HelpIndex


1   /*
2    * HelpIndex.java - Index for help searching feature
3    * :tabSize=8:indentSize=8:noTabs=false:
4    * :folding=explicit:collapseFolds=1:
5    *
6    * Copyright (C) 2002 Slava Pestov
7    *
8    * This program is free software; you can redistribute it and/or
9    * modify it under the terms of the GNU General Public License
10   * as published by the Free Software Foundation; either version 2
11   * of the License, or any later version.
12   *
13   * This program is distributed in the hope that it will be useful,
14   * but WITHOUT ANY WARRANTY; without even the implied warranty of
15   * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
16   * GNU General Public License for more details.
17   *
18   * You should have received a copy of the GNU General Public License
19   * along with this program; if not, write to the Free Software
20   * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
21   */
22  
23  package org.gjt.sp.jedit.help;
24  
25  //{{{ Imports
26  import java.io.*;
27  import java.net.*;
28  import java.util.zip.*;
29  import java.util.*;
30  import org.gjt.sp.jedit.io.*;
31  import org.gjt.sp.jedit.*;
32  import org.gjt.sp.util.Log;
33  //}}}
34  
35  class HelpIndex
36  {
37      //{{{ HelpIndex constructor
38      public HelpIndex()
39      {
40          words = new HashMap();
41          files = new ArrayList();
42  
43          ignoreWord("a");
44          ignoreWord("an");
45          ignoreWord("and");
46          ignoreWord("are");
47          ignoreWord("as");
48          ignoreWord("be");
49          ignoreWord("by");
50          ignoreWord("can");
51          ignoreWord("do");
52          ignoreWord("for");
53          ignoreWord("from");
54          ignoreWord("how");
55          ignoreWord("i");
56          ignoreWord("if");
57          ignoreWord("in");
58          ignoreWord("is");
59          ignoreWord("it");
60          ignoreWord("not");
61          ignoreWord("of");
62          ignoreWord("on");
63          ignoreWord("or");
64          ignoreWord("s");
65          ignoreWord("that");
66          ignoreWord("the");
67          ignoreWord("this");
68          ignoreWord("to");
69          ignoreWord("will");
70          ignoreWord("with");
71          ignoreWord("you");
72      } //}}}
73  
74      /* //{{{ HelpIndex constructor
75      public HelpIndex(String fileListPath, String wordIndexPath)
76      {
77          this();
78      } //}}} */
79  
80      //{{{ indexEditorHelp() method
81      /**
82       * Indexes all available help, including the jEdit user's guide, FAQ,]
83       * and plugin documentation.
84       */
85      public void indexEditorHelp()
86      {
87          try
88          {
89              String   jEditHome = jEdit.getJEditHome();
90              if(jEditHome != null)
91              {
92                  indexDirectory(MiscUtilities.constructPath(jEditHome,"doc","users-guide"));
93                  indexDirectory(MiscUtilities.constructPath(jEditHome,"doc","FAQ"));
94                  indexDirectory(MiscUtilities.constructPath(jEditHome,"doc","news43"));
95              }
96          }
97          catch(Throwable   e)
98          {
99              Log.log(Log.ERROR,this,"Error indexing editor help");
100             Log.log(Log.ERROR,this,e);
101         }
102 
103         PluginJAR[] jars = jEdit.getPluginJARs();
104         for(int i = 0; i < jars.length; i++)
105         {
106             try
107             {
108                 indexJAR(jars[i].getZipFile());
109             }
110             catch(Throwable   e)
111             {
112                 Log.log(Log.ERROR,this,"Error indexing JAR: "
113                     + jars[i].getPath());
114                 Log.log(Log.ERROR,this,e);
115             }
116         }
117 
118         Log.log(Log.DEBUG,this,"Indexed " + words.size() + " words");
119     } //}}}
120 
121     //{{{ indexDirectory() method
122     /**
123      * Indexes all HTML and text files in the specified directory.
124      * @param dir The directory
125      */
126     public void indexDirectory(String   dir) throws Exception  
127     {
128         String  [] files = VFSManager.getFileVFS()
129             ._listDirectory(null,dir,"*.{html,txt}",true,null);
130 
131         for(int i = 0; i < files.length; i++)
132         {
133             indexURL(files[i]);
134         }
135     } //}}}
136 
137     //{{{ indexJAR() method
138     /**
139      * Indexes all HTML and text files in the specified JAR file.
140      * @param jar The JAR file
141      */
142     public void indexJAR(ZipFile jar) throws Exception  
143     {
144         Enumeration e = jar.entries();
145         while(e.hasMoreElements())
146         {
147             ZipEntry entry = (ZipEntry)e.nextElement();
148             String   name = entry.getName();
149             String   lname = name.toLowerCase();
150             if(lname.endsWith(".html")/*  || lname.endsWith(".txt") */)
151             {
152                 // only works for jEdit plugins
153                 String   url = "jeditresource:/" +
154                     MiscUtilities.getFileName(jar.getName())
155                     + "!/" + name;
156                 Log.log(Log.DEBUG,this,url);
157                 indexStream(jar.getInputStream(entry),url);
158             }
159         }
160     } //}}}
161 
162     //{{{ indexURL() method
163     /**
164      * Reads the specified HTML file and adds all words defined therein to the
165      * index.
166      * @param url The HTML file's URL
167      */
168     public void indexURL(String   url) throws Exception  
169     {
170         InputStream _in;
171 
172         if(MiscUtilities.isURL(url))
173             _in =  new URL(url).openStream();
174         else
175         {
176             _in = new FileInputStream(url);
177             // hack since HelpViewer needs a URL...
178             url = "file:" + url;
179         }
180 
181         indexStream(_in,url);
182     } //}}}
183 
184     //{{{ lookupWord() method
185     public Word lookupWord(String   word)
186     {
187         Object   o = words.get(word);
188         if(o == IGNORE)
189             return null;
190         else
191             return (Word)o;
192     } //}}}
193 
194     //{{{ getFile() method
195     public HelpFile getFile(int index)
196     {
197         return (HelpFile)files.get(index);
198     } //}}}
199 
200     //{{{ Private members
201     // used to mark words to ignore (see constructor for the list)
202     private static Object   IGNORE = new Object  ();
203     private HashMap words;
204     private ArrayList files;
205 
206     //{{{ ignoreWord() method
207     private void ignoreWord(String   word)
208     {
209         words.put(word,IGNORE);
210     } //}}}
211 
212     //{{{ indexStream() method
213     /**
214      * Reads the specified HTML file and adds all words defined therein to the
215      * index.
216      * @param _in The input stream
217      * @param file The file
218      */
219     private void indexStream(InputStream _in, String   fileName)
220         throws Exception  
221     {
222         HelpFile file = new HelpFile(fileName);
223         files.add(file);
224         int index = files.size() - 1;
225 
226         StringBuffer   titleText = new StringBuffer  ();
227 
228         BufferedReader in = new BufferedReader(
229             new InputStreamReader(_in));
230 
231         try
232         {
233             StringBuffer   word = new StringBuffer  ();
234             boolean insideTag = false;
235             boolean insideEntity = false;
236 
237             boolean title = false;
238 
239             int c;
240             while((c = in.read()) != -1)
241             {
242                 char ch = (char)c;
243                 if(insideTag)
244                 {
245                     if(ch == '>')
246                     {
247                         if(word.toString().equals("title"))
248                             title = true;
249                         insideTag = false;
250                         word.setLength(0);
251                     }
252                     else
253                         word.append(ch);
254                 }
255                 else if(insideEntity)
256                 {
257                     if(ch == ';')
258                         insideEntity = false;
259                 }
260                 else if(ch == '<')
261                 {
262                     if(title)
263                         title = false;
264 
265                     if(word.length() != 0)
266                     {
267                         addWord(word.toString(),index,title);
268                         word.setLength(0);
269                     }
270 
271                     insideTag = true;
272                 }
273                 else if(ch == '&')
274                     insideEntity = true;
275                 else if(title)
276                     titleText.append(ch);
277                 else if(!Character.isLetterOrDigit(ch))
278                 {
279                     if(word.length() != 0)
280                     {
281                         addWord(word.toString(),index,title);
282                         word.setLength(0);
283                     }
284                 }
285                 else
286                     word.append(ch);
287             }
288         }
289         finally
290         {
291             in.close();
292         }
293 
294         if(titleText.length() == 0)
295             file.title = fileName;
296         else
297             file.title = titleText.toString();
298     } //}}}
299 
300     //{{{ addWord() method
301     private void addWord(String   word, int file, boolean title)
302     {
303         word = word.toLowerCase();
304 
305         Object   o = words.get(word);
306         if(o == IGNORE)
307             return;
308 
309         if(o == null)
310             words.put(word,new Word(word,file,title));
311         else
312             ((Word)o).addOccurrence(file,title);
313     } //}}}
314 
315     //}}}
316 
317     //{{{ Word class
318     static class Word
319     {
320         // how much an occurrence in the title is worth
321         static final int TITLE_OCCUR = 10;
322 
323         // the word
324         String   word;
325 
326         // files it occurs in
327         int occurCount = 0;
328         Occurrence[] occurrences;
329 
330         Word(String   word, int file, boolean title)
331         {
332             this.word = word;
333             occurrences = new Occurrence[5];
334             addOccurrence(file,title);
335         }
336 
337         void addOccurrence(int file, boolean title)
338         {
339             for(int i = 0; i < occurCount; i++)
340             {
341                 if(occurrences[i].file == file)
342                 {
343                     occurrences[i].count += (title ? TITLE_OCCUR : 1);
344                     return;
345                 }
346             }
347 
348             if(occurCount >= occurrences.length)
349             {
350                 Occurrence[] newOccur = new Occurrence[occurrences.length * 2];
351                 System.arraycopy(occurrences,0,newOccur,0,occurCount);
352                 occurrences = newOccur;
353             }
354 
355             occurrences[occurCount++] = new Occurrence(file,title);
356         }
357 
358         static class Occurrence
359         {
360             int file;
361             int count;
362 
363             Occurrence(int file, boolean title)
364             {
365                 this.file = file;
366                 this.count = (title ? TITLE_OCCUR : 1);
367             }
368         }
369     } //}}}
370 
371     //{{{ HelpFile class
372     static class HelpFile
373     {
374         String   file;
375         String   title;
376 
377         HelpFile(String   file)
378         {
379             this.file = file;
380         }
381 
382         public String   toString()
383         {
384             return title;
385         }
386 
387         public boolean equals(Object   o)
388         {
389             if(o instanceof HelpFile)
390                 return ((HelpFile)o).file.equals(file);
391             else
392                 return false;
393         }
394     } //}}}
395 }
396
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags