JAL-3121 round trip GFF attributes including map-valued attributes

[jalview.git] / src / jalview / io / vcf / VCFLoader.java

diff --git a/src/jalview/io/vcf/VCFLoader.java b/src/jalview/io/vcf/VCFLoader.java
index ac707d8..cbdd66c 100644 (file)
--- a/src/jalview/io/vcf/VCFLoader.java
+++ b/src/jalview/io/vcf/VCFLoader.java
@@ -19,11 +19,10 @@ import jalview.io.gff.SequenceOntologyI;
 import jalview.util.MapList;
 import jalview.util.MappingUtils;
 import jalview.util.MessageManager;
+import jalview.util.StringUtils;
 
 import java.io.File;
 import java.io.IOException;
-import java.io.UnsupportedEncodingException;
-import java.net.URLDecoder;
 import java.util.ArrayList;
 import java.util.HashMap;
 import java.util.HashSet;
@@ -57,17 +56,7 @@ import htsjdk.variant.vcf.VCFInfoHeaderLine;
  */
 public class VCFLoader
 {
-  private static final String ENCODED_COMMA = "%2C";
-
-  private static final String ENCODED_PERCENT = "%25";
-
-  private static final String ENCODED_EQUALS = "%3D";
-
-  private static final String ENCODED_SEMICOLON = "%3B";
-
-  private static final String ENCODED_COLON = "%3A";
-
-  private static final String UTF_8 = "UTF-8";
+  private static final String VCF_ENCODABLE = ":;=%,";
 
   /*
    * Jalview feature attributes for VCF fixed column data
@@ -1336,42 +1325,17 @@ public class VCFLoader
       String value = getAttributeValue(variant, key, index);
       if (value != null && isValid(variant, key, value))
       {
-        value = decodeSpecialCharacters(value);
+        /*
+         * decode colon, semicolon, equals sign, percent sign, comma (only)
+         * as required by the VCF specification (para 1.2)
+         */
+        value = StringUtils.urlDecode(value, VCF_ENCODABLE);
         addFeatureAttribute(sf, key, value);
       }
     }
   }
 
   /**
-   * Decodes colon, semicolon, equals sign, percent sign, comma to their decoded
-   * form. The VCF specification (para 1.2) requires these to be encoded where not
-   * used with their special meaning in the VCF syntax. Note that general URL
-   * decoding should not be applied, since this would incorrectly decode (for
-   * example) a '+' sign.
-   * 
-   * @param value
-   * @return
-   */
-  protected static String decodeSpecialCharacters(String value)
-  {
-    /*
-     * avoid regex compilation if it is not needed!
-     */
-    if (!value.contains(ENCODED_COLON) && !value.contains(ENCODED_SEMICOLON)
-            && !value.contains(ENCODED_EQUALS)
-            && !value.contains(ENCODED_PERCENT)
-            && !value.contains(ENCODED_COMMA))
-    {
-      return value;
-    }
-
-    value = value.replace(ENCODED_COLON, ":")
-            .replace(ENCODED_SEMICOLON, ";").replace(ENCODED_EQUALS, "=")
-            .replace(ENCODED_PERCENT, "%").replace(ENCODED_COMMA, ",");
-    return value;
-  }
-
-  /**
    * Answers true for '.', null, or an empty value, or if the INFO type is String.
    * If the INFO type is Integer or Float, answers false if the value is not in
    * valid format.
@@ -1489,12 +1453,7 @@ public class VCFLoader
                * VCF spec requires encoding of special characters e.g. '='
                * so decode them here before storing
                */
-              try
-              {
-                field = URLDecoder.decode(field, UTF_8);
-              } catch (UnsupportedEncodingException e)
-              {
-              }
+              field = StringUtils.urlDecode(field, VCF_ENCODABLE);
               csqValues.put(id, field);
             }
           }