more lenient handling of CSQ fields. See #122

CHANGES.md

@@ -2,6 +2,7 @@ v0.2.8
 ======
 + [internal] use same zip library for make-gnotate as gnotate. this should
   improve speed for reading many small chromosomes.
++ don't quit on CSQ/ANN/BCSQ fields that don't have enough information to parse (#122)
 
 v0.2.7
 ======

src/slivarpkg/comphet.nim

@@ -222,8 +222,8 @@ proc main*(dropfirst:bool=false) =
     if f == "": continue
     try:
       var gf:GeneIndexes
-      ivcf.set_csq_fields(f, gf)
-      gene_fields.add(gf)
+      if ivcf.set_csq_fields(f, gf).len > 0:
+        gene_fields.add(gf)
       # add this to the field names so we can clear it as needed
     except KeyError:
       continue

src/slivarpkg/evaluator.nim

@@ -322,8 +322,9 @@ proc newEvaluator*(ivcf:VCF, samples: seq[Sample], groups: seq[Group], float_exp
     try:
       var gf:GeneIndexes
       var fields = ivcf.set_csq_fields(f, gf)
-      result.gene_fields.add(gf)
-      result.VCF[f] = fields
+      if fields.len > 0:
+        result.gene_fields.add(gf)
+        result.VCF[f] = fields
       # add this to the field names so we can clear it as needed
     except KeyError:
       continue

src/slivarpkg/tsv.nim

@@ -111,7 +111,12 @@ proc set_csq_fields*(ivcf:VCF, field:string, gene_fields: var GeneIndexes, csq_c
   var spl = (if "Format: '" in desc: "Format: '" else: "Format: ")
   if spl notin desc:
     spl = ": '"
-  var adesc = desc.split(spl)[1].split("'")[0].strip().strip(chars={'"', '\''}).multiReplace(("[", ""), ("]", ""), ("'", ""), ("*", "")).split("|")
+  var adesc:seq[string]
+  try:
+    adesc = desc.split(spl)[1].split("'")[0].strip().strip(chars={'"', '\''}).multiReplace(("[", ""), ("]", ""), ("'", ""), ("*", "")).split("|")
+  except IndexDefect:
+    # format field description not as expected. return emptyr result and don't fill gene fields
+    return result
 
   for v in adesc.mitems: v = v.toUpperAscii.strip()
   result = adesc
@@ -132,7 +137,8 @@ proc set_csq_fields*(ivcf:VCF, field:string, gene_fields: var GeneIndexes, csq_c
     if gene_fields.transcript != -1: break
 
   if gene_fields.gene == -1:
-    quit &"[slivar] unable to find gene field in {field}"
+    stderr.write_line &"[slivar] warning: found {field} but it did not contain a description that indicated gene field. skipping"
+    raise newException(KeyError, "&[slivar] gene field not found")
   if gene_fields.consequence == -1:
     quit &"[slivar] unable to find consequence field in {field}"
   if gene_fields.transcript == -1:
@@ -268,8 +274,8 @@ or gene->pLI with:
     impact_order = adjustOrder(opts.impact_order.readFile)
 
   if opts.csq_field != "":
-    set_csq_fields(ivcf, opts.csq_field, gene_fields, opts.csq_column)
-    tsv_header.add(["gene", "highest_impact"])
+    if set_csq_fields(ivcf, opts.csq_field, gene_fields, opts.csq_column).len > 0:
+      tsv_header.add(["gene", "highest_impact"])
 
   for f in opts.sample_field:
     doAssert ivcf.header.get(f, BCF_HEADER_TYPE.BCF_HL_INFO)["Type"] == "String"

src/slivarpkg/version.nim

@@ -1,2 +1,2 @@
-const slivarVersion* = "0.2.7"
+const slivarVersion* = "0.2.8"
 const slivarGitCommit* = staticExec("git rev-parse --verify HEAD")