add notes to master metadata to trace samples through pipeline #19 #18 …

…#12
virus-evolution · Jun 16, 2021 · 274accb · 274accb
1 parent f772aba
commit 274accb
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 7 deletions.
diff --git a/modules/extract_protected_sequences.nf b/modules/extract_protected_sequences.nf
@@ -77,10 +77,11 @@ process annotate_metadata {
         writer.writeheader()
         for row in reader:
             note = []
-            if row["sequence_name"] not in seqs:
-                note.append("sample not recent")
             if row["note"]:
-                note.append(row["note"])
+                note.extend(row["note"].split("|"))
+            statement = "sample not recent"
+            if row["sequence_name"] not in seqs and statement not in note:
+                note.append("sample not recent")
             row["note"] = "|".join(note)
             writer.writerow(row)
     """

diff --git a/modules/preprocess.nf b/modules/preprocess.nf
@@ -245,10 +245,11 @@ process annotate_metadata {
         writer.writeheader()
         for row in reader:
             note = []
-            if row["sequence_name"] in tips:
-                note.append("tip in input tree")
             if row["note"]:
-                note.append(row["note"])
+                note.extend(row["note"].split("|"))
+            statement = "tip in input tree"
+            if row["sequence_name"] and statement not in note:
+                note.append(statement)
             row["note"] = "|".join(note)
             writer.writerow(row)
     """

diff --git a/workflows/process_usher_pangolin.nf b/workflows/process_usher_pangolin.nf
@@ -52,7 +52,10 @@ workflow {
         ch_protobuf = build_protobuf.out.protobuf
     }
 
-    ch_protected = extract_protected_sequences(ch_clean_fasta, ch_clean_metadata)
+    extract_protected_sequences(ch_clean_fasta, ch_clean_metadata)
+    ch_protected = extract_protected_sequences.out.fasta
+    ch_protected_metadata = extract_protected_sequences.out.metadata
+
     update_protobuf(ch_protected, ch_protobuf, ch_clean_metadata)
     train_usher_pangolin(update_protobuf.out.protobuf)
 }