Rename function and add filter by release, taxonomy and assemblies

eva-accession-release-automation/run_release_in_embassy/release_metadata.py

@@ -92,23 +92,30 @@ def get_release_inventory_info_for_assembly(taxonomy_id, assembly_accession, rel
     return results[0][0]
 
 
-def get_release_pending(release_species_inventory_table, metadata_connection_handle, status=None):
-    def format_status(status_list):
-        return f"({str(status_list).strip('[]')})"
-
-    if 'Pending' in status:
-        status.remove('Pending')
-        if status:
-            status_statement = f"and (release_status in {format_status(status)} or release_status is null)"
+def get_release_for_status_and_version(release_species_inventory_table, metadata_connection_handle, status=None,
+                                       release_version=None, taxonomy_id=None, assembly_accessions=None):
+    def format_list(list_to_format):
+        return f"({str(list_to_format).strip('[]')})"
+
+    if status:
+        if 'Pending' in status:
+            status.remove('Pending')
+            if status:
+                status_statement = f"and (release_status in {format_list(status)} or release_status is null)"
+            else:
+                status_statement = f"and release_status is null"
         else:
-            status_statement = f"and release_status is null"
+            status_statement = f"and release_status in {format_list(status)}"
     else:
-        status_statement = f"and release_status in {format_status(status)}"
+        status_statement = None
 
     query = (f"select taxonomy, assembly_accession, release_version from {release_species_inventory_table} "
-              "where should_be_released "
-              "and num_rs_to_release > 0 "
-              f"{status_statement}" if status_statement else ''
-              "ORDER BY release_version, taxonomy, assembly_accession")
+             f"where should_be_released "
+             f"and num_rs_to_release > 0 "
+             f"{status_statement}" if status_statement else ''
+             f"and release_version={release_version} " if release_version else ''
+             f"and taxonomy={taxonomy_id} " if taxonomy_id else ''
+             f"and assembly_accessions in {format_list(assembly_accessions)} " if assembly_accessions else ''
+             "ORDER BY release_version, taxonomy, assembly_accession")
     results = get_all_results_for_query(metadata_connection_handle, query)
     yield results

eva-accession-release-automation/run_release_in_embassy/run_release_for_species.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+import sys
 from argparse import ArgumentParser
 
 import os
@@ -22,7 +23,7 @@
 from ebi_eva_common_pyutils.logger import logging_config
 from ebi_eva_internal_pyutils.metadata_utils import get_metadata_connection_handle
 from ebi_eva_common_pyutils.config import cfg
-from run_release_in_embassy.release_metadata import get_release_assemblies_for_taxonomy, get_release_pending
+from run_release_in_embassy.release_metadata import get_release_assemblies_for_taxonomy, get_release_for_status_and_version
 from run_release_in_embassy.release_common_utils import get_release_folder_name
 
 
@@ -106,15 +107,17 @@ def run_release_for_species(taxonomy_id, release_assemblies, release_version, re
             run_command_with_output(f"Running workflow file {workflow_file_path}", workflow_command)
 
 
-def list_pending_release(status):
+def list_pending_release(status, release_version, taxonomy_id, assembly_accessions):
     private_config_xml_file = cfg.query("maven", "settings_file")
     profile = cfg.query("maven", "environment")
     release_species_inventory_table = cfg.query('release', 'inventory_table')
     with get_metadata_connection_handle(profile, private_config_xml_file) as metadata_connection_handle:
         header = ['taxonomy', 'assembly_accession', 'release_version']
         table = []
-        for taxonomy, assembly_accession, release_version in get_release_pending(release_species_inventory_table,
-                                                                                 metadata_connection_handle, status):
+        for taxonomy, assembly_accession, release_version in get_release_for_status_and_version(
+                release_species_inventory_table, metadata_connection_handle, status, release_version,
+                taxonomy_id, assembly_accessions
+        ):
             table.append((taxonomy, assembly_accession, release_version))
         pretty_print(header, table)
 
@@ -129,20 +132,27 @@ def load_config(*args):
 
 def main():
     argparse = ArgumentParser()
-    argparse.add_argument("--list", nargs='+',
+    argparse.add_argument_group()
+    argparse.add_argument("--list_status", nargs='+',
                           help="Generate the list of species and assembly that needs to be released",
                           choices=['Pending', 'Started', 'Completed'])
-    argparse.add_argument("--taxonomy_id", help="ex: 9913", required=True)
+    argparse.add_argument("--taxonomy_id", help="ex: 9913")
     argparse.add_argument("--assembly_accessions", nargs='+', help="ex: GCA_000003055.3")
-    argparse.add_argument("--release_version", required=True)
-    argparse.add_argument("--resume", default=False, required=False,
+    argparse.add_argument("--release_version")
+    argparse.add_argument("--resume", default=False,
                           help="Resume the nextflow pipeline for the specified taxonomy and assembly")
     args = argparse.parse_args()
     load_config()
     logging_config.add_stdout_handler()
-    if args.list:
-        list_pending_release(args.list)
+    if args.list_status:
+        list_pending_release(args.list, args.release_version, args.taxonomy_id, args.assembly_accessions)
     else:
+        if not args.taxonomy_id:
+            logger.error('--taxonomy_id is required when running the release')
+            sys.exit(1)
+        if not args.release_version:
+            logger.error('--release_version is required when running the release')
+            sys.exit(1)
         run_release_for_species(args.taxonomy_id, args.assembly_accessions, args.release_version, args.resume)