JAL-3365 expand range of allowed DSSP secondary structure symbols in Stockholm files

[jalview.git] / src / jalview / io / StockholmFile.java

diff --git a/src/jalview/io/StockholmFile.java b/src/jalview/io/StockholmFile.java
index 8b26757..5d645ca 100644 (file)
--- a/src/jalview/io/StockholmFile.java
+++ b/src/jalview/io/StockholmFile.java
@@ -31,6 +31,7 @@ import java.util.Enumeration;
 import java.util.Hashtable;
 import java.util.LinkedHashMap;
 import java.util.List;
+import java.util.Locale;
 import java.util.Map;
 import java.util.Vector;
 
@@ -55,8 +56,6 @@ import jalview.util.DBRefUtils;
 import jalview.util.Format;
 import jalview.util.MessageManager;
 
-// import org.apache.log4j.*;
-
 /**
  * This class is supposed to parse a Stockholm format file into Jalview There
  * are TODOs in this class: we do not know what the database source and version
@@ -77,20 +76,21 @@ public class StockholmFile extends AlignFile
 {
   private static final String ANNOTATION = "annotation";
 
-//  private static final Regex OPEN_PAREN = new Regex("(<|\\[)", "(");
-//
-//  private static final Regex CLOSE_PAREN = new Regex("(>|\\])", ")");
+  // private static final Regex OPEN_PAREN = new Regex("(<|\\[)", "(");
+  //
+  // private static final Regex CLOSE_PAREN = new Regex("(>|\\])", ")");
 
   public static final Regex DETECT_BRACKETS = new Regex(
           "(<|>|\\[|\\]|\\(|\\)|\\{|\\})");
 
-  // WUSS extended symbols. Avoid ambiguity with protein SS annotations by using NOT_RNASS first.
+  // WUSS extended symbols. Avoid ambiguity with protein SS annotations by using
+  // NOT_RNASS first.
   public static final String RNASS_BRACKETS = "<>[](){}AaBbCcDdEeFfGgHhIiJjKkLlMmNnOoPpQqRrSsTtUuVvWwXxYyZz";
 
   // use the following regex to decide an annotations (whole) line is NOT an RNA
   // SS (it contains only E,H,e,h and other non-brace/non-alpha chars)
   private static final Regex NOT_RNASS = new Regex(
-          "^[^<>[\\](){}A-DF-Za-df-z]*$");
+          "^[^<>[\\](){}ADFJ-RUVWYZadfj-ruvwyz]*$");
 
   StringBuffer out; // output buffer
 
@@ -240,8 +240,8 @@ public class StockholmFile extends AlignFile
     Regex openparen = new Regex("(<|\\[)", "(");
     Regex closeparen = new Regex("(>|\\])", ")");
 
-//    // Detect if file is RNA by looking for bracket types
-//    Regex detectbrackets = new Regex("(<|>|\\[|\\]|\\(|\\))");
+    // // Detect if file is RNA by looking for bracket types
+    // Regex detectbrackets = new Regex("(<|>|\\[|\\]|\\(|\\))");
 
     rend.optimize();
     p.optimize();
@@ -337,7 +337,8 @@ public class StockholmFile extends AlignFile
             if (dbr != null)
             {
               // we could get very clever here - but for now - just try to
-              // guess accession type from type of sequence, source of alignment plus
+              // guess accession type from type of sequence, source of alignment
+              // plus
               // structure
               // of accession
               guessDatabaseFor(seqO, dbr, dbsource);
@@ -525,8 +526,10 @@ public class StockholmFile extends AlignFile
               treeName = an.stringMatched(2);
               treeString = new StringBuffer();
             }
-            // TODO: JAL-3532 - this is where GF comments and database references are lost
-            // suggest overriding this method for Stockholm files to catch and properly
+            // TODO: JAL-3532 - this is where GF comments and database
+            // references are lost
+            // suggest overriding this method for Stockholm files to catch and
+            // properly
             // process CC, DR etc into multivalued properties
             setAlignmentProperty(an.stringMatched(1), an.stringMatched(2));
           }
@@ -758,7 +761,8 @@ public class StockholmFile extends AlignFile
     }
     if (dbsource == null)
     {
-      // make up an origin based on whether the sequence looks like it is nucleotide
+      // make up an origin based on whether the sequence looks like it is
+      // nucleotide
       // or protein
       dbsource = (seqO.isProtein()) ? "PFAM" : "RFAM";
     }
@@ -942,6 +946,7 @@ public class StockholmFile extends AlignFile
     return ref.getSource().toString() + " ; "
             + ref.getAccessionId().toString();
   }
+
   @Override
   public String print(SequenceI[] s, boolean jvSuffix)
   {
@@ -1110,7 +1115,7 @@ public class StockholmFile extends AlignFile
         }
         else
         {
-          key = type2id(aa.label.toLowerCase());
+          key = type2id(aa.label.toLowerCase(Locale.ROOT));
           if (key == null)
           {
             label = aa.label;