Work around CWL content size limit by chunking

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

2 files changed, 43 insertions, 10 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]
diff --git a/workflows/pangenome-generate/relabel-seqs.py b/workflows/pangenome-generate/relabel-seqs.py
index 1188ceb..970540f 100644
--- a/workflows/pangenome-generate/relabel-seqs.py
+++ b/workflows/pangenome-generate/relabel-seqs.py
@@ -1,5 +1,15 @@
-reads = $(inputs.readsFA)
-subjects = $(inputs.subjects)
+import os
+import json
+
+reads = []
+b = 1
+while os.path.exists("block%i" % b):
+    with open("block%i" % b) as f:
+        reads.extend(json.load(f))
+    b += 1
+
+with open("subjects") as f:
+    subjects = json.load(f)
 
 relabeled_fasta = open("relabeledSeqs.fasta", "wt")
 original_labels = open("originalLabels.ttl", "wt")
@@ -7,12 +17,12 @@ original_labels = open("originalLabels.ttl", "wt")
 for i, r in enumerate(reads):
     with open(r["path"], "rt") as fa:
         label = fa.readline()
-        original_labels.write("<%s> <http://biohackathon.org/bh20-seq-schema/original_fasta_label> \\"%s\\" .\\n" % (subjects[i], label[1:].strip().replace('"', '\\\\"')))
-        relabeled_fasta.write(">"+subjects[i]+"\\n")
+        original_labels.write("<%s> <http://biohackathon.org/bh20-seq-schema/original_fasta_label> \"%s\" .\n" % (subjects[i], label[1:].strip().replace('"', '\\"')))
+        relabeled_fasta.write(">"+subjects[i]+"\n")
         data = fa.read(8096)
         while data:
             relabeled_fasta.write(data)
-            endswithnewline = data.endswith("\\n")
+            endswithnewline = data.endswith("\n")
             data = fa.read(8096)
         if not endswithnewline:
-            relabeled_fasta.write("\\n")
+            relabeled_fasta.write("\n")