Tweak contig pattern to be more compliant to spec

chrovis · Oct 3, 2023 · 6a2d342 · 6a2d342
1 parent 28b7303
commit 6a2d342
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 2 deletions.
diff --git a/src/cljam/io/vcf/util/validator.clj b/src/cljam/io/vcf/util/validator.clj
@@ -41,8 +41,11 @@
 (defn- prep-contig-defs [_validator contigs]
   (into {} (map (juxt :id identity)) contigs))
 
+(def ^:private ^:const contig-pattern
+  #"[0-9A-Za-z!#$%&+./:;?@^_|~-][0-9A-Za-z!#$%&*+./:;=?@^_|~-]*")
+
 (defn- validate-chrom [validator {:keys [chr]}]
-  (if (and (string? chr) (re-matches #"[^<>, \t]+" chr))
+  (if (and (string? chr) (re-matches contig-pattern chr))
     (when-not (get-in validator [:defs :contig chr])
       (error-on-bcf validator [:chr] "Contig %s not defined in meta info" chr))
     (error [:chr]

diff --git a/test/cljam/io/vcf/util/validator_test.clj b/test/cljam/io/vcf/util/validator_test.clj
@@ -36,11 +36,13 @@
     (let [validator (v/map->VCFValidator {})
           prepped (#'v/prep-contig-defs validator
                                         [{:id "1" :length 248956422}
-                                         {:id "2" :length 242193529}])
+                                         {:id "2" :length 242193529}
+                                         {:id "HLA-A*01:01:01:01" :length 3503}])
           validator (assoc validator :defs {:contig prepped})]
       (are [variant expected]
            (= expected (#'v/validate-chrom validator variant))
         {:chr "1"} nil
+        {:chr "HLA-A*01:01:01:01"} nil
 
         {:chr 1}
         {:errors