-package jalview.io;\r
-\r
-import jalview.datamodel.*;\r
-import jalview.util.*;\r
-\r
-import java.io.*;\r
-import java.util.*;\r
-\r
-public class MSFfile extends AlignFile {\r
-\r
- public MSFfile()\r
- {}\r
-\r
- public MSFfile(String inStr) {\r
- super(inStr);\r
- }\r
-\r
- public MSFfile(String inFile, String type) throws IOException {\r
- super(inFile,type);\r
- }\r
-\r
- private static com.stevesoft.pat.Regex gapre = new com.stevesoft.pat.Regex("\\~","-");\r
- private static com.stevesoft.pat.Regex re2gap = new com.stevesoft.pat.Regex("["+jalview.util.Comparison.GapChars+"]","\\~");\r
-\r
- public void parse() {\r
- int i = 0;\r
- boolean seqFlag = false;\r
- String key = new String();\r
- Vector headers = new Vector();\r
- Hashtable seqhash = new Hashtable();\r
- String line;\r
-\r
- try {\r
- while ((line = nextLine()) != null) {\r
-\r
- StringTokenizer str = new StringTokenizer(line);\r
-\r
- while (str.hasMoreTokens()) {\r
-\r
- String inStr = str.nextToken();\r
-\r
- //If line has header information add to the headers vector\r
- if (inStr.indexOf("Name:") != -1) {\r
- key = str.nextToken();\r
- headers.addElement(key);\r
- }\r
-\r
- //if line has // set SeqFlag to 1 so we know sequences are coming\r
- if (inStr.indexOf("//") != -1) {\r
- seqFlag = true;\r
- }\r
-\r
- //Process lines as sequence lines if seqFlag is set\r
- if (( inStr.indexOf("//") == -1) && (seqFlag == true)) {\r
- //seqeunce id is the first field\r
- key = inStr;\r
- StringBuffer tempseq;\r
-\r
- //Get sequence from hash if it exists\r
- if (seqhash.containsKey(key)) {\r
- tempseq = (StringBuffer)seqhash.get(key);\r
- } else {\r
- tempseq = new StringBuffer();\r
- seqhash.put(key,tempseq);\r
- }\r
-\r
- //loop through the rest of the words\r
- while (str.hasMoreTokens()) {\r
- //append the word to the sequence\r
- tempseq.append(str.nextToken());\r
- }\r
- }\r
- }\r
- }\r
- } catch (IOException e) {\r
- System.err.println("Exception parsing MSFFile " + e);\r
- e.printStackTrace();\r
- }\r
-\r
- this.noSeqs = headers.size();\r
-\r
- //Add sequences to the hash\r
- for (i = 0; i < headers.size(); i++ ) {\r
-\r
- if ( seqhash.get(headers.elementAt(i)) != null) {\r
- String head = headers.elementAt(i).toString();\r
- String seq = seqhash.get(head).toString();\r
-\r
- int start = 1;\r
- int end = seq.length();\r
-\r
- if (maxLength < head.length() ) {\r
- maxLength = head.length();\r
- }\r
-\r
- if (head.indexOf("/") > 0 ) {\r
-\r
- StringTokenizer st = new StringTokenizer(head,"/");\r
-\r
- if (st.countTokens() == 2) {\r
-\r
- head = st.nextToken();\r
- String tmp = st.nextToken();\r
- st = new StringTokenizer(tmp,"-");\r
- if (st.countTokens() == 2) {\r
- start = Integer.valueOf(st.nextToken()).intValue();\r
- end = Integer.valueOf(st.nextToken()).intValue();\r
- }\r
- }\r
- }\r
- // Replace ~ with a sensible gap character\r
- seq = gapre.replaceAll(seq);\r
- Sequence newSeq = new Sequence(head,seq,start,end);\r
-\r
- seqs.addElement(newSeq);\r
-\r
- } else {\r
- System.err.println("MSFFile Parser: Can't find sequence for " + headers.elementAt(i));\r
- }\r
- }\r
-\r
- }\r
-\r
- public static int checkSum(String seq) {\r
- //String chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz.*~&@";\r
- int check = 0;\r
-\r
- String index = "--------------------------------------&---*---.-----------------@ABCDEFGHIJKLMNOPQRSTUVWXYZ------ABCDEFGHIJKLMNOPQRSTUVWXYZ----@";\r
- index += "--------------------------------------------------------------------------------------------------------------------------------";\r
-\r
- for(int i = 0; i < seq.length(); i++) {\r
- try {\r
- if (i <seq.length()) {\r
- int pos = index.indexOf(seq.substring(i,i+1));\r
- if (!index.substring(pos,pos+1).equals("_")) {\r
- check += ((i % 57) + 1) * pos;\r
- }\r
- }\r
- } catch (Exception e) {\r
- System.err.println("Exception during MSF Checksum calculation");\r
- e.printStackTrace();\r
- }\r
- }\r
- return check % 10000;\r
- }\r
-\r
- public static String print(SequenceI[] s) {\r
- StringBuffer out = new StringBuffer("PileUp\n\n");\r
-\r
- int max = 0;\r
- int maxid = 0;\r
-\r
- int i = 0;\r
- String big = "";\r
- while (i < s.length && s[i] != null) {\r
- big += s[i].getSequence();\r
- i++;\r
- }\r
- i = 0;\r
- int bigcheck = checkSum(big);\r
-\r
- out.append(" MSF: " + s[0].getSequence().length() + " Type: P Check: " + bigcheck + " ..\n\n\n");\r
-\r
- while (i < s.length && s[i] != null) {\r
- String seq = s[i].getSequence();\r
- String name = s[i].getName()+ "/" + s[i].getStart() + "-" + s[i].getEnd();\r
- int check = checkSum(s[i].getSequence());\r
- out.append(" Name: " + name + " oo Len: " + s[i].getSequence().length() + " Check: " + check + " Weight: 1.00\n");\r
- if (seq.length() > max) {\r
- max = seq.length();\r
- }\r
- if (name.length() > maxid) {\r
- maxid = name.length();\r
- }\r
- i++;\r
- }\r
-\r
- if (maxid < 10) {\r
- maxid = 10;\r
- }\r
- maxid++;\r
- out.append( "\n\n//\n\n");\r
-\r
- int len = 50;\r
-\r
- int nochunks = max / len + 1;\r
- if (max%len == 0) {\r
- nochunks--;\r
- }\r
- for (i = 0; i < nochunks; i++) {\r
- int j = 0;\r
- while (j < s.length && s[j] != null) {\r
- String name = s[j].getName();\r
- out.append( new Format("%-" + maxid + "s").form(name + "/" + s[j].getStart() + "-" + s[j].getEnd()) + " ");\r
- for (int k = 0; k < 5; k++) {\r
-\r
- int start = i*50 + k*10;\r
- int end = start + 10;\r
-\r
- if (end < s[j].getSequence().length() && start < s[j].getSequence().length() ) {\r
- out.append(re2gap.replaceAll(s[j].getSequence().substring(start,end)));\r
- if (k < 4) {\r
- // out.append(" ");\r
- } else {\r
- out.append("\n");\r
- }\r
- } else {\r
- if (start < s[j].getSequence().length()) {\r
- out.append(re2gap.replaceAll(s[j].getSequence().substring(start)));\r
- out.append("\n");\r
- } else {\r
- if (k == 0) {\r
- out.append("\n");\r
- }\r
- }\r
- }\r
- }\r
- j++;\r
- }\r
- out.append("\n");\r
-\r
- }\r
- return out.toString();\r
- }\r
- public String print() {\r
- return print(getSeqsAsArray());\r
- }\r
-}\r
-\r
-\r
-\r
-\r
-\r
-\r
-\r
+/*
+ * Jalview - A Sequence Alignment Editor and Viewer (Version 2.7)
+ * Copyright (C) 2011 J Procter, AM Waterhouse, J Engelhardt, LM Lui, G Barton, M Clamp, S Searle
+ *
+ * This file is part of Jalview.
+ *
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ * PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along with Jalview. If not, see <http://www.gnu.org/licenses/>.
+ */
+package jalview.io;
+
+import java.io.*;
+import java.util.*;
+
+import jalview.datamodel.*;
+import jalview.util.*;
+
+/**
+ * DOCUMENT ME!
+ *
+ * @author $author$
+ * @version $Revision$
+ */
+public class MSFfile extends AlignFile
+{
+
+ /**
+ * Creates a new MSFfile object.
+ */
+ public MSFfile()
+ {
+ }
+
+ /**
+ * Creates a new MSFfile object.
+ *
+ * @param inFile
+ * DOCUMENT ME!
+ * @param type
+ * DOCUMENT ME!
+ *
+ * @throws IOException
+ * DOCUMENT ME!
+ */
+ public MSFfile(String inFile, String type) throws IOException
+ {
+ super(inFile, type);
+ }
+
+ public MSFfile(FileParse source) throws IOException
+ {
+ super(source);
+ }
+
+ {
+ // TODO Auto-generated constructor stub
+ }
+
+ /**
+ * DOCUMENT ME!
+ */
+ public void parse() throws IOException
+ {
+ int i = 0;
+ boolean seqFlag = false;
+ String key = new String();
+ Vector headers = new Vector();
+ Hashtable seqhash = new Hashtable();
+ String line;
+
+ try
+ {
+ while ((line = nextLine()) != null)
+ {
+ StringTokenizer str = new StringTokenizer(line);
+
+ while (str.hasMoreTokens())
+ {
+ String inStr = str.nextToken();
+
+ // If line has header information add to the headers vector
+ if (inStr.indexOf("Name:") != -1)
+ {
+ key = str.nextToken();
+ headers.addElement(key);
+ }
+
+ // if line has // set SeqFlag to 1 so we know sequences are coming
+ if (inStr.indexOf("//") != -1)
+ {
+ seqFlag = true;
+ }
+
+ // Process lines as sequence lines if seqFlag is set
+ if ((inStr.indexOf("//") == -1) && (seqFlag == true))
+ {
+ // seqeunce id is the first field
+ key = inStr;
+
+ StringBuffer tempseq;
+
+ // Get sequence from hash if it exists
+ if (seqhash.containsKey(key))
+ {
+ tempseq = (StringBuffer) seqhash.get(key);
+ }
+ else
+ {
+ tempseq = new StringBuffer();
+ seqhash.put(key, tempseq);
+ }
+
+ // loop through the rest of the words
+ while (str.hasMoreTokens())
+ {
+ // append the word to the sequence
+ tempseq.append(str.nextToken());
+ }
+ }
+ }
+ }
+ } catch (IOException e)
+ {
+ System.err.println("Exception parsing MSFFile " + e);
+ e.printStackTrace();
+ }
+
+ this.noSeqs = headers.size();
+
+ // Add sequences to the hash
+ for (i = 0; i < headers.size(); i++)
+ {
+ if (seqhash.get(headers.elementAt(i)) != null)
+ {
+ String head = headers.elementAt(i).toString();
+ String seq = seqhash.get(head).toString();
+
+ if (maxLength < head.length())
+ {
+ maxLength = head.length();
+ }
+
+ // Replace ~ with a sensible gap character
+ seq = seq.replace('~', '-');
+
+ Sequence newSeq = parseId(head);
+
+ newSeq.setSequence(seq);
+
+ seqs.addElement(newSeq);
+ }
+ else
+ {
+ System.err.println("MSFFile Parser: Can't find sequence for "
+ + headers.elementAt(i));
+ }
+ }
+ }
+
+ /**
+ * DOCUMENT ME!
+ *
+ * @param seq
+ * DOCUMENT ME!
+ *
+ * @return DOCUMENT ME!
+ */
+ public int checkSum(String seq)
+ {
+ int check = 0;
+ String sequence = seq.toUpperCase();
+
+ for (int i = 0; i < sequence.length(); i++)
+ {
+ try
+ {
+
+ int value = sequence.charAt(i);
+ if (value != -1)
+ {
+ check += (i % 57 + 1) * value;
+ }
+ } catch (Exception e)
+ {
+ System.err.println("Exception during MSF Checksum calculation");
+ e.printStackTrace();
+ }
+ }
+
+ return check % 10000;
+ }
+
+ /**
+ * DOCUMENT ME!
+ *
+ * @param s
+ * DOCUMENT ME!
+ * @param is_NA
+ * DOCUMENT ME!
+ *
+ * @return DOCUMENT ME!
+ */
+ public String print(SequenceI[] seqs)
+ {
+
+ boolean is_NA = jalview.util.Comparison.isNucleotide(seqs);
+
+ SequenceI[] s = new SequenceI[seqs.length];
+
+ StringBuffer out = new StringBuffer("!!" + (is_NA ? "NA" : "AA")
+ + "_MULTIPLE_ALIGNMENT 1.0");
+ // TODO: JBPNote : Jalview doesn't remember NA or AA yet.
+ out.append(newline);
+ out.append(newline);
+ int max = 0;
+ int maxid = 0;
+ int i = 0;
+
+ while ((i < seqs.length) && (seqs[i] != null))
+ {
+ // Replace all internal gaps with . and external spaces with ~
+ s[i] = new Sequence(seqs[i].getName(), seqs[i].getSequenceAsString()
+ .replace('-', '.'), seqs[i].getStart(),seqs[i].getEnd());
+
+ StringBuffer sb = new StringBuffer();
+ sb.append(s[i].getSequence());
+
+ for (int ii = 0; ii < sb.length(); ii++)
+ {
+ if (sb.charAt(ii) == '.')
+ {
+ sb.setCharAt(ii, '~');
+ }
+ else
+ {
+ break;
+ }
+ }
+
+ for (int ii = sb.length() - 1; ii > 0; ii--)
+ {
+ if (sb.charAt(ii) == '.')
+ {
+ sb.setCharAt(ii, '~');
+ }
+ else
+ {
+ break;
+ }
+ }
+
+ s[i].setSequence(sb.toString());
+
+ if (s[i].getSequence().length > max)
+ {
+ max = s[i].getSequence().length;
+ }
+
+ i++;
+ }
+
+ Format maxLenpad = new Format("%" + (new String("" + max)).length()
+ + "d");
+ Format maxChkpad = new Format("%" + (new String("1" + max)).length()
+ + "d");
+ i = 0;
+
+ int bigChecksum = 0;
+ int[] checksums = new int[s.length];
+ while (i < s.length)
+ {
+ checksums[i] = checkSum(s[i].getSequenceAsString());
+ bigChecksum += checksums[i];
+ i++;
+ }
+
+ long maxNB = 0;
+ out.append(" MSF: " + s[0].getSequence().length + " Type: "
+ + (is_NA ? "N" : "P") + " Check: " + (bigChecksum % 10000)
+ + " ..");
+ out.append(newline);
+ out.append(newline);
+ out.append(newline);
+
+ String[] nameBlock = new String[s.length];
+ String[] idBlock = new String[s.length];
+
+ i = 0;
+ while ((i < s.length) && (s[i] != null))
+ {
+
+ nameBlock[i] = new String(" Name: " + printId(s[i]) + " ");
+
+ idBlock[i] = new String("Len: "
+ + maxLenpad.form(s[i].getSequence().length) + " Check: "
+ + maxChkpad.form(checksums[i]) + " Weight: 1.00"+newline);
+
+ if (s[i].getName().length() > maxid)
+ {
+ maxid = s[i].getName().length();
+ }
+
+ if (nameBlock[i].length() > maxNB)
+ {
+ maxNB = nameBlock[i].length();
+ }
+
+ i++;
+ }
+
+ if (maxid < 10)
+ {
+ maxid = 10;
+ }
+
+ if (maxNB < 15)
+ {
+ maxNB = 15;
+ }
+
+ Format nbFormat = new Format("%-" + maxNB + "s");
+
+ for (i = 0; (i < s.length) && (s[i] != null); i++)
+ {
+ out.append(nbFormat.form(nameBlock[i]) + idBlock[i]);
+ }
+
+ maxid++;
+ out.append(newline);
+ out.append(newline);out.append("//");
+ out.append(newline);
+ out.append(newline);
+ int len = 50;
+
+ int nochunks = (max / len) + 1;
+
+ if ((max % len) == 0)
+ {
+ nochunks--;
+ }
+
+ for (i = 0; i < nochunks; i++)
+ {
+ int j = 0;
+
+ while ((j < s.length) && (s[j] != null))
+ {
+ String name = printId(s[j]);
+
+ out.append(new Format("%-" + maxid + "s").form(name + " "));
+
+ for (int k = 0; k < 5; k++)
+ {
+ int start = (i * 50) + (k * 10);
+ int end = start + 10;
+
+ if ((end < s[j].getSequence().length)
+ && (start < s[j].getSequence().length))
+ {
+ out.append(s[j].getSequence(start, end));
+
+ if (k < 4)
+ {
+ out.append(" ");
+ }
+ else
+ {
+ out.append(newline);
+ }
+ }
+ else
+ {
+ if (start < s[j].getSequence().length)
+ {
+ out.append(s[j].getSequenceAsString().substring(start));
+ out.append(newline);
+ }
+ else
+ {
+ if (k == 0)
+ {
+ out.append(newline);
+ }
+ }
+ }
+ }
+
+ j++;
+ }
+
+ out.append(newline);
+ }
+
+ return out.toString();
+ }
+
+ /**
+ * DOCUMENT ME!
+ *
+ * @return DOCUMENT ME!
+ */
+ public String print()
+ {
+ return print(getSeqsAsArray());
+ }
+}
git://source.jalview.org / jalview.git / blobdiff