001: // plasmaWordCon.java
002: // -----------------------
003: // part of The Kelondro Database
004: // (C) by Michael Peter Christen; mc@anomic.de
005: // first published on http://www.anomic.de
006: // Frankfurt, Germany, 2004
007: // last major change: 22.09.2004
008: //
009: // This program is free software; you can redistribute it and/or modify
010: // it under the terms of the GNU General Public License as published by
011: // the Free Software Foundation; either version 2 of the License, or
012: // (at your option) any later version.
013: //
014: // This program is distributed in the hope that it will be useful,
015: // but WITHOUT ANY WARRANTY; without even the implied warranty of
016: // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
017: // GNU General Public License for more details.
018: //
019: // You should have received a copy of the GNU General Public License
020: // along with this program; if not, write to the Free Software
021: // Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
022: //
023: // Using this software in any meaning (reading, learning, copying, compiling,
024: // running) means that you agree that the Author(s) is (are) not responsible
025: // for cost, loss of data or any harm that may be caused directly or indirectly
026: // by usage of this softare or this documentation. The usage of this software
027: // is on your own risk. The installation and usage (starting/running) of this
028: // software may allow other people or application to access your computer and
029: // any attached devices and is highly dependent on the configuration of the
030: // software which must be done by the user of the software; the author(s) is
031: // (are) also not responsible for proper configuration and usage of the
032: // software, even if provoked by documentation provided together with
033: // the software.
034: //
035: // Any changes to this file according to the GPL as documented in the file
036: // gpl.txt aside this file in the shipment you received can be done to the
037: // lines that follows this copyright notice here, but changes must not be
038: // done inside the copyright notive above. A re-distribution must contain
039: // the intact and unchanged copyright notice.
040: // Contributions and changes to the program code must be marked as such.
041:
042: /*
043: */
044:
045: package de.anomic.plasma;
046:
047: import java.io.File;
048: import java.io.IOException;
049:
050: import de.anomic.kelondro.kelondroDynTree;
051: import de.anomic.kelondro.kelondroNaturalOrder;
052: import de.anomic.kelondro.kelondroRow;
053:
054: public class plasmaWordConnotation {
055:
056: private static final int wordlength = 32;
057: private static final int countlength = 4;
058: private static final int nodesize = 4048;
059: private kelondroDynTree refDB;
060:
061: public plasmaWordConnotation(File refDBfile, long preloadTime,
062: char fillChar) {
063: refDB = new kelondroDynTree(refDBfile, preloadTime, wordlength,
064: nodesize, new kelondroRow("byte[] word-" + wordlength
065: + ", Cardinal count-" + countlength,
066: kelondroNaturalOrder.naturalOrder, 0),
067: fillChar, true);
068: }
069:
070: private void addSingleRef(String word, String reference)
071: throws IOException {
072: //word = word.toLowerCase();
073: //reference = reference.toLowerCase();
074: kelondroRow.Entry record = refDB
075: .get(word, reference.getBytes());
076: long c;
077: if (record == null)
078: c = 0;
079: else
080: c = record.getColLong(1);
081: record.setCol(1, c++);
082: refDB.put(word, record);
083: }
084:
085: public void addSentence(String[] words) throws IOException {
086: for (int i = 0; i < words.length; i++)
087: words[i] = words[i].toLowerCase();
088: for (int i = 0; i < words.length; i++) {
089: for (int j = 0; j < words.length; j++) {
090: if ((i != j) && (words[i].length() > 2)
091: && (words[j].length() > 2))
092: addSingleRef(words[i], words[j]);
093: }
094: }
095: }
096:
097: public void addSentence(String sentence) throws IOException {
098: addSentence(sentence.split(" "));
099: }
100:
101: /*
102: public String[] getConnotation(String word, int count) {
103: TreeMap map = new TreeMap();
104: return null;
105: }
106: */
107: }
|