-/*\r
- * Jalview - A Sequence Alignment Editor and Viewer\r
- * Copyright (C) 2007 AM Waterhouse, J Procter, G Barton, M Clamp, S Searle\r
- *\r
- * This program is free software; you can redistribute it and/or\r
- * modify it under the terms of the GNU General Public License\r
- * as published by the Free Software Foundation; either version 2\r
- * of the License, or (at your option) any later version.\r
- *\r
- * This program is distributed in the hope that it will be useful,\r
- * but WITHOUT ANY WARRANTY; without even the implied warranty of\r
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\r
- * GNU General Public License for more details.\r
- *\r
- * You should have received a copy of the GNU General Public License\r
- * along with this program; if not, write to the Free Software\r
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA\r
- */\r
-package jalview.io;\r
-\r
-import java.io.*;\r
-import java.util.*;\r
-\r
-import jalview.datamodel.*;\r
-import jalview.util.*;\r
-\r
-/**\r
- * DOCUMENT ME!\r
- *\r
- * @author $author$\r
- * @version $Revision$\r
- */\r
-public class MSFfile\r
- extends AlignFile\r
-{\r
-\r
- /**\r
- * Creates a new MSFfile object.\r
- */\r
- public MSFfile()\r
- {\r
- }\r
-\r
- /**\r
- * Creates a new MSFfile object.\r
- *\r
- * @param inFile DOCUMENT ME!\r
- * @param type DOCUMENT ME!\r
- *\r
- * @throws IOException DOCUMENT ME!\r
- */\r
- public MSFfile(String inFile, String type)\r
- throws IOException\r
- {\r
- super(inFile, type);\r
- }\r
-\r
- /**\r
- * DOCUMENT ME!\r
- */\r
- public void parse()\r
- throws IOException\r
- {\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
- {\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
- {\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
- {\r
- seqFlag = true;\r
- }\r
-\r
- //Process lines as sequence lines if seqFlag is set\r
- if ( (inStr.indexOf("//") == -1) && (seqFlag == true))\r
- {\r
- //seqeunce id is the first field\r
- key = inStr;\r
-\r
- StringBuffer tempseq;\r
-\r
- //Get sequence from hash if it exists\r
- if (seqhash.containsKey(key))\r
- {\r
- tempseq = (StringBuffer) seqhash.get(key);\r
- }\r
- else\r
- {\r
- tempseq = new StringBuffer();\r
- seqhash.put(key, tempseq);\r
- }\r
-\r
- //loop through the rest of the words\r
- while (str.hasMoreTokens())\r
- {\r
- //append the word to the sequence\r
- tempseq.append(str.nextToken());\r
- }\r
- }\r
- }\r
- }\r
- }\r
- catch (IOException e)\r
- {\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
- {\r
- String head = headers.elementAt(i).toString();\r
- String seq = seqhash.get(head).toString();\r
-\r
- if (maxLength < head.length())\r
- {\r
- maxLength = head.length();\r
- }\r
-\r
- // Replace ~ with a sensible gap character\r
- seq = seq.replace('~', '-');\r
-\r
- Sequence newSeq = parseId(head);\r
-\r
- newSeq.setSequence(seq);\r
-\r
- seqs.addElement(newSeq);\r
- }\r
- else\r
- {\r
- System.err.println("MSFFile Parser: Can't find sequence for " +\r
- headers.elementAt(i));\r
- }\r
- }\r
- }\r
-\r
- /**\r
- * DOCUMENT ME!\r
- *\r
- * @param seq DOCUMENT ME!\r
- *\r
- * @return DOCUMENT ME!\r
- */\r
- public int checkSum(String seq)\r
- {\r
- int check = 0;\r
- String sequence = seq.toUpperCase();\r
-\r
- for (int i = 0; i < sequence.length(); i++)\r
- {\r
- try\r
- {\r
-\r
- int value = sequence.charAt(i);\r
- if (value != -1)\r
- {\r
- check += (i % 57 + 1) * value;\r
- }\r
- }\r
- catch (Exception e)\r
- {\r
- System.err.println("Exception during MSF Checksum calculation");\r
- e.printStackTrace();\r
- }\r
- }\r
-\r
- return check % 10000;\r
- }\r
-\r
- /**\r
- * DOCUMENT ME!\r
- *\r
- * @param s DOCUMENT ME!\r
- * @param is_NA DOCUMENT ME!\r
- *\r
- * @return DOCUMENT ME!\r
- */\r
- public String print(SequenceI[] seqs)\r
- {\r
-\r
- boolean is_NA = jalview.util.Comparison.isNucleotide(seqs);\r
-\r
- SequenceI[] s = new SequenceI[seqs.length];\r
-\r
- StringBuffer out = new StringBuffer("!!" + (is_NA ? "NA" : "AA") +\r
- "_MULTIPLE_ALIGNMENT 1.0\n\n"); // TODO: JBPNote : Jalview doesn't remember NA or AA yet.\r
-\r
- int max = 0;\r
- int maxid = 0;\r
- int i = 0;\r
-\r
- while ( (i < seqs.length) && (seqs[i] != null))\r
- {\r
- // Replace all internal gaps with . and external spaces with ~\r
- s[i] = new Sequence(seqs[i].getName(),\r
- seqs[i].getSequenceAsString().replace('-', '.'));\r
-\r
- StringBuffer sb = new StringBuffer();\r
- sb.append(s[i].getSequence());\r
-\r
- for (int ii = 0; ii < sb.length(); ii++)\r
- {\r
- if (sb.charAt(ii) == '.')\r
- {\r
- sb.setCharAt(ii, '~');\r
- }\r
- else\r
- {\r
- break;\r
- }\r
- }\r
-\r
- for (int ii = sb.length() - 1; ii > 0; ii--)\r
- {\r
- if (sb.charAt(ii) == '.')\r
- {\r
- sb.setCharAt(ii, '~');\r
- }\r
- else\r
- {\r
- break;\r
- }\r
- }\r
-\r
- s[i].setSequence(sb.toString());\r
-\r
- if (s[i].getSequence().length > max)\r
- {\r
- max = s[i].getSequence().length;\r
- }\r
-\r
- i++;\r
- }\r
-\r
- Format maxLenpad = new Format("%" + (new String("" + max)).length() +\r
- "d");\r
- Format maxChkpad = new Format("%" + (new String("1" + max)).length() +\r
- "d");\r
- i = 0;\r
-\r
- int bigChecksum = 0;\r
- int[] checksums = new int[s.length];\r
- while (i < s.length)\r
- {\r
- checksums[i] = checkSum(s[i].getSequenceAsString());\r
- bigChecksum += checksums[i];\r
- i++;\r
- }\r
-\r
- long maxNB = 0;\r
- out.append(" MSF: " + s[0].getSequence().length + " Type: " +\r
- (is_NA ? "N" : "P") + " Check: " + (bigChecksum % 10000) +\r
- " ..\n\n\n");\r
-\r
- String[] nameBlock = new String[s.length];\r
- String[] idBlock = new String[s.length];\r
-\r
- i = 0;\r
- while ( (i < s.length) && (s[i] != null))\r
- {\r
-\r
- nameBlock[i] = new String(" Name: " + printId(s[i]) + " ");\r
-\r
- idBlock[i] = new String("Len: " +\r
- maxLenpad.form(s[i].getSequence().length) +\r
- " Check: " +\r
- maxChkpad.form(checksums[i]) + " Weight: 1.00\n");\r
-\r
- if (s[i].getName().length() > maxid)\r
- {\r
- maxid = s[i].getName().length();\r
- }\r
-\r
- if (nameBlock[i].length() > maxNB)\r
- {\r
- maxNB = nameBlock[i].length();\r
- }\r
-\r
- i++;\r
- }\r
-\r
- if (maxid < 10)\r
- {\r
- maxid = 10;\r
- }\r
-\r
- if (maxNB < 15)\r
- {\r
- maxNB = 15;\r
- }\r
-\r
- Format nbFormat = new Format("%-" + maxNB + "s");\r
-\r
- for (i = 0; (i < s.length) && (s[i] != null); i++)\r
- {\r
- out.append(nbFormat.form(nameBlock[i]) + idBlock[i]);\r
- }\r
-\r
- maxid++;\r
- out.append("\n\n//\n\n");\r
-\r
- int len = 50;\r
-\r
- int nochunks = (max / len) + 1;\r
-\r
- if ( (max % len) == 0)\r
- {\r
- nochunks--;\r
- }\r
-\r
- for (i = 0; i < nochunks; i++)\r
- {\r
- int j = 0;\r
-\r
- while ( (j < s.length) && (s[j] != null))\r
- {\r
- String name = printId(s[j]);\r
-\r
- out.append(new Format("%-" + maxid + "s").form(name + " "));\r
-\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) &&\r
- (start < s[j].getSequence().length))\r
- {\r
- out.append(s[j].getSequence(start, end));\r
-\r
- if (k < 4)\r
- {\r
- out.append(" ");\r
- }\r
- else\r
- {\r
- out.append("\n");\r
- }\r
- }\r
- else\r
- {\r
- if (start < s[j].getSequence().length)\r
- {\r
- out.append(s[j].getSequenceAsString().substring(start));\r
- out.append("\n");\r
- }\r
- else\r
- {\r
- if (k == 0)\r
- {\r
- out.append("\n");\r
- }\r
- }\r
- }\r
- }\r
-\r
- j++;\r
- }\r
-\r
- out.append("\n");\r
- }\r
-\r
- return out.toString();\r
- }\r
-\r
- /**\r
- * DOCUMENT ME!\r
- *\r
- * @return DOCUMENT ME!\r
- */\r
- public String print()\r
- {\r
- return print(getSeqsAsArray());\r
- }\r
-}\r
+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
+ * 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/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
+package jalview.io;
+
+import jalview.datamodel.Sequence;
+import jalview.datamodel.SequenceI;
+import jalview.util.Comparison;
+import jalview.util.Format;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Hashtable;
+import java.util.List;
+import java.util.StringTokenizer;
+
+/**
+ * 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, DataSourceType type) throws IOException
+ {
+ super(inFile, type);
+ }
+
+ public MSFfile(FileParse source) throws IOException
+ {
+ super(source);
+ }
+
+ /**
+ * Read and parse MSF sequence data
+ */
+ @Override
+ public void parse() throws IOException
+ {
+ boolean seqFlag = false;
+ List<String> headers = new ArrayList<String>();
+ Hashtable<String, StringBuilder> seqhash = new Hashtable<String, StringBuilder>();
+
+ try
+ {
+ String line;
+ while ((line = nextLine()) != null)
+ {
+ StringTokenizer str = new StringTokenizer(line);
+
+ String key = null;
+ 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.add(key);
+ }
+
+ // if line has // set SeqFlag 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)
+ {
+ // sequence id is the first field
+ key = inStr;
+
+ StringBuilder tempseq;
+
+ // Get sequence from hash if it exists
+ if (seqhash.containsKey(key))
+ {
+ tempseq = seqhash.get(key);
+ }
+ else
+ {
+ tempseq = new StringBuilder(64);
+ seqhash.put(key, tempseq);
+ }
+
+ // loop through the rest of the words
+ while (str.hasMoreTokens())
+ {
+ // append the word to the sequence
+ String sequenceBlock = str.nextToken();
+ tempseq.append(sequenceBlock);
+ }
+ }
+ }
+ }
+ } catch (IOException e)
+ {
+ System.err.println("Exception parsing MSFFile " + e);
+ e.printStackTrace();
+ }
+
+ this.noSeqs = headers.size();
+
+ // Add sequences to the hash
+ for (int i = 0; i < headers.size(); i++)
+ {
+ if (seqhash.get(headers.get(i)) != null)
+ {
+ String head = headers.get(i);
+ String seq = seqhash.get(head).toString();
+
+ if (maxLength < head.length())
+ {
+ maxLength = head.length();
+ }
+
+ /*
+ * replace ~ (leading/trailing positions) with the gap character;
+ * use '.' as this is the internal gap character required by MSF
+ */
+ 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.get(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!
+ */
+ @Override
+ public String print(SequenceI[] sqs, boolean jvSuffix)
+ {
+
+ boolean is_NA = Comparison.isNucleotide(sqs);
+
+ SequenceI[] s = new SequenceI[sqs.length];
+
+ StringBuilder out = new StringBuilder(256);
+ out.append("!!").append(is_NA ? "NA" : "AA")
+ .append("_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 < sqs.length) && (sqs[i] != null))
+ {
+ /*
+ * modify to MSF format: uses '.' for internal gaps,
+ * and '~' for leading or trailing gaps
+ */
+ String seqString = sqs[i].getSequenceAsString().replace('-', '.');
+
+ StringBuilder sb = new StringBuilder(seqString);
+
+ 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] = new Sequence(sqs[i].getName(), sb.toString(),
+ sqs[i].getStart(), sqs[i].getEnd());
+
+ if (sb.length() > max)
+ {
+ max = sb.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].getLength() + " 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], jvSuffix) + " ");
+
+ idBlock[i] = new String("Len: " + maxLenpad.form(s[i].getLength())
+ + " 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) + (max % len > 0 ? 1 : 0);
+
+ for (i = 0; i < nochunks; i++)
+ {
+ int j = 0;
+
+ while ((j < s.length) && (s[j] != null))
+ {
+ String name = printId(s[j], jvSuffix);
+
+ 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;
+
+ int length = s[j].getLength();
+ if ((end < length)
+ && (start < length))
+ {
+ out.append(s[j].getSequence(start, end));
+
+ if (k < 4)
+ {
+ out.append(" ");
+ }
+ else
+ {
+ out.append(newline);
+ }
+ }
+ else
+ {
+ if (start < 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();
+ }
+}
git://source.jalview.org / jalview.git / blobdiff