Merge pull request #140 from tcezard/EVA-3207_limit_clustering

EVA-3207 - Prevent remapping/clustering when the target genome is from a different species

eva_submission/eload_ingestion.py

@@ -10,6 +10,7 @@
 from ebi_eva_common_pyutils.config import cfg
 from ebi_eva_common_pyutils.config_utils import get_mongo_uri_for_eva_profile, get_primary_mongo_creds_for_profile, \
     get_accession_pg_creds_for_profile, get_count_service_creds_for_profile
+from ebi_eva_common_pyutils.ena_utils import get_assembly_name_and_taxonomy_id
 from ebi_eva_common_pyutils.metadata_utils import resolve_variant_warehouse_db_name, insert_new_assembly_and_taxonomy, \
     get_assembly_set_from_metadata
 from ebi_eva_common_pyutils.pg_utils import get_all_results_for_query, execute_query
@@ -90,7 +91,8 @@ def ingest(
 
         if 'optional_remap_and_cluster' in tasks:
             target_assembly = self._get_target_assembly()
-            if target_assembly:
+            # EVA-3207: Temporary limitation while we sort out the remapping across species
+            if target_assembly and self._target_assembly_from_same_taxonomy(target_assembly):
                 self.run_remap_and_cluster_workflow(target_assembly, resume=resume)
 
         if do_variant_load or annotation_only:
@@ -399,6 +401,15 @@ def _get_target_assembly(self):
         self.warning(f'Could not find any current supported assembly for {self.taxonomy}, skipping clustering')
         return None
 
+    def _target_assembly_from_same_taxonomy(self, target_assembly):
+        # Find taxonomy of the target assembly
+        _, taxonomy_id_from_target = get_assembly_name_and_taxonomy_id(target_assembly)
+        if int(taxonomy_id_from_target) != int(self.taxonomy):
+            self.warning(f'Target assembly {target_assembly} is from a different taxonomy {taxonomy_id_from_target} '
+                         f'compared to the current project {self.taxonomy}. Therefore remapping will not be carried out!')
+            return False
+        return True
+
     def create_extraction_properties(self, output_file_path, taxonomy):
         properties = self.properties_generator.get_remapping_extraction_properties(
             taxonomy=taxonomy,

tests/test_eload_ingestion.py

@@ -151,11 +151,13 @@ def test_ingest_all_tasks(self):
                 patch('eva_submission.eload_utils.get_all_results_for_query') as m_get_alias_results, \
                 patch('eva_submission.eload_ingestion.get_vep_and_vep_cache_version') as m_get_vep_versions, \
                 patch('eva_submission.eload_utils.requests.post') as m_post, \
+                patch('eva_submission.eload_ingestion.get_assembly_name_and_taxonomy_id') as m_get_tax, \
                 self._patch_mongo_database():
             m_get_alias_results.return_value = [['alias']]
             m_get_vep_versions.return_value = (100, 100, 'homo_sapiens')
             m_post.return_value.text = self.get_mock_result_for_ena_date()
             m_get_results.side_effect = default_db_results_for_ingestion()
+            m_get_tax.return_value = ('name', '9090')
             self.eload.ingest(1)
 
     def test_ingest_metadata_load(self):
@@ -370,8 +372,10 @@ def test_ingest_clustering(self):
         with self._patch_metadata_handle(), \
                 patch('eva_submission.eload_ingestion.get_all_results_for_query') as m_get_results, \
                 patch('eva_submission.eload_ingestion.command_utils.run_command_with_output', autospec=True) as m_run_command, \
+                patch('eva_submission.eload_ingestion.get_assembly_name_and_taxonomy_id') as m_get_tax, \
                 self._patch_mongo_database():
             m_get_results.side_effect = default_db_results_for_clustering()
+            m_get_tax.return_value = ('name', '9796')
             self.eload.ingest(tasks=['optional_remap_and_cluster'])
             assert self.eload.eload_cfg.query('ingestion', 'remap_and_cluster', 'target_assembly') == 'GCA_123'
             assert m_run_command.call_count == 1
@@ -394,6 +398,7 @@ def test_resume_when_step_fails(self):
                 patch('eva_submission.eload_utils.get_all_results_for_query') as m_get_alias_results, \
                 patch('eva_submission.eload_ingestion.get_vep_and_vep_cache_version') as m_get_vep_versions, \
                 patch('eva_submission.eload_utils.requests.post') as m_post, \
+                patch('eva_submission.eload_ingestion.get_assembly_name_and_taxonomy_id') as m_get_tax, \
                 self._patch_mongo_database():
             m_get_alias_results.return_value = [['alias']]
             m_get_vep_versions.return_value = (100, 100, 'homo_sapiens')
@@ -408,6 +413,7 @@ def test_resume_when_step_fails(self):
                 None,  # clustering
                 None,  # variant load
             ]
+            m_get_tax.return_value = ('name', '9090')
 
             with self.assertRaises(subprocess.CalledProcessError):
                 self.eload.ingest()
@@ -425,11 +431,13 @@ def test_resume_completed_job(self):
                 patch('eva_submission.eload_utils.get_all_results_for_query') as m_get_alias_results, \
                 patch('eva_submission.eload_ingestion.get_vep_and_vep_cache_version') as m_get_vep_versions, \
                 patch('eva_submission.eload_utils.requests.post') as m_post, \
+                patch('eva_submission.eload_ingestion.get_assembly_name_and_taxonomy_id') as m_get_tax, \
                 self._patch_mongo_database():
             m_get_alias_results.return_value = [['alias']]
             m_get_vep_versions.return_value = (100, 100, 'homo_sapiens')
             m_post.return_value.text = self.get_mock_result_for_ena_date()
             m_get_results.side_effect = default_db_results_for_ingestion() + default_db_results_for_ingestion()
+            m_get_tax.return_value = ('name', '9796')
 
             # Resuming with no existing job execution is fine
             self.eload.ingest(resume=True)