in progress

[jalview.git] / forester / java / src / org / forester / application / gene_tree_preprocess.java

diff --git a/forester/java/src/org/forester/application/gene_tree_preprocess.java b/forester/java/src/org/forester/application/gene_tree_preprocess.java
index 525a3fc..581381c 100644 (file)
--- a/forester/java/src/org/forester/application/gene_tree_preprocess.java
+++ b/forester/java/src/org/forester/application/gene_tree_preprocess.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.application;
 
@@ -29,32 +29,31 @@ import java.io.BufferedWriter;
 import java.io.File;
 import java.io.FileWriter;
 import java.io.IOException;
-import java.util.HashSet;
-import java.util.Set;
 import java.util.SortedSet;
+import java.util.TreeSet;
 
-import org.forester.archaeopteryx.tools.SequenceDataRetriver;
 import org.forester.io.parsers.util.ParserUtils;
 import org.forester.io.writers.PhylogenyWriter;
 import org.forester.phylogeny.Phylogeny;
-import org.forester.phylogeny.PhylogenyMethods;
 import org.forester.phylogeny.PhylogenyNode;
 import org.forester.phylogeny.factories.ParserBasedPhylogenyFactory;
 import org.forester.phylogeny.factories.PhylogenyFactory;
 import org.forester.phylogeny.iterators.PhylogenyNodeIterator;
 import org.forester.util.CommandLineArguments;
 import org.forester.util.ForesterUtil;
+import org.forester.ws.seqdb.SequenceDbWsTools;
 
 public class gene_tree_preprocess {
 
-    final static private String HELP_OPTION_1 = "help";
-    final static private String HELP_OPTION_2 = "h";
-    final static private String PRG_NAME      = "gene_tree_preprocess";
-    final static private String PRG_DESC      = "gene tree preprocessing for SDI analysis";
-    final static private String PRG_VERSION   = "1.00";
-    final static private String PRG_DATE      = "2012.06.07";
-    final static private String E_MAIL        = "phylosoft@gmail.com";
-    final static private String WWW           = "www.phylosoft.org/forester/";
+    final static private String HELP_OPTION_1           = "help";
+    final static private String HELP_OPTION_2           = "h";
+    final static private String PRG_NAME                = "gene_tree_preprocess";
+    final static private String PRG_DESC                = "gene tree preprocessing for SDI analysis";
+    final static private String PRG_VERSION             = "1.01";
+    final static private String PRG_DATE                = "2012.06.07";
+    final static private String E_MAIL                  = "phylosoft@gmail.com";
+    final static private String WWW                     = "www.phylosoft.org/forester";
+    private final static int    DEFAULT_LINES_TO_RETURN = 50;
 
     public static void main( final String[] args ) {
         try {
@@ -74,7 +73,7 @@ public class gene_tree_preprocess {
                                          "failed to read phylogeny from [" + in + "]: " + e.getLocalizedMessage() );
             }
             final File outtree = new File( ForesterUtil.removeSuffix( in.toString() )
-                    + "_preprocessed_gene_tree.phylo.xml" );
+                                           + "_preprocessed_gene_tree.phylo.xml" );
             final File removed_nodes = new File( ForesterUtil.removeSuffix( in.toString() ) + "_removed_nodes.txt" );
             final File present_species = new File( ForesterUtil.removeSuffix( in.toString() ) + "_species_present.txt" );
             checkForOutputFileWriteability( outtree );
@@ -82,13 +81,17 @@ public class gene_tree_preprocess {
             checkForOutputFileWriteability( present_species );
             if ( phy.getNumberOfExternalNodes() < 2 ) {
                 ForesterUtil.fatalError( PRG_NAME, "phylogeny has " + phy.getNumberOfExternalNodes()
-                        + " external node(s), aborting" );
+                                         + " external node(s), aborting" );
             }
-            final SortedSet<String> not_found = SequenceDataRetriver.obtainSeqInformation( phy, true );
+            final SortedSet<String> not_found = SequenceDbWsTools.obtainSeqInformation( phy,
+                                                                                        true,
+                                                                                        false,
+                                                                                        DEFAULT_LINES_TO_RETURN );
             for( final String remove_me : not_found ) {
-               // System.out.println( " not found: " + remove_me );
-                PhylogenyMethods.removeNode( phy.getNode( remove_me ), phy );
+                phy.deleteSubtree( phy.getNode( remove_me ), true );
             }
+            phy.clearHashIdToNodeMap();
+            phy.externalNodesHaveChanged();
             if ( phy.getNumberOfExternalNodes() < 2 ) {
                 ForesterUtil.fatalError( PRG_NAME,
                                          "after removal of unresolvable external nodes, phylogeny has "
@@ -102,21 +105,21 @@ public class gene_tree_preprocess {
                 ForesterUtil.fatalError( PRG_NAME, "failed to write to [" + outtree + "]: " + e.getLocalizedMessage() );
             }
             ForesterUtil.programMessage( PRG_NAME, "wrote output phylogeny to: " + outtree );
-            final Set<String> species_found = new HashSet<String>();
+            final SortedSet<String> species_set = new TreeSet<String>();
+            for( final PhylogenyNodeIterator iter = phy.iteratorExternalForward(); iter.hasNext(); ) {
+                final PhylogenyNode node = iter.next();
+                if ( node.getNodeData().isHasTaxonomy() ) {
+                    final String sn = node.getNodeData().getTaxonomy().getScientificName();
+                    if ( !ForesterUtil.isEmpty( sn ) ) {
+                        species_set.add( sn );
+                    }
+                }
+            }
             try {
                 final BufferedWriter out = new BufferedWriter( new FileWriter( present_species ) );
-                for( final PhylogenyNodeIterator iter = phy.iteratorExternalForward(); iter.hasNext(); ) {
-                    final PhylogenyNode node = iter.next();
-                    if ( node.getNodeData().isHasTaxonomy() ) {
-                        final String sn = node.getNodeData().getTaxonomy().getScientificName();
-                        if ( !ForesterUtil.isEmpty( sn ) ) {
-                            if ( !species_found.contains( sn ) ) {
-                                species_found.add( sn );
-                                out.write( node.getNodeData().getTaxonomy().getScientificName() );
-                                out.newLine();
-                            }
-                        }
-                    }
+                for( final String species : species_set ) {
+                    out.write( species );
+                    out.newLine();
                 }
                 out.close();
             }
@@ -145,7 +148,7 @@ public class gene_tree_preprocess {
         }
     }
 
-    public static void checkForOutputFileWriteability( final File outfile ) {
+    private static void checkForOutputFileWriteability( final File outfile ) {
         final String error = ForesterUtil.isWritableFile( outfile );
         if ( !ForesterUtil.isEmpty( error ) ) {
             ForesterUtil.fatalError( PRG_NAME, error );