X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=forester%2Fjava%2Fsrc%2Forg%2Fforester%2Futil%2FSequenceIdParser.java;h=d75a6dd8e6e2237a02a551cd56698f5d8b24f9c4;hb=32b66090883b7ca47347fef19152d84bd02c0b9d;hp=7486d8a43359898f457f3f2c7aaadceccedf0325;hpb=813491b06a1cf8364202bfc803a4dcff8b89977c;p=jalview.git

diff --git a/forester/java/src/org/forester/util/SequenceIdParser.java b/forester/java/src/org/forester/util/SequenceIdParser.java
index 7486d8a..d75a6dd 100644
--- a/forester/java/src/org/forester/util/SequenceIdParser.java
+++ b/forester/java/src/org/forester/util/SequenceIdParser.java
@@ -32,7 +32,6 @@ import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 
 import org.forester.phylogeny.data.Identifier;
-import org.forester.ws.uniprot.DatabaseTools;
 
 public final class SequenceIdParser {
 
@@ -58,20 +57,69 @@ public final class SequenceIdParser {
                                                                          .compile( "(?:\\A|.*[^a-zA-Z0-9])([A-Z]{2}\\d{6})(?:[^a-zA-Z0-9]|\\Z)" );
     private final static Pattern GENBANK_PROTEIN_AC_PATTERN      = Pattern
                                                                          .compile( "(?:\\A|.*[^a-zA-Z0-9])([A-Z]{3}\\d{5})(?:[^a-zA-Z0-9]|\\Z)" );
-    private final static boolean DEBUG                           = false;
+   
+    // RefSeq accession numbers can be distinguished from GenBank accessions 
+    // by their distinct prefix format of 2 characters followed by an
+    // underscore character ('_'). For example, a RefSeq protein accession is NP_015325. 
+    private final static Pattern REFSEQ_PATTERN      = Pattern
+    .compile( "(?:\\A|.*[^a-zA-Z0-9])([A-Z]{2}_\\d{6,})(?:[^a-zA-Z0-9]|\\Z)" );
 
+   
     
-    
+    /**
+     * Returns null if no match.
+     * 
+     */
     public final static Identifier parse( final String s ) {
-        String v = DatabaseTools.parseGenbankAccessor( s );
+        String v = parseGenbankAccessor( s );
+        if ( !ForesterUtil.isEmpty( v ) ) {
+
+            return new Identifier( v, Identifier.NCBI );
+        }
+        v = parseRefSeqAccessor( s );
         if ( !ForesterUtil.isEmpty( v ) ) {
-            return new Identifier( v, "genbank" );
+            return new Identifier( v, Identifier.REFSEQ );
         }
-        
         return null;
     }
     
+    /**
+     * Returns null if no match.
+     * 
+     */
+    public static String parseGenbankAccessor( final String query ) {
+        Matcher m = GENBANK_NUCLEOTIDE_AC_PATTERN_1.matcher( query );
+        if ( m.lookingAt() ) {
+            return m.group( 1 );
+        }
+        else {
+            m = GENBANK_NUCLEOTIDE_AC_PATTERN_2.matcher( query );
+            if ( m.lookingAt() ) {
+                return m.group( 1 );
+            }
+            else {
+                m = GENBANK_PROTEIN_AC_PATTERN.matcher( query );
+                if ( m.lookingAt() ) {
+                    return m.group( 1 );
+                }
+                else {
+                    return null;
+                }
+            }
+        }
+    }
     
+    /**
+     * Returns null if no match.
+     * 
+     */
+    private final static String parseRefSeqAccessor( final String query ) {
+        Matcher m = REFSEQ_PATTERN.matcher( query );
+        if ( m.lookingAt() ) {
+            return m.group( 1 );
+        }
+        return null;
+    }