCBI Public Domain"
 repository-code: "https://github.com/ncbi/elastic-blast/"
 authors: 

bin/aws-get-auto-scaling-events.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+# aws-get-auto-scaling-events.sh: Get autoscaling events for ElasticBLAST's
+# AWS Batch compute environment
+#
+# Author: Greg Boratyn ([email protected])
+# Created: Fri Aug 12 17:28:20 EDT 2022
+
+# The script assumes that elastic-blast.log file exists
+logfile=${1:-elastic-blast.log}
+COMP_ENV_NAME=$(grep ComputeEnvName $logfile | tr '/' '\t' | cut -f 2 | tail -n 1)
+if [ ! -z "${COMP_ENV_NAME}" ] ; then
+    AUTO_SCALE_GRP_NAME=$(aws autoscaling describe-auto-scaling-groups --output json | jq -Mr '.AutoScalingGroups[].AutoScalingGroupName' | grep $COMP_ENV_NAME)
+    if [ $? -eq 0 ] ; then
+        aws autoscaling describe-scaling-activities --auto-scaling-group-name $AUTO_SCALE_GRP_NAME
+    else
+        echo "Failed to find an AWS auto scaling group for the AWS Batch Compute environment $COMP_ENV_NAME"
+        exit 1
+    fi
+else
+    echo "Failed to find an AWS Batch Compute environment in $logfile"
+    exit 1
+fi

bin/blast-tuner.py

@@ -75,7 +75,7 @@ def main():
         if args.db is not None:
             try:
                 db_metadata = get_db_metadata(args.db, sp.get_db_mol_type(args.program),
-                                              db_source)
+                                              db_source, gcp_prj=args.gcp_project)
             except FileNotFoundError:
                 raise UserReportError(returncode=BLASTDB_ERROR,
                                       message=f'Metadata for BLAST database "{args.db}" was not found or database molecular type is not the same as required by BLAST program: "{args.program}"')
@@ -169,6 +169,7 @@ def create_arg_parser():
                         help='Number of residues or bases in query sequecnes')
     optional.add_argument("--db-source", type=str, help="Where NCBI-provided databases are downloaded from, default: AWS", choices=['AWS', 'GCP', 'NCBI'])
     optional.add_argument("--region", type=str, help=f'Cloud Service Provider region. Defaults: {ELB_DFLT_AWS_REGION} for AWS; {ELB_DFLT_GCP_REGION} for GCP')
+    optional.add_argument("--gcp-project", type=str, help=f'GCP project, required if --db-source or --csp-target is GCP')
     optional.add_argument("--options", type=str, help='BLAST options', default='')
     optional.add_argument("--db-mem-limit-factor", type=float,
                           help='This number times database bytes-to-cache will produce memory limit for a BLAST search. (default: 0.0: for AWS, 1.1 for GCP)')

docker-blast/Dockerfile-build-from-local-sources

@@ -52,6 +52,7 @@ RUN apt-get -y -m update && \
 
 RUN mkdir -p /blast/bin /blast/lib
 COPY --from=blast /blast/bin /blast/bin
+COPY update_blastdb.pl /blast/bin
 COPY --from=blast /blast/lib /blast/lib
 COPY --from=blast /root/edirect /root/edirect
 COPY splitq_download_db_search /blast/bin/

docker-blast/Makefile

@@ -30,7 +30,7 @@ GCP_IMG?=gcr.io/ncbi-sandbox-blast/${IMG}
 AWS_SERVER?=public.ecr.aws/i6v3i0i9
 AWS_IMG?=${AWS_SERVER}/elasticblast-elb
 AWS_REGION?=us-east-1
-VERSION?=1.1.0
+VERSION?=1.1.1
 
 ifeq (, $(shell which vmtouch 2>/dev/null))
 NOVMTOUCH?=--no-vmtouch

docker-blast/splitq_download_db_search

@@ -249,15 +249,15 @@ def _download_database(args, is_user, db_done):
     verbose = ' --verbose --verbose --verbose --verbose --verbose --verbose' if args.verbose else ''
     creds = ' --no-sign-request' if args.no_creds else ''
     nprocs_to_download_db = min(MAX_PROCS_TO_DOWNLOAD_DB, int(os.cpu_count()/args.num_threads))
+    p = safe_exec(f"time update_blastdb.pl taxdb --decompress --source ncbi {verbose} --num_threads {nprocs_to_download_db}")
+    print(p.stdout.decode(), end='')
+    print(p.stderr.decode(), end='')
     if is_user:
-        p = safe_exec(f"time update_blastdb.pl taxdb --source {args.source}{verbose} --num_threads {nprocs_to_download_db}")
-        print(p.stdout.decode(), end='')
-        print(p.stderr.decode(), end='')
         p = safe_exec(f"time aws s3 cp --only-show-errors{creds} {os.path.join(args.db_path,'')} . --recursive --exclude * --include {args.db}.* --include taxdb.*")
         print(p.stdout.decode(), end='')
         print(p.stderr.decode(), end='')
     else:
-        p = safe_exec(f"time update_blastdb.pl {args.db} --source {args.source}{verbose} --num_threads {nprocs_to_download_db}")
+        p = safe_exec(f"time update_blastdb.pl {args.db} --decompress --source {args.source}{verbose} --num_threads {nprocs_to_download_db}")
         print(p.stdout.decode(), end='')
         print(p.stderr.decode(), end='')
     print('End database download')