From d16edc08a4fee1810142a3ee9d5a9ef3979219e4 Mon Sep 17 00:00:00 2001
From: Jennifer Chang <jennifer.chang.bioinform@gmail.com>
Date: Fri, 5 Apr 2024 11:21:54 -0700
Subject: [PATCH] fixup: move hard-coded columns to a shared workflow variable
 or config params so they don't get out of sync between rules

---
 ingest/defaults/config.yaml | 4 +++-
 ingest/rules/nextclade.smk  | 9 ++++++---
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/ingest/defaults/config.yaml b/ingest/defaults/config.yaml
index d65f8d5a..4afafebc 100644
--- a/ingest/defaults/config.yaml
+++ b/ingest/defaults/config.yaml
@@ -110,4 +110,6 @@ nextclade:
   min_length: 1000 # E gene length is approximately 1400nt
   min_seed_cover: 0.1
   nextclade_field: 'nextclade_subtype'
-  gene: ['E','NS1']
\ No newline at end of file
+  gene: ['E','NS1']
+  input_nextclade_fields: "seqName,clade,alignmentStart,alignmentEnd,coverage,failedCdses"
+  output_nextclade_fields: "alignmentStart,alignmentEnd,genome_coverage,failedCdses"
\ No newline at end of file
diff --git a/ingest/rules/nextclade.smk b/ingest/rules/nextclade.smk
index e153c418..32a60a7d 100644
--- a/ingest/rules/nextclade.smk
+++ b/ingest/rules/nextclade.smk
@@ -63,13 +63,15 @@ rule concat_nextclade_subtype_results:
     params:
         id_field=config["curate"]["id_field"],
         nextclade_field=config["nextclade"]["nextclade_field"],
+        input_nextclade_fields=config["nextclade"]["input_nextclade_fields"],
+        output_nextclade_fields=config["nextclade"]["output_nextclade_fields"],
     shell:
         """
-        echo "{params.id_field},{params.nextclade_field},alignmentStart,alignmentEnd,genome_coverage,failedCdses" \
+        echo "{params.id_field},{params.nextclade_field},{params.output_nextclade_fields}" \
         | tr ',' '\t' \
         > {output.nextclade_subtypes}
 
-        tsv-select -H -f "seqName,clade,alignmentStart,alignmentEnd,coverage,failedCdses" {input.nextclade_results_files} \
+        tsv-select -H -f "{params.input_nextclade_fields}" {input.nextclade_results_files} \
         | awk 'NR>1 {{print}}' \
         >> {output.nextclade_subtypes}
         """
@@ -86,12 +88,13 @@ rule append_nextclade_columns:
     params:
         id_field=config["curate"]["id_field"],
         nextclade_field=config["nextclade"]["nextclade_field"],
+        output_nextclade_fields=config["nextclade"]["output_nextclade_fields"],
     shell:
         """
         tsv-join -H \
             --filter-file {input.nextclade_subtypes} \
             --key-fields {params.id_field} \
-            --append-fields {params.nextclade_field},alignmentStart,alignmentEnd,genome_coverage,failedCdses \
+            --append-fields {params.nextclade_field},{params.output_nextclade_fields} \
             --write-all ? \
             {input.metadata} \
         > {output.metadata_all}