Merge pull request #862 from hubmapconsortium/mruffalo/multiome-dag-s…

…nareseq-fix Mruffalo/multiome dag snareseq fix
hubmapconsortium · May 9, 2024 · d8d600a · d8d600a
2 parents cbe9969 + aa2aaaf
commit d8d600a
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 5 deletions.
diff --git a/src/ingest-pipeline/airflow/dags/cwl/portal-containers b/src/ingest-pipeline/airflow/dags/cwl/portal-containers
diff --git a/src/ingest-pipeline/airflow/dags/multiome.py b/src/ingest-pipeline/airflow/dags/multiome.py
@@ -43,6 +43,8 @@
         "pipeline_name",
         "assay_rna",
         "assay_atac",
+        "assay_azimuth",
+        "requires_one_atac_metadata_file",
     ],
 )
 
@@ -123,9 +125,12 @@ def build_cwltool_cmd1(**kwargs):
                     command.append(f"--fastq_dir_{component.lower()}")
                     command.append(data_dir / Path(f"raw/fastq/{component}"))
 
-            for data_dir in data_dirs:
+            atac_metadata_files = [find_atac_metadata_file(data_dir) for data_dir in data_dirs]
+            if params.requires_one_atac_metadata_file:
+                if (count := len(atac_metadata_files)) != 1:
+                    raise ValueError(f"Need 1 ATAC-seq metadata file, found {count}")
                 command.append("--atac_metadata_file")
-                command.append(find_atac_metadata_file(data_dir))
+                command.append(atac_metadata_files[0])
 
             return join_quote_command_str(command)
 
@@ -150,7 +155,7 @@ def build_cwltool_cmd2(**kwargs):
                 "--secondary-analysis-matrix",
                 "secondary_analysis.h5mu",
                 "--assay",
-                params.assay_rna,
+                params.assay_azimuth,
             ]
 
             return join_quote_command_str(command)
@@ -339,15 +344,19 @@ def get_simple_multiome_dag_params(assay: str) -> MultiomeSequencingDagParameter
         pipeline_name=f"multiome-{assay}",
         assay_rna=assay,
         assay_atac=assay,
+        assay_azimuth=assay,
+        requires_one_atac_metadata_file=False,
     )
 
 
 multiome_dag_params: List[MultiomeSequencingDagParameters] = [
     MultiomeSequencingDagParameters(
         dag_id="multiome_10x",
         pipeline_name="multiome-10x",
-        assay_rna="10x_v3_sn",
+        assay_rna="multiome_10x",
         assay_atac="multiome_10x",
+        assay_azimuth="10x_v3_sn",
+        requires_one_atac_metadata_file=True,
     ),
     get_simple_multiome_dag_params("snareseq"),
 ]