X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=forester%2Fjava%2Fsrc%2Forg%2Fforester%2Futil%2FForesterUtil.java;h=a8c81b0dd23ae214f2280b324a659ae3f3f544f2;hb=aec065f948a075773794133f102ea19eb1d59f64;hp=cb04e6521e93833a056a9756c4cad769ed6647ac;hpb=cc75486aa58b98ab6fa53d8de4cb9b984a86bf83;p=jalview.git

diff --git a/forester/java/src/org/forester/util/ForesterUtil.java b/forester/java/src/org/forester/util/ForesterUtil.java
index cb04e65..a8c81b0 100644
--- a/forester/java/src/org/forester/util/ForesterUtil.java
+++ b/forester/java/src/org/forester/util/ForesterUtil.java
@@ -21,7 +21,7 @@
 // Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
 //
 // Contact: phylosoft @ gmail . com
-// WWW: www.phylosoft.org/forester
+// WWW: https://sites.google.com/site/cmzmasek/home/software/forester
 
 package org.forester.util;
 
@@ -59,6 +59,7 @@ import java.util.SortedMap;
 import java.util.SortedSet;
 import java.util.TreeMap;
 import java.util.TreeSet;
+import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 
 import org.forester.phylogeny.PhylogenyNode;
@@ -82,6 +83,14 @@ public final class ForesterUtil {
     public static final NumberFormat FORMATTER_6;
     public static final NumberFormat FORMATTER_06;
     public static final NumberFormat FORMATTER_3;
+    public static final String       NCBI_PROTEIN                     = "http://www.ncbi.nlm.nih.gov/protein/";
+    public static final String       NCBI_NUCCORE                     = "http://www.ncbi.nlm.nih.gov/nuccore/";
+    public final static String       UNIPROT_KB                       = "http://www.uniprot.org/uniprot/";
+    public final static Pattern      UNIPROT_KB_PATTERN_1             = Pattern
+                                                                              .compile( "(?:\\b|_)(?:sp|tr)[\\.|\\-_=/\\\\]([A-Z][0-9][A-Z0-9]{3}[0-9])(?:\\b|_)" );
+    public final static Pattern      UNIPROT_KB_PATTERN_2             = Pattern
+                                                                              .compile( "\\b(?:[A-Z0-9]{2,5}|(?:[A-Z][0-9][A-Z0-9]{3}[0-9]))_(([A-Z9][A-Z]{2}[A-Z0-9]{2})|RAT|PIG|PEA)\\b" );
+    public static final String       NCBI_GI                          = "http://www.ncbi.nlm.nih.gov/protein/gi:";
     static {
         final DecimalFormatSymbols dfs = new DecimalFormatSymbols();
         dfs.setDecimalSeparator( '.' );
@@ -95,6 +104,124 @@ public final class ForesterUtil {
     private ForesterUtil() {
     }
 
+    public static String extractRefSeqAccessorAccessor( final PhylogenyNode node ) {
+        String v = null;
+        if ( node.getNodeData().isHasSequence() ) {
+            final Sequence seq = node.getNodeData().getSequence();
+            if ( !isEmpty( seq.getSymbol() ) ) {
+                v = SequenceIdParser.parseRefSeqAccessor( seq.getSymbol() );
+            }
+            if ( isEmpty( v ) && !isEmpty( seq.getName() ) ) {
+                v = SequenceIdParser.parseRefSeqAccessor( seq.getName() );
+            }
+            if ( isEmpty( v ) && ( node.getNodeData().getSequence().getAccession() != null )
+                    && !isEmpty( seq.getAccession().getValue() ) ) {
+                v = SequenceIdParser.parseRefSeqAccessor( seq.getAccession().getValue() );
+            }
+        }
+        if ( isEmpty( v ) && !isEmpty( node.getName() ) ) {
+            v = SequenceIdParser.parseRefSeqAccessor( node.getName() );
+        }
+        return v;
+    }
+
+    public static String extractGenbankAccessor( final PhylogenyNode node ) {
+        String v = null;
+        if ( node.getNodeData().isHasSequence() ) {
+            final Sequence seq = node.getNodeData().getSequence();
+            if ( !isEmpty( seq.getSymbol() ) ) {
+                v = SequenceIdParser.parseGenbankAccessor( seq.getSymbol() );
+            }
+            if ( isEmpty( v ) && !isEmpty( seq.getName() ) ) {
+                v = SequenceIdParser.parseGenbankAccessor( seq.getName() );
+            }
+            if ( isEmpty( v ) && ( node.getNodeData().getSequence().getAccession() != null )
+                    && !isEmpty( seq.getAccession().getValue() ) ) {
+                v = SequenceIdParser.parseGenbankAccessor( seq.getAccession().getValue() );
+            }
+        }
+        if ( isEmpty( v ) && !isEmpty( node.getName() ) ) {
+            v = SequenceIdParser.parseGenbankAccessor( node.getName() );
+        }
+        return v;
+    }
+
+    public static String extractGInumber( final PhylogenyNode node ) {
+        String v = null;
+        if ( node.getNodeData().isHasSequence() ) {
+            final Sequence seq = node.getNodeData().getSequence();
+            if ( isEmpty( v ) && !isEmpty( seq.getName() ) ) {
+                v = SequenceIdParser.parseGInumber( seq.getName() );
+            }
+            if ( isEmpty( v ) && ( node.getNodeData().getSequence().getAccession() != null )
+                    && !isEmpty( seq.getAccession().getValue() ) ) {
+                v = SequenceIdParser.parseGInumber( seq.getAccession().getValue() );
+            }
+        }
+        if ( isEmpty( v ) && !isEmpty( node.getName() ) ) {
+            v = SequenceIdParser.parseGInumber( node.getName() );
+        }
+        return v;
+    }
+
+    public static String extractUniProtKbProteinSeqIdentifier( final PhylogenyNode node ) {
+        String upkb = null;
+        if ( node.getNodeData().isHasSequence() ) {
+            final Sequence seq = node.getNodeData().getSequence();
+            Matcher m;
+            if ( !isEmpty( seq.getSymbol() ) ) {
+                m = UNIPROT_KB_PATTERN_1.matcher( seq.getSymbol() );
+                if ( m.find() ) {
+                    upkb = m.group( 1 );
+                }
+                else {
+                    m = UNIPROT_KB_PATTERN_2.matcher( seq.getSymbol() );
+                    if ( m.find() ) {
+                        upkb = m.group();
+                    }
+                }
+            }
+            if ( isEmpty( upkb ) && !isEmpty( seq.getName() ) ) {
+                m = UNIPROT_KB_PATTERN_1.matcher( seq.getName() );
+                if ( m.find() ) {
+                    upkb = m.group( 1 );
+                }
+                else {
+                    m = UNIPROT_KB_PATTERN_2.matcher( seq.getName() );
+                    if ( m.find() ) {
+                        upkb = m.group();
+                    }
+                }
+            }
+            if ( isEmpty( upkb ) && ( node.getNodeData().getSequence().getAccession() != null )
+                    && !isEmpty( seq.getAccession().getValue() ) ) {
+                m = UNIPROT_KB_PATTERN_1.matcher( seq.getAccession().getValue() );
+                if ( m.find() ) {
+                    upkb = m.group( 1 );
+                }
+                else {
+                    m = UNIPROT_KB_PATTERN_2.matcher( seq.getAccession().getValue() );
+                    if ( m.find() ) {
+                        upkb = m.group();
+                    }
+                }
+            }
+        }
+        if ( isEmpty( upkb ) && !isEmpty( node.getName() ) ) {
+            final Matcher m1 = UNIPROT_KB_PATTERN_1.matcher( node.getName() );
+            if ( m1.find() ) {
+                upkb = m1.group( 1 );
+            }
+            else {
+                final Matcher m2 = UNIPROT_KB_PATTERN_2.matcher( node.getName() );
+                if ( m2.find() ) {
+                    upkb = m2.group();
+                }
+            }
+        }
+        return upkb;
+    }
+
     final public static void appendSeparatorIfNotEmpty( final StringBuffer sb, final char separator ) {
         if ( sb.length() > 0 ) {
             sb.append( separator );
@@ -996,6 +1123,10 @@ public final class ForesterUtil {
         System.out.print( "]" );
     }
 
+    public final static void updateProgress( final int i, final DecimalFormat f ) {
+        System.out.print( "\r[" + f.format( i ) + "]" );
+    }
+
     public final static String wordWrap( final String str, final int width ) {
         final StringBuilder sb = new StringBuilder( str );
         int start = 0;