2 * Jalview - A Sequence Alignment Editor and Viewer
\r
3 * Copyright (C) 2005 AM Waterhouse, J Procter, G Barton, M Clamp, S Searle
\r
5 * This program is free software; you can redistribute it and/or
\r
6 * modify it under the terms of the GNU General Public License
\r
7 * as published by the Free Software Foundation; either version 2
\r
8 * of the License, or (at your option) any later version.
\r
10 * This program is distributed in the hope that it will be useful,
\r
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
\r
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
\r
13 * GNU General Public License for more details.
\r
15 * You should have received a copy of the GNU General Public License
\r
16 * along with this program; if not, write to the Free Software
\r
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
\r
22 import jalview.datamodel.*;
\r
23 import jalview.util.*;
\r
28 public class PfamFile extends AlignFile {
\r
35 public PfamFile(String inStr) {
\r
39 public void initData() {
\r
44 public PfamFile(String inFile, String type) throws IOException {
\r
48 public void parse() throws IOException{
\r
53 Hashtable seqhash = new Hashtable();
\r
54 Vector headers = new Vector();
\r
57 while ((line = nextLine()) != null)
\r
60 if (line.indexOf(" ") != 0)
\r
62 if (line.indexOf("#") != 0)
\r
65 StringTokenizer str = new StringTokenizer(line," ");
\r
68 if (str.hasMoreTokens())
\r
70 id = str.nextToken();
\r
72 StringBuffer tempseq;
\r
74 if (seqhash.containsKey(id))
\r
75 tempseq = (StringBuffer)seqhash.get(id);
\r
78 tempseq = new StringBuffer();
\r
79 seqhash.put(id,tempseq);
\r
82 if (!(headers.contains(id)))
\r
83 headers.addElement(id);
\r
86 tempseq.append(str.nextToken());
\r
92 this.noSeqs = headers.size();
\r
94 throw new IOException("No sequences found (PFAM input)");
\r
96 for (i = 0; i < headers.size(); i++ ) {
\r
98 if ( seqhash.get(headers.elementAt(i)) != null) {
\r
99 if (maxLength < seqhash.get(headers.elementAt(i)).toString().length() )
\r
100 maxLength = seqhash.get(headers.elementAt(i)).toString().length();
\r
102 String head = headers.elementAt(i).toString();
\r
104 int end = seqhash.get(headers.elementAt(i)).toString().length();
\r
106 if (head.indexOf("/") > 0 ) {
\r
107 StringTokenizer st = new StringTokenizer(head,"/");
\r
108 if (st.countTokens() == 2) {
\r
109 ids.addElement(st.nextToken());
\r
110 String tmp = st.nextToken();
\r
111 st = new StringTokenizer(tmp,"-");
\r
112 if (st.countTokens() == 2) {
\r
113 start = Integer.valueOf(st.nextToken()).intValue();
\r
114 end = Integer.valueOf(st.nextToken()).intValue();
\r
121 ids.addElement(headers.elementAt(i));
\r
125 ids.addElement(headers.elementAt(i));
\r
128 Sequence newSeq = null;
\r
129 if (start != -1 && end != -1)
\r
131 newSeq = new Sequence(ids.elementAt(i).toString(),
\r
132 seqhash.get(headers.elementAt(i).toString()).toString(),start,end);
\r
133 seqs.addElement(newSeq);
\r
137 newSeq = new Sequence(ids.elementAt(i).toString(),
\r
138 seqhash.get(headers.elementAt(i).toString()).toString(),1,
\r
139 seqhash.get(headers.elementAt(i).toString()).toString().length());
\r
140 seqs.addElement(newSeq);
\r
143 if(!isValidProteinSequence(newSeq.getSequence()))
\r
144 throw new IOException("Not a valid protein sequence - (PFAM input)");
\r
147 System.err.println("PFAM File reader: Can't find sequence for " + headers.elementAt(i));
\r
153 public static String print(SequenceI[] s) {
\r
154 StringBuffer out = new StringBuffer("");
\r
161 while (i < s.length && s[i] != null) {
\r
162 String tmp = s[i].getName() + "/" + s[i].getStart()+ "-" + s[i].getEnd();
\r
164 if (s[i].getSequence().length() > max) {
\r
165 max = s[i].getSequence().length();
\r
167 if (tmp.length() > maxid) {
\r
168 maxid = tmp.length();
\r
178 while ( j < s.length && s[j] != null) {
\r
179 out.append( new Format("%-" + maxid + "s").form(s[j].getName() + "/" + s[j].getStart() + "-" + s[j].getEnd() ) + " ");
\r
181 out.append(s[j].getSequence() + "\n");
\r
186 return out.toString();
\r
189 public String print() {
\r
190 return print(getSeqsAsArray());
\r