rename method and update calls in base

microbiomedata · Jan 22, 2024 · b09cdcf · b09cdcf
1 parent 2005ba6
commit b09cdcf
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 10 deletions.
diff --git a/nmdc_automation/re_iding/base.py b/nmdc_automation/re_iding/base.py
@@ -23,8 +23,8 @@
                                                get_data_object_record_by_id,
                                                get_omics_processing_id)
 from nmdc_automation.re_iding.file_utils import (find_data_object_type,
-                                                compute_new_paths,
-                                                assembly_file_operations)
+                                                 compute_new_paths_and_link,
+                                                 assembly_file_operations)
 
 NAPA_TEMPLATE = "../../../configs/re_iding_worklfows.yaml"
 DATA_BASE_URL = "https://data.microbiomedata.org/data"
@@ -165,7 +165,7 @@ def update_reads_qc_analysis_activity_set(self, db_record: Dict,
                 logger.info(f"old_do_id: {old_do_id}")
                 old_do_rec = get_data_object_record_by_id(db_record, old_do_id)
                 data_object_type = find_data_object_type(old_do_rec)
-                new_file_path = compute_new_paths(
+                new_file_path = compute_new_paths_and_link(
                 old_do_rec["url"], new_readsqc_base_dir, new_activity_id, self.data_dir
                 )
                 logging.info(f"New file path computed for {data_object_type}: {new_file_path}")
@@ -223,7 +223,7 @@ def update_metagenome_assembly_set(self, db_record: Dict,
                 data_object_type = find_data_object_type(old_do_rec)
                 if not data_object_type:
                     continue
-                new_file_path = compute_new_paths(old_do_rec["url"],new_assembly_base_dir, new_activity_id)
+                new_file_path = compute_new_paths_and_link(old_do_rec["url"], new_assembly_base_dir, new_activity_id)
                 updated_md5, updated_file_size = assembly_file_operations(
                 old_do_rec, data_object_type, new_file_path, new_activity_id,
                     self.data_dir)
@@ -286,7 +286,7 @@ def update_read_based_taxonomy_analysis_activity_set(self, db_record: Dict,
                 data_object_type = find_data_object_type(old_do_rec)
                 if not data_object_type:
                     continue
-                new_file_path = compute_new_paths(
+                new_file_path = compute_new_paths_and_link(
                 old_do_rec["url"], new_readbased_base_dir, new_activity_id, self.data_dir
                 )
                 logging.info(f"New file path computed for {data_object_type}: {new_file_path}")
@@ -360,12 +360,11 @@ def update_metatranscriptome_activity_set(self, db_record: Dict,
                 if not data_object_type:
                     logger.warning(f"Data object type not found for {old_do_id}")
                     # continue
-
-                new_file_path = compute_new_paths(old_do_rec["url"],new_metatranscriptome_base_dir, new_activity_id)
+                # link data object to new location
+                new_file_path = compute_new_paths_and_link(
+                    old_do_rec["url"], new_metatranscriptome_base_dir, new_activity_id, self.data_dir)
                 logging.info(f"New file path computed for {data_object_type}: {new_file_path}")
 
-                # TODO: determine file operations if any, similar to assembly_file_operations
-
                 new_do = self.make_new_data_object(
                     omics_processing_id, activity_type, new_activity_id, old_do_rec, data_object_type
                 )

diff --git a/nmdc_automation/re_iding/file_utils.py b/nmdc_automation/re_iding/file_utils.py
@@ -191,7 +191,7 @@ def assembly_file_operations(data_object_record, data_object_type,
     return md5, size
 
 
-def compute_new_paths(
+def compute_new_paths_and_link(
         old_url: str,
         new_base_dir: Union[str, os.PathLike],
         act_id: str,