Release 0.2.6

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: "0.2.5"
-date-released: 2022-03-24
+version: "0.2.6"
+date-released: 2022-06-08
 license: "NCBI Public Domain"
 repository-code: "https://github.com/ncbi/elastic-blast/"
 authors: 

DISCLAIMER.md

@@ -0,0 +1,12 @@
+Although all reasonable efforts have been taken to ensure the accuracy
+and reliability of the software and data, the NLM and the U.S.
+Government do not and cannot warrant the performance or results that
+may be obtained by using this software or data. The NLM and the U.S.
+Government disclaim all warranties, express or implied, including
+warranties of performance, merchantability or fitness for any
+particular purpose.
+
+Users of ElasticBLAST are solely responsible for any and all cloud service
+provider charges associated with their use of ElasticBLAST. 
+
+See also: [LICENSE](LICENSE.md)

bin/gcp-setup-elastic-blast-janitor.sh

@@ -24,7 +24,7 @@ usage() {
     echo -e "\t-h: Show this message"
 }
 
-while getopts "u:ph" OPT; do
+while getopts "u:p:h" OPT; do
     case $OPT in 
         u) user=${OPTARG}
             ;;

requirements/base.txt

@@ -1,10 +1,10 @@
 wheel == 0.37.1
 setuptools == 56.0.0
-importlib-resources == 5.4.0
-importlib-metadata == 4.11.3
-pex == 2.1.73
-boto3 == 1.21.24
-botocore == 1.24.24
+importlib-resources == 5.7.1
+importlib-metadata == 4.11.4
+pex == 2.1.92
+boto3 == 1.24.3
+botocore == 1.27.3
 awslimitchecker == 12.0.0
 tenacity == 8.0.1
 dataclasses-json == 0.5.7

requirements/test.txt

@@ -1,13 +1,13 @@
 -r base.txt
 
-pytest == 7.1.1
+pytest == 7.1.2
 pytest-cov == 3.0.0
 pytest-mock == 3.7.0
 teamcity-messages == 1.31
-mypy == 0.941
+mypy == 0.961
 pylint == 2.7.4
-tox == 3.24.5
+tox == 3.25.0
 yamllint == 1.26.3
-moto == 3.1.1
+moto == 3.1.12
 docker == 5.0.3
-cfn-lint == 0.58.4
+cfn-lint == 0.61.0

src/elastic_blast/aws.py

@@ -62,7 +62,7 @@
 from .constants import ELB_AWS_JANITOR_CFN_TEMPLATE, ELB_DFLT_JANITOR_SCHEDULE_AWS
 from .constants import ELB_AWS_JANITOR_LAMBDA_DEPLOYMENT_BUCKET, ELB_AWS_JANITOR_LAMBDA_DEPLOYMENT_KEY
 from .constants import CFG_CLOUD_PROVIDER, CFG_CP_AWS_AUTO_SHUTDOWN_ROLE, CSP
-from .constants import AWS_JANITOR_ROLE_NAME
+from .constants import AWS_JANITOR_ROLE_NAME, ELB_JANITOR_SCHEDULE
 from .constants import STATUS_MESSAGE_ERROR, STATUS_MESSAGE_VERBOSE
 from .filehelper import parse_bucket_name_key
 from .aws_traits import get_machine_properties, create_aws_config, get_availability_zones_for
@@ -240,8 +240,8 @@ def _init(self, cfg: ElasticBlastConfig, create: bool):
                     logging.debug(f'Found janitor role for {AWS_JANITOR_ROLE_NAME}: {role.arn}')
                 except:
                     logging.debug(f'Did not find janitor role for {AWS_JANITOR_ROLE_NAME}')
-            if 'ELB_JANITOR_SCHEDULE' in os.environ:
-                janitor_schedule = os.environ['ELB_JANITOR_SCHEDULE']
+            if ELB_JANITOR_SCHEDULE in os.environ:
+                janitor_schedule = os.environ[ELB_JANITOR_SCHEDULE]
                 logging.debug(f'Overriding janitor schedule to "{janitor_schedule}"')
             if 'ELB_DISABLE_AUTO_SHUTDOWN' in os.environ:
                 janitor_schedule = ''
@@ -916,14 +916,6 @@ def upload_job_ids(self) -> None:
         bucket.put_object(Body=self.job_ids.to_json().encode(), Key=key) # type: ignore
         logging.debug(f'Uploaded job IDs to {self.results_bucket}/{ELB_METADATA_DIR}/{ELB_AWS_JOB_IDS}')
 
-        # This code is needed for janitor backward compatibility in version
-        # 0.2.4, and can be removed when the ElasticBLAST janitor is upgraded to version 0.2.4.
-        ELB_AWS_OLD_JOB_IDS = 'job-ids.json'
-        bucket_name, key = parse_bucket_name_key(f'{self.results_bucket}/{ELB_METADATA_DIR}/{ELB_AWS_OLD_JOB_IDS}')
-        bucket = self.s3.Bucket(bucket_name)
-        bucket.put_object(Body=json.dumps(self.job_ids.to_list()).encode(), Key=key)
-        logging.debug(f'Uploaded job IDs to {self.results_bucket}/{ELB_METADATA_DIR}/{ELB_AWS_OLD_JOB_IDS}')
-
 
     def upload_query_length(self, query_length: int) -> None:
         """Save query length in a metadata file in S3"""

src/elastic_blast/config.py

@@ -46,7 +46,7 @@
 from .constants import ELB_DFLT_AWS_SPOT_BID_PERCENTAGE
 from .constants import APP_STATE_RESULTS_MD5, SYSTEM_MEMORY_RESERVE
 from .constants import ELB_S3_PREFIX, ELB_GCS_PREFIX
-from .constants import ELB_DFLT_AWS_REGION, ELB_DFLT_GCP_REGION
+from .constants import ELB_DFLT_AWS_REGION, ELB_DFLT_GCP_REGION, ELB_DFLT_GCP_ZONE
 from .util import UserReportError
 from .filehelper import parse_bucket_name_key
 from typing import List
@@ -70,6 +70,9 @@ def _load_config_from_environment(cfg: configparser.ConfigParser) -> None:
     """Selected environment variables can be used to configure ElasticBLAST"""
     if 'ELB_GCP_PROJECT' in os.environ:
         cfg[CFG_CLOUD_PROVIDER][CFG_CP_GCP_PROJECT] = os.environ['ELB_GCP_PROJECT']
+    # If GCP project is provided by the environment/configuration, leverage it
+    if 'CLOUDSDK_CORE_PROJECT' in os.environ:
+        cfg[CFG_CLOUD_PROVIDER][CFG_CP_GCP_PROJECT] = os.environ['CLOUDSDK_CORE_PROJECT']
     if 'ELB_GCP_REGION' in os.environ:
         cfg[CFG_CLOUD_PROVIDER][CFG_CP_GCP_REGION] = os.environ['ELB_GCP_REGION']
     if 'ELB_GCP_ZONE' in os.environ:
@@ -140,16 +143,6 @@ def configure(args: argparse.Namespace) -> configparser.ConfigParser:
         retval[CFG_CLOUD_PROVIDER][CFG_CP_GCP_REGION] = args.gcp_region
     if hasattr(args, 'gcp_zone') and args.gcp_zone:
         retval[CFG_CLOUD_PROVIDER][CFG_CP_GCP_ZONE] = args.gcp_zone
-
-    # If results bucket was provided, set the default region in the
-    # corresponding cloud service provider if it wasn't specified by the user
-    if CFG_BLAST_RESULTS in retval[CFG_BLAST]:
-        if retval[CFG_BLAST][CFG_BLAST_RESULTS].startswith(ELB_S3_PREFIX):
-            if CFG_CP_AWS_REGION not in retval[CFG_CLOUD_PROVIDER]:
-                retval[CFG_CLOUD_PROVIDER][CFG_CP_AWS_REGION] = ELB_DFLT_AWS_REGION
-        elif retval[CFG_BLAST][CFG_BLAST_RESULTS].startswith(ELB_GCS_PREFIX):
-            if CFG_CP_GCP_REGION not in retval[CFG_CLOUD_PROVIDER]:
-                retval[CFG_CLOUD_PROVIDER][CFG_CP_GCP_REGION] = ELB_DFLT_GCP_REGION
 
     # Exception to prevent unnecessary API calls and ensure testability
     # of some functionality without credentials
@@ -168,7 +161,7 @@ def _validate_csp(cfg: configparser.ConfigParser) -> None:
     Throws a UserReportError in case of invalid configuration.
     """
     if CFG_CLOUD_PROVIDER not in cfg:
-        report_config_error(['Cloud provider configuration is missing'])
+        return
 
     # are gcp or aws entries present in cloud-provider config
     gcp = sum([i.startswith('gcp') for i in cfg[CFG_CLOUD_PROVIDER]]) > 0
@@ -179,8 +172,6 @@ def _validate_csp(cfg: configparser.ConfigParser) -> None:
     # both and none are forbidden
     if gcp and aws:
         msg.append('Cloud provider config contains entries for more than one cloud provider. Only one cloud provider can be used')
-    if not gcp and not aws:
-        msg.append('Cloud provider configuration is missing')
 
     if CFG_CP_NAME in cfg[CFG_CLOUD_PROVIDER]:
         logging.debug(f'Cloud Service Provider {cfg[CFG_CLOUD_PROVIDER][CFG_CP_NAME]}')

src/elastic_blast/constants.py

@@ -198,7 +198,9 @@ def __str__(self):
 
 
 ELB_DFLT_GCP_REGION = 'us-east4'
+ELB_DFLT_GCP_ZONE = 'us-east4-b'
 ELB_DFLT_AWS_REGION = 'us-east-1'
+ELB_UNKNOWN_GCP_PROJECT = 'elb-unknown-gcp-project'
 
 ELB_DOCKER_VERSION = '1.1.0'
 ELB_QS_DOCKER_VERSION = '0.1.2'
@@ -235,6 +237,7 @@ def __str__(self):
 CFG_CP_GCP_ZONE = 'gcp-zone'
 CFG_CP_GCP_NETWORK = 'gcp-network'
 CFG_CP_GCP_SUBNETWORK = 'gcp-subnetwork'
+CFG_CP_GCP_GKE_VERSION = 'gke-version'
 CFG_CP_AWS_REGION = 'aws-region'
 CFG_CP_AWS_KEY_PAIR = 'aws-key-pair'
 CFG_CP_AWS_VPC = 'aws-vpc'
@@ -317,6 +320,8 @@ def __str__(self):
 # https://docs.aws.amazon.com/eventbridge/latest/userguide/eb-create-rule-schedule.html
 ELB_DFLT_JANITOR_SCHEDULE_AWS = "cron(*/5 * * * ? *)"
 
+ELB_JANITOR_SCHEDULE = 'ELB_JANITOR_SCHEDULE'
+
 ELB_AWS_JANITOR_LAMBDA_DEPLOYMENT_BUCKET = 'elb-camacho'
 ELB_AWS_JANITOR_LAMBDA_DEPLOYMENT_KEY = 'functions/'
 ELB_AWS_JANITOR_CFN_TEMPLATE = 'https://elb-camacho.s3.amazonaws.com/templates/elastic-blast-janitor-cf.yaml'

src/elastic_blast/elb_config.py

@@ -57,6 +57,7 @@
 from .constants import CFG_CLOUD_PROVIDER
 from .constants import CFG_CP_GCP_PROJECT, CFG_CP_GCP_REGION, CFG_CP_GCP_ZONE
 from .constants import CFG_CP_GCP_NETWORK, CFG_CP_GCP_SUBNETWORK
+from .constants import CFG_CP_GCP_GKE_VERSION
 from .constants import CFG_CP_AWS_REGION, CFG_CP_AWS_VPC, CFG_CP_AWS_SUBNET
 from .constants import CFG_CP_AWS_JOB_ROLE, CFG_CP_AWS_BATCH_SERVICE_ROLE
 from .constants import CFG_CP_AWS_INSTANCE_ROLE, CFG_CP_AWS_SPOT_FLEET_ROLE
@@ -82,7 +83,9 @@
 from .constants import ELB_DFLT_AWS_NUM_CPUS, ELB_DFLT_GCP_NUM_CPUS
 from .constants import ELB_S3_PREFIX, ELB_GCS_PREFIX, ELB_UNKNOWN_MAX_NUMBER_OF_CONCURRENT_JOBS
 from .constants import AWS_ROLE_PREFIX, CFG_CP_AWS_AUTO_SHUTDOWN_ROLE
-from .constants import BLASTDB_ERROR, ELB_UNKNOWN
+from .constants import BLASTDB_ERROR, ELB_UNKNOWN, ELB_JANITOR_SCHEDULE
+from .constants import ELB_DFLT_GCP_REGION, ELB_DFLT_GCP_ZONE
+from .constants import ELB_DFLT_AWS_REGION, ELB_UNKNOWN_GCP_PROJECT
 from .util import validate_gcp_string, check_aws_region_for_invalid_characters
 from .util import validate_gke_cluster_name, ElbSupportedPrograms
 from .util import get_query_batch_size
@@ -210,12 +213,15 @@ class CloudProviderBaseConfig:
 @dataclass
 class GCPConfig(CloudProviderBaseConfig, ConfigParserToDataclassMapper):
     """GCP config for ElasticBLAST"""
-    project: GCPString
-    region: GCPString
-    zone: GCPString
+    region: GCPString = GCPString(ELB_DFLT_GCP_REGION)
+    project: GCPString = GCPString(ELB_UNKNOWN_GCP_PROJECT)
+    zone: GCPString = GCPString(ELB_DFLT_GCP_ZONE)
     network: Optional[str] = None
     subnet: Optional[str] = None
     user: Optional[str] = None
+    # FIXME: This is a temporary fix for EB-1530. gke_version should be set to
+    # None once the proper fix is implemented.
+    gke_version: Optional[str] = '1.21'
 
     # mapping to class attributes to ConfigParser parameters so that objects
     # can be initialized from ConfigParser objects
@@ -225,7 +231,8 @@ class GCPConfig(CloudProviderBaseConfig, ConfigParserToDataclassMapper):
                'cloud': None,
                'user': None,
                'network': ParamInfo(CFG_CLOUD_PROVIDER, CFG_CP_GCP_NETWORK),
-               'subnet': ParamInfo(CFG_CLOUD_PROVIDER, CFG_CP_GCP_SUBNETWORK)}
+               'subnet': ParamInfo(CFG_CLOUD_PROVIDER, CFG_CP_GCP_SUBNETWORK),
+               'gke_version': ParamInfo(CFG_CLOUD_PROVIDER, CFG_CP_GCP_GKE_VERSION)}
 
     def __post_init__(self):
         self.cloud = CSP.GCP
@@ -235,6 +242,13 @@ def __post_init__(self):
         if p.stdout:
             self.user = p.stdout.decode('utf-8').rstrip()
 
+        if self.project == ELB_UNKNOWN_GCP_PROJECT:
+            proj = get_gcp_project()
+            if not proj:
+                raise ValueError(f'GCP project is unset, please invoke gcloud config set project REPLACE_WITH_YOUR_PROJECT_NAME_HERE')
+            else:
+                self.project = GCPString(proj)
+
     def validate(self, errors: List[str], task: ElbCommand):
         """Validate config"""
         if bool(self.network) != bool(self.subnet):
@@ -244,7 +258,7 @@ def validate(self, errors: List[str], task: ElbCommand):
 @dataclass
 class AWSConfig(CloudProviderBaseConfig, ConfigParserToDataclassMapper):
     """AWS config for ElasticBLAST"""
-    region: AWSRegion
+    region: AWSRegion = AWSRegion(ELB_DFLT_AWS_REGION)
     vpc: Optional[str] = None
     subnet: Optional[str] = None
     security_group: Optional[str] = None
@@ -589,7 +603,8 @@ def __init__(self, *args, **kwargs):
         # post-init activities
 
         try:
-            self.cloud_provider.region.validate(dry_run)
+            if self.cloud_provider.region:
+                self.cloud_provider.region.validate(dry_run)
         except ValueError as err:
             raise UserReportError(returncode=INPUT_ERROR, message=str(err))
 
@@ -713,7 +728,17 @@ def _init_from_ConfigParser(self, cfg: configparser.ConfigParser,
 
         self._validate_config_parser(cfg)
         _validate_csp(cfg)
+        self.cluster = ClusterConfig.create_from_cfg(cfg)
+
+        # determine cloud provider, first by user config, then results bucket
         if sum([i.startswith('aws') for i in cfg[CFG_CLOUD_PROVIDER]]) > 0:
+            cloud = CSP.AWS
+        elif sum([i.startswith('gcp') for i in cfg[CFG_CLOUD_PROVIDER]]) > 0:
+            cloud = CSP.GCP
+        else:
+            cloud = self.cluster.results.get_cloud_provider()
+
+        if cloud == CSP.AWS:
             self.cloud_provider = AWSConfig.create_from_cfg(cfg)
             # for mypy
             self.aws = cast(AWSConfig, self.cloud_provider)
@@ -722,8 +747,6 @@ def _init_from_ConfigParser(self, cfg: configparser.ConfigParser,
             # for mypy
             self.gcp = cast(GCPConfig, self.cloud_provider)
 
-        self.cluster = ClusterConfig.create_from_cfg(cfg)
-
         if task == ElbCommand.SUBMIT:
             self.blast = BlastConfig.create_from_cfg(cfg)
 
@@ -853,6 +876,13 @@ def validate(self, task: ElbCommand = ElbCommand.SUBMIT, dry_run=False):
                     if instance_props.memory - SYSTEM_MEMORY_RESERVE < bytes_to_cache_gb:
                         errors.append(f'BLAST database {self.blast.db} memory requirements exceed memory available on selected machine type "{self.cluster.machine_type}". Please select machine type with at least {bytes_to_cache_gb + SYSTEM_MEMORY_RESERVE}GB available memory.')
 
+            # validate janitor schedule if provided
+            if ELB_JANITOR_SCHEDULE in os.environ:
+                try:
+                    validate_janitor_schedule(os.environ[ELB_JANITOR_SCHEDULE], self.cloud_provider.cloud)
+                except ValueError as err:
+                    errors.append(str(err))
+
         if errors:
             raise UserReportError(returncode=INPUT_ERROR,
                                   message='\n'.join(errors))
@@ -1108,6 +1138,31 @@ def get_instance_props(cloud_provider: CSP, region: str, machine_type: str) -> I
     return instance_props
 
 
+def validate_janitor_schedule(val: str, cloud_provider: CSP) -> None:
+    """Validate cron schedule for janitor job. Raises ValueError if validation fails."""
+    special = r'@(yearly|annually|monthly|weekly|daily|midnight|hourly)'
+    minute = r'\*|(\*|([1-5]?[0-9]))((,(\*|([1-5]?[0-9])))*([/-][1-5]?[0-9])?)*'
+    hour = r'\*|(\*|([1-2]?[0-9]))((,(\*|([1-2]?[0-9])))*([/-][1-2]?[0-9])?)*'
+    day_of_month_gcp = r'\*|(\*|([1-3]?[0-9]))((,(\*|([1-3]?[0-9])))*([/-][1-3]?[0-9])?)*'
+    day_of_month_aws = r'\*|\?|(\*|([1-3]?[0-9]L?W?))((,(\*|([1-3]?[0-9]L?W?)))*([/-][1-3]?[0-9])?)*'
+    month = r'\*|(\*|(1?[0-9]))((,(\*|(1?[0-9])))*([/-]1?[0-9])?)*'
+    day_of_week_gcp = r'\*|((\*|[0-7]|mon|tue|wed|thu|fri|sat|sun)((,(\*|[0-7]|mon|tue|wed|thu|fri|sat|sun))*([/-]([1-6]|mon|tue|wed|thu|fri|sat|sun))?)*)'
+    day_of_week_aws = r'\*|\?|(((\*|[1-7]|MON|TUE|WED|THU|FRI|SAT|SUN)L?)(([,#](([1-7]|MON|TUE|WED|THU|FRI|SAT|SUN)L?))*([/-]([1-6]|MON|TUE|WED|THU|FRI|SAT|SUN))?)*)'
+    year = r'\*|(\*|(2[01][0-9]{2}))((,(\*|(2[01][0-9]{2})))*(-2[01][0-9]{2})?(/\d{1,3})?)*'
+
+
+    if cloud_provider == CSP.GCP:
+        pattern = special + '|' + '((' + minute + r')\s(' + hour + r')\s(' + day_of_month_gcp + r')\s(' + month + r')\s(' + day_of_week_gcp + '))'
+        url = 'https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/#cron-schedule-syntax'
+    else:
+        pattern = r'cron\((' + minute + r')\s(' + hour + r')\s(' + day_of_month_aws + r')\s(' + month + r')\s(' + day_of_week_aws + r')\s(' + year + r')\)'
+        url = 'https://docs.aws.amazon.com/eventbridge/latest/userguide/eb-create-rule-schedule.html'
+
+    r = re.fullmatch(pattern, val)
+    if r is None:
+        raise ValueError(f'Invalid value of environment variable {ELB_JANITOR_SCHEDULE} "{val}". The string must match the regular expression "{pattern}". For more information, please see {url}')
+
+
 class JSONEnumEncoder(json.JSONEncoder):
     """JSON encoder that handles basic types and Enum"""
     def default(self, o):
@@ -1116,3 +1171,28 @@ def default(self, o):
             return o.name
         else:
             return json.JSONEncoder(self, o)
+
+
+def get_gcp_project() -> Optional[str]:
+    """Return current GCP project or None if the property is unset.
+
+    Raises:
+        util.SafeExecError on problems with command line gcloud
+        RuntimeError if gcloud run is successful, but the result is empty"""
+    cmd: str = 'gcloud config get-value project'
+    p = safe_exec(cmd)
+    result: Optional[str]
+
+    # the result should not be empty, for unset properties gcloud returns the
+    # string: '(unset)' to stderr
+    if not p.stdout and not p.stderr:
+        raise RuntimeError('Current GCP project could not be established')
+
+    result = p.stdout.decode().split('\n')[0]
+
+    # return None if project is unset
+    if result == '(unset)':
+        result = None
+    return result
+
+