inprogress

[jalview.git] / forester / java / src / org / forester / io / parsers / FastaParser.java

diff --git a/forester/java/src/org/forester/io/parsers/FastaParser.java b/forester/java/src/org/forester/io/parsers/FastaParser.java
index 7b7bfac..348bec4 100644 (file)
--- a/forester/java/src/org/forester/io/parsers/FastaParser.java
+++ b/forester/java/src/org/forester/io/parsers/FastaParser.java
@@ -22,7 +22,7 @@
 // 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.io.parsers;
 
@@ -41,13 +41,8 @@ import java.util.regex.Pattern;
 import org.forester.msa.BasicMsa;
 import org.forester.msa.Msa;
 import org.forester.msa.MsaFormatException;
-import org.forester.phylogeny.Phylogeny;
-import org.forester.phylogeny.PhylogenyNode;
-import org.forester.phylogeny.data.Accession;
-import org.forester.phylogeny.iterators.PhylogenyNodeIterator;
 import org.forester.sequence.BasicSequence;
 import org.forester.sequence.Sequence;
-import org.forester.util.ForesterUtil;
 
 public class FastaParser {
 
@@ -55,7 +50,7 @@ public class FastaParser {
     private static final Pattern SEQ_REGEX       = Pattern.compile( "^\\s*(.+)" );
     private static final Pattern ANYTHING_REGEX  = Pattern.compile( "[\\d\\s]+" );
     //>gi|71834668|ref|NP_001025424.1| Bcl2 [Danio rerio]
-    private static final Pattern FASTA_DESC_LINE = Pattern
+    public static final Pattern  FASTA_DESC_LINE = Pattern
                                                          .compile( ">?\\s*([^|]+)\\|([^|]+)\\S*\\s+(.+)\\s+\\[(.+)\\]" );
 
     public static void main( final String[] args ) {
@@ -73,6 +68,10 @@ public class FastaParser {
         }
     }
 
+    static public boolean isLikelyFasta( final File f ) throws IOException {
+        return isLikelyFasta( new FileInputStream( f ) );
+    }
+
     static public boolean isLikelyFasta( final InputStream is ) throws IOException {
         final BufferedReader reader = new BufferedReader( new InputStreamReader( is, "UTF-8" ) );
         String line = null;
@@ -185,36 +184,4 @@ public class FastaParser {
         }
         return line;
     }
-
-    public static void extractFastaInformation( final Phylogeny phy ) {
-        for( final PhylogenyNodeIterator iter = phy.iteratorExternalForward(); iter.hasNext(); ) {
-            final PhylogenyNode node = iter.next();
-            if ( !ForesterUtil.isEmpty( node.getName() ) ) {
-                final Matcher name_m = FASTA_DESC_LINE.matcher( node.getName() );
-                if ( name_m.lookingAt() ) {
-                    System.out.println();
-                    // System.out.println( name_m.group( 1 ) );
-                    // System.out.println( name_m.group( 2 ) );
-                    // System.out.println( name_m.group( 3 ) );
-                    // System.out.println( name_m.group( 4 ) );
-                    final String acc_source = name_m.group( 1 );
-                    final String acc = name_m.group( 2 );
-                    final String seq_name = name_m.group( 3 );
-                    final String tax_sn = name_m.group( 4 );
-                    if ( !ForesterUtil.isEmpty( acc_source ) && !ForesterUtil.isEmpty( acc ) ) {
-                        ForesterUtil.ensurePresenceOfSequence( node );
-                        node.getNodeData().getSequence( 0 ).setAccession( new Accession( acc, acc_source ) );
-                    }
-                    if ( !ForesterUtil.isEmpty( seq_name ) ) {
-                        ForesterUtil.ensurePresenceOfSequence( node );
-                        node.getNodeData().getSequence( 0 ).setName( seq_name );
-                    }
-                    if ( !ForesterUtil.isEmpty( tax_sn ) ) {
-                        ForesterUtil.ensurePresenceOfTaxonomy( node );
-                        node.getNodeData().getTaxonomy( 0 ).setScientificName( tax_sn );
-                    }
-                }
-            }
-        }
-    }
 }