Add remove contaminants database by reading fasta file and compared accession

Add compattibility of contaminants selection for Mascot

git-svn-id: https://subversion.renater.fr/xtandempipeline/trunk@205 b8ef2a07-7df7-436f-90b9-41648038564b

xtandempipeline/src/fr/inra/pappso/xtandempipeline/class_msms/config.java

@@ -178,7 +178,7 @@ public class config {
 	}
 
 	public void set_database_filter(File F) {
-		properties.setProperty(database_filter, F.getName());
+		properties.setProperty(database_filter, F.getPath());
 	}
 
 	public void set_protein_evalue(float F) {

xtandempipeline/src/fr/inra/pappso/xtandempipeline/grouping/GroupSet.java

 package fr.inra.pappso.xtandempipeline.grouping;
 
+import java.io.BufferedReader;
 import java.io.File;
+import java.io.FileInputStream;
+import java.io.InputStreamReader;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.HashSet;
@@ -90,23 +93,61 @@ public class GroupSet {
 	}
 
 	public void removeGroupFromDatabase(File database) {
-		// TODO Remove Group from this database
-		logger.info("Remove Group from Database: " + database.getName());
-		for (Group group : this.getGroupList()) {
-			boolean isconta = false;
-			for (SubGroup sg : group.getSubGroupSet()) {
-				for (HashSampleScanSetProt hashProt : sg
-						.getHashSampleScanSetProtSet()) {
-					if (hashProt.getMatch().get_protein_match().getDatabase().getDatabasePath()
-							.contains(database.getName())) {
-						isconta = true;
+		//Read fasta file to get accession list
+		if(database.exists()){
+			logger.info("Read fasta file : "+database.getName());
+			HashSet<String> access = new HashSet<String>();
+			try{
+				FileInputStream is = new FileInputStream(database);
+				InputStreamReader isr = new InputStreamReader(is);
+				BufferedReader br = new BufferedReader(isr);
+				
+				String line;
+				while((line= br.readLine())!=null){
+					line.trim();
+					if(line.length()>0){
+						if(line.startsWith(">")){
+							access.add(line.split(" ")[0].replaceFirst(">", ""));
+						}
 					}
 				}
+			}catch (Exception e) {
+				logger.error(e);
+			}
+			//Search contaminants proteins
+			logger.info("Remove Group from Database in new mode: " + database.getName());
+			for (Group group : this.getGroupList()) {
+				boolean isconta = false;
+				for (SubGroup sg : group.getSubGroupSet()) {
+					for (HashSampleScanSetProt hashProt : sg
+							.getHashSampleScanSetProtSet()) {
+						if (access.contains(hashProt.getMatch().get_protein_match().get_cle_accession())) {
+							isconta = true;
+						}
+					}
+				}
+				if (isconta)
+					this.groupList.remove(group);
+			}
+		}else{
+			logger.error("fasta file doesnt not exist: " + database.getAbsolutePath());
+			logger.info("Remove Group from Database in old mode: " + database.getName());
+			for (Group group : this.getGroupList()) {
+				boolean isconta = false;
+				for (SubGroup sg : group.getSubGroupSet()) {
+					for (HashSampleScanSetProt hashProt : sg
+							.getHashSampleScanSetProtSet()) {
+						if (hashProt.getMatch().get_protein_match().getDatabase().getDatabasePath()
+								.contains(database.getName())) {
+							isconta = true;
+						}
+					}
+				}
+				if (isconta)
+					this.groupList.remove(group);
 			}
-			if (isconta)
-				this.groupList.remove(group);
 		}
-		//On renummerote après avoir modifier les groupes
+		//Numbering after contaminant removing
 		this.numbering();
 	}