|
1
43 package net.jforum.dao.generic;
44
45 import java.sql.Connection  ;
46 import java.sql.PreparedStatement ;
47 import java.sql.ResultSet ;
48 import java.sql.SQLException ;
49 import java.sql.Statement ;
50 import java.util.ArrayList ;
51 import java.util.Arrays ;
52 import java.util.HashSet ;
53 import java.util.Iterator ;
54 import java.util.List ;
55 import java.util.Set ;
56 import java.util.StringTokenizer ;
57 import net.jforum.entities.Post;
58 import net.jforum.util.preferences.ConfigKeys;
59 import net.jforum.util.preferences.SystemGlobals;
60 import org.apache.log4j.Logger;
61
62
66 public class GenericSearchIndexerDAO extends AutoKeys implements net.jforum.dao.SearchIndexerDAO
67 {
68 private static final Logger log = Logger.getLogger(GenericSearchIndexerDAO.class);
69
70 private Connection conn;
71
72
75 public void setConnection(Connection conn)
76 {
77 this.conn = conn;
78 }
79
80
83 public void insertSearchWords(List posts) throws Exception
84 {
85 int minWordSize = SystemGlobals.getIntValue(ConfigKeys.SEARCH_MIN_WORD_SIZE);
86 int maxWordSize = SystemGlobals.getIntValue(ConfigKeys.SEARCH_MAX_WORD_SIZE);
87 int searchMaxWordsMessage = SystemGlobals.getIntValue(ConfigKeys.SEARCH_MAX_WORDS_MESSAGE);
88 String wordFilterRegex = SystemGlobals.getValue(ConfigKeys.SEARCH_WORD_FILTER_REGEX);
89 StringBuffer sb = new StringBuffer (512);
90
91 String excludeWordsString = SystemGlobals.getValue("search.exclude.words");
93 Set excludeWords = new HashSet ();
94 if (excludeWordsString != null) {
95 String [] words = excludeWordsString.split(",");
96 for (int i = 0; words != null && i < words.length; i++) {
97 excludeWords.add(words[i].trim());
98 }
99 }
100
101 String matchSql = SystemGlobals.getSql("SearchModel.associateWordToPost");
102 PreparedStatement words = this.conn.prepareStatement(SystemGlobals.getSql("SearchModel.insertWords"));
103
104 for (Iterator iter = posts.iterator(); iter.hasNext(); ) {
105 Post p = (Post)iter.next();
106
107 String text = new StringBuffer (p.getText()).append(" ")
108 .append(p.getSubject()).toString();
109
110 text = text.toLowerCase().replaceAll("[\\.\\\\\\/~\\^\\&\\(\\)\\-_+=!@;#\\$%\"\'\\[\\]\\{\\}\\?<\\:>,\\*\n\r\t]", " ");
111
112 Set allWords = new HashSet ();
113
114 sb.delete(0, sb.length());
115
116 StringTokenizer st = new StringTokenizer (text, " ");
117
118 while (st.hasMoreTokens() &&
120 (searchMaxWordsMessage < 1 || allWords.size() < searchMaxWordsMessage)) {
121 String w = st.nextToken().trim();
122
123 if (w.length() < minWordSize) {
124 continue;
125 }
126 else if (w.length() > maxWordSize) {
127 w = w.substring(0, maxWordSize);
128 }
129
130 if (!allWords.contains(w) && !excludeWords.contains(w) &&
131 (wordFilterRegex == null || w.matches(wordFilterRegex))) {
132 allWords.add(w);
133 sb.append('\'').append(w).append('\'').append(",");
134 }
135 }
136
137 String in = sb.substring(0, sb.length() - 1);
138
139 String sql = SystemGlobals.getSql("SearchModel.selectExistingWords");
140 sql = sql.replaceAll("#IN#", in);
141
142 Statement s = this.conn.createStatement();
143 ResultSet rs = s.executeQuery(sql);
144
145 List newWords = new ArrayList ();
146
147 while (rs.next()) {
148 newWords.add(rs.getString("word"));
149 }
150
151 rs.close();
152 s.close();
153
154 allWords.removeAll(newWords);
155
156 for (Iterator witer = allWords.iterator(); witer.hasNext(); ) {
158 String ww = (String )witer.next();
159
160 words.setString(1, ww);
161 words.setInt(2, ww.hashCode());
162
163 try {
164 words.executeUpdate();
165 }
166 catch (SQLException e) {
167 log.error("Cannot index word: \"" + ww + "\"", e);
168 throw e;
169 }
170 }
171
172 sql = matchSql.replaceAll("#ID#", String.valueOf(p.getId())).replaceAll("#IN#", in);
173
174 Statement match = this.conn.createStatement();
175 match.executeUpdate(sql);
176 match.close();
177 }
178
179 words.close();
180 }
181
182
185 public void insertSearchWords(final Post post) throws Exception
186 {
187 this.insertSearchWords(Arrays.asList(new Post[] { post }));
188 }
189 }
190 |
Popular Tags |
Java API By Example, From Geeks To Geeks.