From 4dfd20147b1753e436e71ea673e2ac9a7fae420d Mon Sep 17 00:00:00 2001
From: jprocter <jprocter@compbio.dundee.ac.uk>
Date: Fri, 7 Dec 2012 15:50:23 +0000
Subject: [PATCH] Anne's refined RNAML parser. Seems to have some bugs still
 (merge with jprocter's bugfixed parser)

---
 src/jalview/io/RnamlFile.java |  501 +++++++++++++++--------------------------
 1 file changed, 180 insertions(+), 321 deletions(-)

diff --git a/src/jalview/io/RnamlFile.java b/src/jalview/io/RnamlFile.java
index bc6b385..0029e60 100644
--- a/src/jalview/io/RnamlFile.java
+++ b/src/jalview/io/RnamlFile.java
@@ -17,34 +17,22 @@
  */
 package jalview.io;
 
-import jalview.datamodel.AlignmentAnnotation;
-import jalview.datamodel.Annotation;
-import jalview.datamodel.SecondaryStructureAnnotation;
-import jalview.datamodel.Sequence;
-import jalview.datamodel.SequenceFeature;
-import jalview.datamodel.SequenceI;
-import jalview.ws.jws1.Annotate3D;
-
 import java.io.BufferedReader;
 import java.io.FileNotFoundException;
-import java.io.IOException;
-import java.io.Reader;
 import java.io.FileReader;
-import java.lang.reflect.Array;
 import java.util.ArrayList;
-import java.util.Collections;
-import java.util.Hashtable;
-import java.util.List;
 import java.util.Vector;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
 
-import javax.xml.parsers.ParserConfigurationException;
-import javax.xml.parsers.SAXParser;
-import javax.xml.parsers.SAXParserFactory;
-
-import org.xml.sax.InputSource;
-import org.xml.sax.SAXException;
-
-import com.stevesoft.pat.Regex;
+import jalview.analysis.SecStrConsensus;
+import jalview.analysis.SecStrConsensus.SimpleBP;
+import jalview.datamodel.AlignmentAnnotation;
+import jalview.datamodel.Annotation;
+import jalview.datamodel.SecondaryStructureAnnotation;
+import jalview.datamodel.Sequence;
+import jalview.datamodel.SequenceFeature;
+import jalview.datamodel.SequenceI;
 
 import fr.orsay.lri.varna.exceptions.ExceptionFileFormatOrSyntax;
 import fr.orsay.lri.varna.exceptions.ExceptionLoadingFailed;
@@ -55,318 +43,189 @@ import fr.orsay.lri.varna.factories.RNAFactory.RNAFileType;
 import fr.orsay.lri.varna.models.rna.RNA;
 import fr.orsay.lri.varna.utils.RNAMLParser;
 
+public class RnamlFile extends AlignFile {
+	public int id;
+	public String namefile;
+	public String name;
+	protected ArrayList<RNA> result;
 
+	public RnamlFile() {
+		super();
 
+	}
+
+	public RnamlFile(String inFile, String type) throws Exception {
+		super(inFile, type);
 
-public class RnamlFile extends AlignFile
-{
-	 public String id;
-	 protected ArrayList<RNA> result;
-	 
-	 public RnamlFile()
-	  {
-		 super();
-		
-	  }
-	 
-	 public RnamlFile(String inFile, String type) throws IOException, ExceptionFileFormatOrSyntax, ParserConfigurationException, SAXException, ExceptionPermissionDenied, ExceptionLoadingFailed, InterruptedException, ExceptionUnmatchedClosingParentheses
-	  {
-	    super(inFile, type);
-	    
-	  }
-
-	  public RnamlFile(FileParse source) throws IOException, ExceptionFileFormatOrSyntax, ParserConfigurationException, SAXException, ExceptionPermissionDenied, ExceptionLoadingFailed, InterruptedException, ExceptionUnmatchedClosingParentheses
-	  {
-	    super(source);
-	    
-	  }
-	  
-//	  public RnamlFile(BufferedReader r) throws IOException, ExceptionFileFormatOrSyntax, ParserConfigurationException, SAXException, ExceptionPermissionDenied, ExceptionLoadingFailed
-//	  {
-//		  	super();	    
-//		    parse(r);
-//		    // sets the index of each sequence in the alignment
-//		    for( int i=0,c=seqs.size(); i<c; i++ ) {  
-//		    	seqs.get(i).setIndex(i);  
-//		    }
-//		  
-//	    
-//	  }
-	  
-	public BufferedReader CreateReader() throws FileNotFoundException
-	{
-		 FileReader fr = null;
-		  fr = new FileReader(inFile); 
-		  
-		  BufferedReader r = new BufferedReader (fr);
-		  return r;
 	}
-	
-	
+
+	public RnamlFile(FileParse source) throws Exception {
+		super(source);
+
+	}
+
+	public BufferedReader CreateReader() throws FileNotFoundException {
+		FileReader fr = null;
+		fr = new FileReader(inFile);
+
+		BufferedReader r = new BufferedReader(fr);
+		return r;
+	}
+
 	@SuppressWarnings("unchecked")
-	public void parse() throws FileNotFoundException, ExceptionPermissionDenied, ExceptionLoadingFailed, ExceptionFileFormatOrSyntax
-	{
-		 result = RNAFactory.loadSecStrRNAML(getReader());
-		  
-		 System.out.println("this is the secondary scructure:" +result.size());
-		  //System.out.println("this is the secondary scructure:" +result.toString());
-		  //System.out.println("this is the secondary scructure:" +result.size());
-		 SequenceI[] seqs = new SequenceI[result.size()];
-		  System.out.println(type); //the type is "File"
-		  System.out.println(inFile );//inFile is the path
-		  
-		 for(int i=0;i<result.size();i++)
-			{
-			 RNA current = result.get(i);
-		  System.out.println(current.getSeq());
-		  //System.out.println(result.get(i).getStructBPSEQ());
-		  System.out.println(result.get(i).getStructDBN(true));
-		  System.out.println(i);
-		  String rna =current.getStructDBN(true);
-		  //System.out.println(result.get(0).getAnnotations());
-		  String seq = current.getSeq();
-	
-		  int begin=1;
-		  int end = seq.length();
-		  
-		  id = current.getName();
-		  seqs[i] = new Sequence(id, seq, begin, end);
-		  //System.out.println(seq);
-		  System.out.println("Rna is still "+rna);
-		  //SequenceFeature feat = new SequenceFeature(type, "",5, 13, 0f, null);
-		  //seqs[i].addSequenceFeature(feat);
-		  
-		  String[] annot=new String[rna.length()] ;
-		  Annotation[] ann = new Annotation[rna.length()];
-		
-		  for(int j=0;j<rna.length();j++)
-			{
-			  annot[j] =rna.substring(j,j+1);
-		  
+	public void parse() throws FileNotFoundException,
+			ExceptionPermissionDenied, ExceptionLoadingFailed,
+			ExceptionFileFormatOrSyntax {
+
+		result = RNAFactory.loadSecStrRNAML(getReader());
+
+		ArrayList<ArrayList> allarray = new ArrayList();
+		ArrayList<ArrayList<SimpleBP>> BP = new ArrayList();
+		ArrayList strucinarray = new ArrayList();
+		SequenceI[] seqs = new SequenceI[result.size()];
+		namefile = inFile.getName();
+
+		for (int i = 0; i < result.size(); i++) {
+
+			RNA current = result.get(i);
+			String rna = current.getStructDBN(true);
+			String seq = current.getSeq();
+			int begin = 1;
+			int end = 1000;
+			id = i;
+			System.out.println("id=" + i);
+			name = this.safeName(namefile, i);
+			seqs[i] = new Sequence(name, seq, begin, end);
+
+			seqs[i].setEnd(seqs[i].findPosition(seqs[i].getLength()));
+			String[] annot = new String[rna.length()];
+			Annotation[] ann = new Annotation[rna.length()];
+
+			for (int j = 0; j < rna.length(); j++) {
+				annot[j] = "" + rna.charAt(j);
+
 			}
-		  
-		  for(int k=0;k<rna.length();k++)
-		{
-			  ann[k] = new Annotation(annot[k], "", jalview.schemes.ResidueProperties.getRNASecStrucState(annot[k]).charAt(0), 0f);
-			  			  
-		}
-		  AlignmentAnnotation align = new AlignmentAnnotation("Sec. str.",current.getID(),ann);
-		  
-		  seqs[i].addAlignmentAnnotation(align);
-		  seqs[i].setRNA(result.get(i));
-		  this.annotations.addElement(align);
+			for (int k = 0; k < rna.length(); k++) {
+				ann[k] = new Annotation(annot[k], "",
+						jalview.schemes.ResidueProperties.getRNASecStrucState(
+								annot[k]).charAt(0), 0f);
+			}
+
+			AlignmentAnnotation align = new AlignmentAnnotation("Sec. str.",
+					current.getID(), ann);
+
+			seqs[i].addAlignmentAnnotation(align);
+			seqs[i].setRNA(result.get(i));
+
+			char[] struc = align.getRNAStruc().toCharArray();
+
+			allarray.add(strucinarray);
+
+			this.annotations.addElement(align);
+			BP.add(align.bps);
+			
 		}
-   	  this.setSeqs(seqs);
-		 
-		 
-		
+		this.setSeqs(seqs);
+		int[] tab = SecStrConsensus.extractConsensus(BP);
+
 
 	}
-	  
-	
-	  public void parse(BufferedReader r) throws ExceptionPermissionDenied, ExceptionLoadingFailed, ExceptionFileFormatOrSyntax 
-	  {
-		   System.out.println("parse with Buffer (RnamlFile");
-		  Hashtable seqhash = new Hashtable();
-		  Vector headers = new Vector();
-		  //ArrayList<String> seq =new ArrayList();
-		  //System.out.println(r);
-		  
-		  result = RNAFactory.loadSecStr(r,RNAFileType.FILE_TYPE_RNAML);
-		  
-		  System.out.println("this is the secondary scructure:" +result.toArray());
-		  System.out.println("this is the secondary scructure:" +result.toString());
-		  System.out.println("this is the secondary scructure:" +result.size());
-			for(int i=0;i<result.size();i++)
-			{
-		  System.out.println(result.get(0).getSeq());
-		  System.out.println(result.get(0).getStructBPSEQ());
-		  System.out.println(result.get(0).getStructDBN(true));
-		  String rna1 =result.get(0).getStructDBN(true);
-		  System.out.println(result.get(0).getAnnotations());
-		  String seq = result.get(0).getSeq();
-		  System.out.println(type); //the type is "File"
-		  System.out.println(inFile );//inFile is the path
-		  int begin=0;
-		  int end = seq.length()-1;
-		  SequenceI[] seqs = new SequenceI[1];
-		  id = safeName(getDataName());
-		  seqs[0] = new Sequence(id, seq, begin, end);
-		  SequenceFeature feat = new SequenceFeature(type, "",5, 13, 0f, null);
-		  seqs[0].addSequenceFeature(feat);
-		  this.setSeqs(seqs);
-		  
-		  String[] annot=new String[10000] ;
-		  Annotation[] ann = new Annotation[1000];
-		  for(int j=0;j<rna1.length();j++)
-			{
-		  annot[j] =rna1.substring(j,j+1);
-		  //System.out.println(annot[j]);
+
+	public void parse(BufferedReader r) throws ExceptionPermissionDenied,
+			ExceptionLoadingFailed, ExceptionFileFormatOrSyntax {
+
+		result = RNAFactory.loadSecStrRNAML(r);
+
+		ArrayList<ArrayList> allarray = new ArrayList();
+		ArrayList<ArrayList<SimpleBP>> BP = new ArrayList();
+		ArrayList strucinarray = new ArrayList();
+		SequenceI[] seqs = new SequenceI[result.size()];
+		namefile = inFile.getName();
+
+		for (int i = 0; i < result.size(); i++) {
+
+			RNA current = result.get(i);
+			String rna = current.getStructDBN(true);
+
+			String seq = current.getSeq();
+			int begin = 1;
+			int end = 1000;
+			id = i;
+			name = this.safeName(namefile, i);
+			seqs[i] = new Sequence(name, seq, begin, end);
+
+			seqs[i].setEnd(seqs[i].findPosition(seqs[i].getLength()));
+			String[] annot = new String[rna.length()];
+			Annotation[] ann = new Annotation[rna.length()];
+
+			for (int j = 0; j < rna.length(); j++) {
+				annot[j] = "" + rna.charAt(j);
+
 			}
-		  
-		  for(int k=0;k<rna1.length();k++)
-		{
-			  ann[k] = new Annotation(annot[k], "", ' ', 0f);
-			  
-			  ann[k].secondaryStructure = jalview.schemes.ResidueProperties.getRNASecStrucState(annot[k]).charAt(0);
-			  
-		}
-		  AlignmentAnnotation align = new AlignmentAnnotation("Secondary structure","",ann);
-		  seqs[0].addAlignmentAnnotation(align);
-		  this.annotations.addElement(align);
+			for (int k = 0; k < rna.length(); k++) {
+				ann[k] = new Annotation(annot[k], "",
+						jalview.schemes.ResidueProperties.getRNASecStrucState(
+								annot[k]).charAt(0), 0f);
+			}
+
+			AlignmentAnnotation align = new AlignmentAnnotation("Sec. str.",
+					current.getID(), ann);
+
+			seqs[i].addAlignmentAnnotation(align);
+			seqs[i].setRNA(result.get(i));
+
+			char[] struc = align.getRNAStruc().toCharArray();
+			for (int y = 0; y < struc.length; y++) {
+				strucinarray.add(struc[y]);
+				// System.out.println("structy"+struc[y]);
+			}
+			// System.out.println("Sequence"+i+"\t : "+strucinarray.toString());
+			allarray.add(strucinarray);
+
+			this.annotations.addElement(align);
+			System.out.println(rna);
+			BP.add(align.bps);
+			System.out.println(align.bps.size());
 		}
-		//Annotate3D tintin = new Annotate3D();
-		  
-		//System.out.println(tintin);  
-		  
-		  
-		  
-		  
-		  
-		  
-		 
+		this.setSeqs(seqs);
+		int[] tab = SecStrConsensus.extractConsensus(BP);
+
 	
-//					for(RNAMLParser.RNATmp rnaTmp : RNAMLData.getMolecules())
-//					{
-//						// Retrieving parsed data
-//						seq = rnaTmp.getSequence(); //contient la séquence sous forme de liste
-//						System.out.println(seq);
-//						String seqAsStr = "";
-//						for(String s:seq)
-//						{
-//							// Only consider last character, e.g. in case of modified nucleotide
-//							seqAsStr += s.charAt(s.length()-1);
-//						}
-//						int begin=0;
-//						int end = seqAsStr.length()-1;
-//						SequenceI[] seqs = new SequenceI[1];
-//						id = safeName(getDataName());
-//						seqs[0] = new Sequence(id, seqAsStr, begin, end);
-//						SequenceFeature feat = new SequenceFeature(type, "",5, 13, 0f, null);
-//
-//	                    seqs[0].addSequenceFeature(feat);
-//						this.setSeqs(seqs);
-						
-						//String pos = annot.substring(1, 2);
-						//System.out.println(annot);
-						//System.out.println(pos);
-						
-					   // ann[1] = new Annotation(pos, "", ' ', 0f);
-					   // ann[1].secondaryStructure = 'A';
-					    //System.out.println(ann[1]);
-					    
-					   // toto.annotations=ann;
-					    
-
-					    
-////						 Annotation[] els = new Annotation[seq.size()];
-//						
-//						//System.err.println(""+seq.size());
-//					    Vector<RNAMLParser.BPTemp> allbpsTmp = rnaTmp.getStructure();
-//					    String[] annot = new String[10000];
-//					  
-//						//Annotation[] ann = new Annotation[allbpsTmp.size()];
-//						for (int i = 0; i < allbpsTmp.size(); i++) {
-//						RNAMLParser.BPTemp bp = allbpsTmp.get(i);
-//						//System.err.println("toto ="+bp);
-//						int bp5 = bp.pos5;
-//						annot[bp5]= "(";
-//						//annot[bp5]= '(';
-//						System.out.println(annot[bp5]);
-//						//System.out.println("bp5 ="+bp5);
-//						int bp3 = bp.pos3;
-//					
-//						annot[bp3]= ")";
-//						System.out.println(annot[bp3]);
-//						//System.out.println(annot.get(bp3));
-//						
-//						//String pos1 = annot.substring(bp5, bp5+1);
-//						//String pos2 = annot.substring(bp3, bp3+1);
-//						//ann[i] = new Annotation(pos1, "", ' ', 0f);
-//						//ann[i+1] = new Annotation(pos2, "", ' ', 0f);
-//						//System.out.println(pos1);
-//						
-//						//System.out.println("bp3 ="+bp3);
-//						//ModeleBP newStyle = bp.createBPStyle(mb, part);
-//						}	
-//						Annotation[] ann = new Annotation[annot.length+1];
-//						for(int i=0;i<annot.length;i++)
-//						{		
-//							if(annot[i]==null)
-//							{
-//								annot[i]=".";
-//							}
-//							
-//							ann[i] = new Annotation(annot[i], "", ' ', 0f);
-//						//System.out.print(annot[i]);
-//						}
-//						
-//						AlignmentAnnotation toto = new AlignmentAnnotation("Secondary structure","",ann);
-//						this.annotations.addElement(toto);
-					
-						// Creating empty structure of suitable size
-//						int[] str = new int[seq.size()];
-//						for (int i=0;i<str.length;i++)
-//						{ str[i] = -1; }
-//						Vector<RNAMLParser.BPTemp> allbpsTmp = rnaTmp.getStructure();
-//						for (int i = 0; i < allbpsTmp.size(); i++) {
-//							RNAMLParser.BPTemp bp = allbpsTmp.get(i);
-//							//System.err.println("toto ="+bp);
-//							int bp5 = bp.pos5;
-//							//System.out.println("bp5 ="+bp5);
-//							int bp3 = bp.pos3;
-							//System.out.println("bp3 ="+bp3);
-							//ModeleBP newStyle = bp.createBPStyle(mb, part);
-//						}
-						//Annotation ann;
-					    //ann = new Annotation(bp5, "", ' ', 0f);
-				
-					//}
+	}
 
-			
+	
 
-//				return result;
-	  
-}
+	public static String print(SequenceI[] s) {
+		return "not yet implemented";
+	}
+
+	public String print() {
+		System.out.print("print :");
+		return print(getSeqsAsArray());
+	}
+
+	public ArrayList getRNA() {
+		return result;
+	}
+
+	public Vector getAnnot() {
+		return annotations;
+	}
+
+	// public static void main(String[] args) {
+	// Pattern p= Pattern.compile("(.+)[.][^.]+");
+	// Matcher m = p.matcher("toto.xml.zip");
+	// System.out.println(m.matches());
+	// System.out.println(m.group(1));
+	// }
+	public String safeName(String namefile, int id) {
+
+		Pattern p = Pattern.compile("(.*)[.][^.]+");
+		Matcher m = p.matcher(namefile);
+		name = m.group(1) + "_" + id;
+
+		return name;
+
+	}
 
-	  
-	  public static String print(SequenceI[] s)
-	  {
-	    return "not yet implemented";
-	  }
-	  
-	  public String print()
-	  {
-		System.out.print("affiche :");
-	    return print(getSeqsAsArray());
-	  }
-	  
-	  /**
-	   * make a friendly ID string.
-	   * 
-	   * @param dataName
-	   * @return truncated dataName to after last '/'
-	   */
-	  private String safeName(String dataName)
-	  {
-	    int b = 0;
-	    while ((b = dataName.indexOf("/")) > -1 && b < dataName.length() )
-	    {	    	   
-	    	dataName = dataName.substring(b + 1).trim();
-	    	
-	    }
-	    int e = (dataName.length() - dataName.indexOf("."))+1;
-	    dataName = dataName.substring(1,e).trim();
-	    return dataName;
-	  }
-	  
-	  public ArrayList getRNA()
-	  {
-		  return result;
-	  }
-	  
-	  public Vector getAnnot()
-	  {
-		  return annotations;
-	  }
 }
-- 
1.7.10.2