Work around CWL content size limit by chunking

Arvados-DCO-1.1-Signed-off-by: Peter Amstutz <peter.amstutz@curii.com>

1 files changed, 27 insertions, 4 deletions

diff --git a/workflows/pangenome-generate/relabel-seqs.cwl b/workflows/pangenome-generate/relabel-seqs.cwl
index 2b780d4..01196f6 100644
--- a/workflows/pangenome-generate/relabel-seqs.cwl
+++ b/workflows/pangenome-generate/relabel-seqs.cwl
@@ -3,6 +3,10 @@ class: CommandLineTool
 inputs:
   readsFA: File[]
   subjects: string[]
+  script:
+    type: File
+    default: {class: File, location: relabel-seqs.py}
+    inputBinding: {}
 outputs:
   relabeledSeqs:
     type: File
@@ -15,11 +19,30 @@ outputs:
 requirements:
   InlineJavascriptRequirement: {}
   InitialWorkDirRequirement:
-    listing:
-      - entry: {$include: relabel-seqs.py}
-        entryname: relabel-seqs.py
+    listing: |
+          ${
+          var i = 0;
+          var b = 1;
+          var out = [];
+          for (; i < inputs.readsFA.length; i++) {
+            var block = [];
+            for (; i < (b*100) && i < inputs.readsFA.length; i++) {
+              block.push(inputs.readsFA[i]);
+            }
+            out.push({
+              entryname: "block"+b,
+              entry: JSON.stringify(block)
+            });
+            b++;
+          }
+          out.push({
+            entry: JSON.stringify(inputs.subjects),
+            entryname: "subjects"
+          });
+          return out;
+          }
 hints:
   DockerRequirement:
     dockerPull: commonworkflowlanguage/cwltool_module
 stdout:
-baseCommand: [python, relabel-seqs.py]
+baseCommand: [python]