8 files changed, 191 insertions, 13 deletions
diff --git a/workflows/pangenome-generate/merge-metadata.cwl b/workflows/pangenome-generate/merge-metadata.cwl
index 9164c09..4d9c808 100644
--- a/workflows/pangenome-generate/merge-metadata.cwl
+++ b/workflows/pangenome-generate/merge-metadata.cwl
@@ -5,14 +5,48 @@ hints:
     dockerPull: commonworkflowlanguage/cwltool_module
 inputs:
   metadata: File[]
-  metadataSchema: File
   subjects: string[]
+  metadataSchema:
+    type: File
+    inputBinding: {position: 2}
+  originalLabels:
+    type: File
+    inputBinding: {position: 3}
+  dups:
+    type: File?
+    inputBinding: {position: 4}
+  script:
+    type: File
+    inputBinding: {position: 1}
+    default: {class: File, location: merge-metadata.py}
 outputs:
   merged: stdout
 stdout: mergedmetadata.ttl
 requirements:
+  InlineJavascriptRequirement: {}
   InitialWorkDirRequirement:
-    listing:
-      - entry: {$include: merge-metadata.py}
-        entryname: merge-metadata.py
-baseCommand: [python3, merge-metadata.py]
+    listing: |
+          ${
+          var i = 0;
+          var b = 1;
+          var out = [];
+          for (; i < inputs.metadata.length; i++) {
+            var block = [];
+            var sub = [];
+            for (; i < (b*150) && i < inputs.metadata.length; i++) {
+              block.push(inputs.metadata[i]);
+              sub.push(inputs.subjects[i]);
+            }
+            out.push({
+              entryname: "block"+b,
+              entry: JSON.stringify(block)
+            });
+            out.push({
+              entryname: "subs"+b,
+              entry: JSON.stringify(sub)
+            });
+            b++;
+          }
+          return out;
+          }
+baseCommand: python
diff --git a/workflows/pangenome-generate/merge-metadata.py b/workflows/pangenome-generate/merge-metadata.py
index 64275b1..65d08a6 100644
--- a/workflows/pangenome-generate/merge-metadata.py
+++ b/workflows/pangenome-generate/merge-metadata.py
@@ -1,9 +1,28 @@
+import re
 import schema_salad.schema
 import schema_salad.jsonld_context
+import json
+import sys
+import os
+import logging
 
-metadataSchema = '$(inputs.metadataSchema.path)'
-metadata = $(inputs.metadata)
-subjects = $(inputs.subjects)
+metadataSchema = sys.argv[1]
+originalLabels = sys.argv[2]
+dups = None
+if len(sys.argv) == 4:
+    dups = sys.argv[3]
+
+def readitems(stem):
+    items = []
+    b = 1
+    while os.path.exists("%s%i" % (stem, b)):
+        with open("%s%i" % (stem, b)) as f:
+            items.extend(json.load(f))
+        b += 1
+    return items
+
+metadata = readitems("block")
+subjects = readitems("subs")
 
 (document_loader,
  avsc_names,
@@ -11,7 +30,20 @@ subjects = $(inputs.subjects)
  metaschema_loader) = schema_salad.schema.load_schema(metadataSchema)
 
 for i, m in enumerate(metadata):
-    doc, metadata = schema_salad.schema.load_and_validate(document_loader, avsc_names, m["path"], True)
+    doc, metadata = schema_salad.schema.load_and_validate(document_loader, avsc_names, m["path"], False, False)
     doc["id"] = subjects[i]
     g = schema_salad.jsonld_context.makerdf(subjects[i], doc, document_loader.ctx)
     print(g.serialize(format="ntriples").decode("utf-8"))
+
+if dups:
+    sameseqs = open(dups, "rt")
+    for d in sameseqs:
+        logging.warn(d)
+        g = re.match(r"\d+\t(.*)", d)
+        logging.warn("%s", g.group(1))
+        sp = g.group(1).split(",")
+        for n in sp[1:]:
+            print("<%s> <http://biohackathon.org/bh20-seq-schema/has_duplicate_sequence> <%s> ." % (n.strip(), sp[0].strip()))
+
+orig = open(originalLabels, "rt")
+print(orig.read())
diff --git a/workflows/pangenome-generate/minimap2.cwl b/workflows/pangenome-generate/minimap2.cwl
index bf19ef7..42d1dce 100644
--- a/workflows/pangenome-generate/minimap2.cwl
+++ b/workflows/pangenome-generate/minimap2.cwl
@@ -12,7 +12,7 @@ hints:
   ResourceRequirement:
     coresMin: 8
     coresMax: 32
-    ramMin: $(7 * 1024)
+    ramMin: $(9 * 1024)
     outdirMin: $(Math.ceil(inputs.readsFA.size/(1024*1024*1024) + 20))
 stdout: $(inputs.readsFA.nameroot).paf
 baseCommand: minimap2
diff --git a/workflows/pangenome-generate/pangenome-generate.cwl b/workflows/pangenome-generate/pangenome-generate.cwl
index 2710743..6794e2d 100644
--- a/workflows/pangenome-generate/pangenome-generate.cwl
+++ b/workflows/pangenome-generate/pangenome-generate.cwl
@@ -18,13 +18,22 @@ outputs:
   odgiRDF:
     type: File
     outputSource: odgi2rdf/rdf
+  readsMergeDedup:
+    type: File
+    outputSource: dedup/readsMergeDedup
   mergedMetadata:
     type: File
     outputSource: mergeMetadata/merged
 steps:
+  relabel:
+    in:
+      readsFA: inputReads
+      subjects: subjects
+    out: [relabeledSeqs, originalLabels]
+    run: relabel-seqs.cwl
   dedup:
-    in: {readsFA: inputReads}
-    out: [readsMergeDedup]
+    in: {readsFA: relabel/relabeledSeqs}
+    out: [readsMergeDedup, dups]
     run: seqkit-rmdup.cwl
   overlapReads:
     in: {readsFA: dedup/readsMergeDedup}
@@ -53,5 +62,7 @@ steps:
       metadata: metadata
       metadataSchema: metadataSchema
       subjects: subjects
+      dups: dedup/dups
+      originalLabels: relabel/originalLabels
     out: [merged]
     run: merge-metadata.cwl
diff --git a/workflows/pangenome-generate/relabel-seqs.cwl b/workflows/pangenome-generate/relabel-seqs.cwl
new file mode 100644
index 0000000..c1f17a4
--- /dev/null
+++ b/workflows/pangenome-generate/relabel-seqs.cwl
@@ -0,0 +1,50 @@
+cwlVersion: v1.1
+class: CommandLineTool
+inputs:
+  readsFA: File[]
+  subjects: string[]
+  script:
+    type: File
+    default: {class: File, location: relabel-seqs.py}
+    inputBinding: {}
+outputs:
+  relabeledSeqs:
+    type: File
+    outputBinding:
+      glob: relabeledSeqs.fasta
+  originalLabels:
+    type: File
+    outputBinding:
+      glob: originalLabels.ttl
+requirements:
+  InlineJavascriptRequirement: {}
+  InitialWorkDirRequirement:
+    listing: |
+          ${
+          var i = 0;
+          var b = 1;
+          var out = [];
+          for (; i < inputs.readsFA.length; i++) {
+            var block = [];
+            var sub = [];
+            for (; i < (b*150) && i < inputs.readsFA.length; i++) {
+              block.push(inputs.readsFA[i]);
+              sub.push(inputs.subjects[i]);
+            }
+            out.push({
+              entryname: "block"+b,
+              entry: JSON.stringify(block)
+            });
+            out.push({
+              entryname: "subs"+b,
+              entry: JSON.stringify(sub)
+            });
+            b++;
+          }
+          return out;
+          }
+hints:
+  DockerRequirement:
+    dockerPull: commonworkflowlanguage/cwltool_module
+stdout:
+baseCommand: [python]
diff --git a/workflows/pangenome-generate/relabel-seqs.py b/workflows/pangenome-generate/relabel-seqs.py
new file mode 100644
index 0000000..6b022a0
--- /dev/null
+++ b/workflows/pangenome-generate/relabel-seqs.py
@@ -0,0 +1,30 @@
+import os
+import json
+
+def readitems(stem):
+    items = []
+    b = 1
+    while os.path.exists("%s%i" % (stem, b)):
+        with open("%s%i" % (stem, b)) as f:
+            items.extend(json.load(f))
+        b += 1
+    return items
+
+reads = readitems("block")
+subjects = readitems("subs")
+
+relabeled_fasta = open("relabeledSeqs.fasta", "wt")
+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")
+        data = fa.read(8096)
+        while data:
+            relabeled_fasta.write(data)
+            endswithnewline = data.endswith("\n")
+            data = fa.read(8096)
+        if not endswithnewline:
+            relabeled_fasta.write("\n")
diff --git a/workflows/pangenome-generate/seqkit-rmdup.cwl b/workflows/pangenome-generate/seqkit-rmdup.cwl
index d3626f5..071fa66 100644
--- a/workflows/pangenome-generate/seqkit-rmdup.cwl
+++ b/workflows/pangenome-generate/seqkit-rmdup.cwl
@@ -1,12 +1,16 @@
 cwlVersion: v1.1
 class: CommandLineTool
 inputs:
-  readsFA: File[]
+  readsFA: File
 outputs:
   readsMergeDedup:
     type: File
     outputBinding:
       glob: readsMergeDedup.fasta
+  dups:
+    type: File?
+    outputBinding:
+      glob: dups.txt
 requirements:
   InlineJavascriptRequirement: {}
 hints:
@@ -28,5 +32,6 @@ baseCommand: seqkit
 arguments: [rmdup,
             --by-seq,
             --ignore-case,
+            --dup-num-file, dups.txt,
             -o, readsMergeDedup.fasta,
             $(inputs.readsFA)]
diff --git a/workflows/pangenome-generate/testjob.yml b/workflows/pangenome-generate/testjob.yml
new file mode 100644
index 0000000..a48aff8
--- /dev/null
+++ b/workflows/pangenome-generate/testjob.yml
@@ -0,0 +1,16 @@
+inputReads:
+  - class: File
+    location: ../../example/sequence.fasta
+  - class: File
+    location: ../../example/sequence.fasta
+metadata:
+  - class: File
+    location: ../../example/metadata.yaml
+  - class: File
+    location: ../../example/metadata.yaml
+metadataSchema:
+  class: File
+  location: ../../bh20sequploader/bh20seq-schema.yml
+subjects:
+  - http://arvados.org/keep/seq1
+  - http://arvados.org/keep/seq2