this.seqs = seqs

[jalview.git] / src / jalview / analysis / SequenceIdMatcher.java

diff --git a/src/jalview/analysis/SequenceIdMatcher.java b/src/jalview/analysis/SequenceIdMatcher.java
index 1efe498..9d33996 100755 (executable)
--- a/src/jalview/analysis/SequenceIdMatcher.java
+++ b/src/jalview/analysis/SequenceIdMatcher.java
@@ -26,7 +26,8 @@ import jalview.datamodel.*;
  * <p>Title: </p>\r
  * SequenceIdMatcher\r
  * <p>Description: </p>\r
- * Routine which does approximate Sequence Id resolution by name using string containment rather than equivalence\r
+ * Routine which does approximate Sequence Id resolution by name using\r
+ * string containment (on word boundaries) rather than equivalence\r
  * <p>Copyright: Copyright (c) 2004</p>\r
  *\r
  * <p>Company: Dundee University</p>\r
@@ -41,7 +42,6 @@ public class SequenceIdMatcher
   public SequenceIdMatcher(SequenceI[] seqs)\r
   {\r
     names = new Hashtable();\r
-\r
     for (int i = 0; i < seqs.length; i++)\r
     {\r
       names.put(new SeqIdName(seqs[i].getName()), seqs[i]);\r
@@ -83,13 +83,13 @@ public class SequenceIdMatcher
    */\r
   SequenceI[] findIdMatch(SequenceI[] seqs)\r
   {\r
-    SequenceI[] namedseqs = new SequenceI[seqs.length];\r
-\r
+    SequenceI[] namedseqs = null;\r
     int i = 0;\r
     SeqIdName nam;\r
 \r
     if (seqs.length > 0)\r
     {\r
+      namedseqs = new SequenceI[seqs.length];\r
       do\r
       {\r
         nam = new SeqIdName(seqs[i].getName());\r
@@ -103,7 +103,7 @@ public class SequenceIdMatcher
           namedseqs[i] = null;\r
         }\r
       }\r
-      while (i++ < seqs.length);\r
+      while (++i < seqs.length);\r
     }\r
 \r
     return namedseqs;\r
@@ -115,12 +115,15 @@ public class SequenceIdMatcher
 \r
     SeqIdName(String s)\r
     {\r
-      id = new String(s);\r
+      if (s!=null)\r
+        id = new String(s);\r
+      else\r
+        id = "";\r
     }\r
 \r
     public int hashCode()\r
     {\r
-      return (id.substring(0, 4).hashCode());\r
+      return ((id.length()>=4) ? id.substring(0, 4).hashCode() : id.hashCode());\r
     }\r
 \r
     public boolean equals(Object s)\r
@@ -140,24 +143,45 @@ public class SequenceIdMatcher
       return false;\r
     }\r
 \r
+    /**\r
+     * Characters that define the end of a unique sequence ID at\r
+     * the beginning of an arbitrary ID string\r
+     * JBPNote: This is a heuristic that will fail for arbritrarily extended sequence id's\r
+     * (like portions of an aligned set of repeats from one sequence)\r
+     */\r
+    private String WORD_SEP="~. |#\\/<>!\"£$%^*)}[@',?";\r
+\r
+   /**\r
+    * matches if one ID properly contains another at a whitespace boundary.\r
+    * TODO: (JBPNote) These are not efficient. should use char[] for speed\r
+    * todo: (JBPNote) Set separator characters appropriately\r
+    * @param s SeqIdName\r
+    * @return boolean\r
+    */\r
     public boolean equals(SeqIdName s)\r
     {\r
-      if (id.startsWith(s.id) || s.id.startsWith(id))\r
-      {\r
-        return true;\r
-      }\r
-\r
-      return false;\r
+      if (id.length()>s.id.length()) {\r
+        return id.startsWith(s.id) ?\r
+            (WORD_SEP.indexOf(id.charAt(s.id.length()))>-1)\r
+            : false;\r
+      } else\r
+        return s.id.startsWith(id) ?\r
+            (s.id.equals(id) ? true :\r
+             (WORD_SEP.indexOf(s.id.charAt(id.length()))>-1))\r
+            : false;\r
     }\r
 \r
     public boolean equals(String s)\r
     {\r
-      if (id.startsWith(s) || s.startsWith(id))\r
-      {\r
-        return true;\r
-      }\r
-\r
-      return false;\r
+      if (id.length()>s.length()) {\r
+        return id.startsWith(s) ?\r
+            (WORD_SEP.indexOf(id.charAt(s.length()))>-1)\r
+            : false;\r
+      } else\r
+        return s.startsWith(id) ?\r
+            (s.equals(id) ? true :\r
+             (WORD_SEP.indexOf(s.charAt(id.length()))>-1))\r
+            : false;\r
     }\r
   }\r
 }\r