TupleSet


1   /*
2    * (c) Copyright 2003, 2004, 2005 Hewlett-Packard Development Company, LP
3    * [See end of file]
4    */
5    
6   package com.hp.hpl.jena.util.tuple ;
7   
8   import java.io.* ;
9   import java.util.* ;
10  
11  import org.apache.commons.logging.Log;
12  import org.apache.commons.logging.LogFactory;
13  
14  /**
15   * @author  Andy Seaborne
16   * @version $Id: TupleSet.java,v 1.6 2005/02/21 12:19:27 andy_seaborne Exp $
17   */
18  
19  public class TupleSet implements Iterator
20  {
21      BufferedReader in ;
22      public String   line = null ;
23      public int lineNumber = 0 ;
24  
25      static final char COMMENTCHAR = '#' ;
26      List current = null ;
27      boolean finished = false ;
28  
29      protected static Log logger = LogFactory.getLog( TupleSet.class );
30      
31      /** Creates new TupleSet */
32      public TupleSet(Reader r)
33      {
34          if ( ! ( r instanceof BufferedReader ) )
35              in = new BufferedReader(r) ;
36          else
37              in = (BufferedReader)r;
38      }
39  
40      public boolean hasNext()
41      {
42          if ( finished ) return false ;
43  
44          if ( current == null )
45              current = tuple() ;
46          return current != null ;
47      }
48  
49      public Object   next()
50      {
51          if ( hasNext() )
52          {
53              List x = current ;
54              current = null ;
55              return x ;
56          }
57          else
58              return null ;
59      }
60  
61  
62      public void remove()
63      {
64          throw new java.lang.UnsupportedOperationException  ("TupleSet.remove") ;
65      }
66  
67      private List tuple()
68      {
69  
70          try {
71              lineNumber ++ ;
72              line = in.readLine() ;
73          } catch (IOException e) {}
74  
75          if ( line == null )
76          {
77              finished = true ;
78              return null ;
79          }
80  
81          //System.out.println("Line: "+line) ;
82          List tuple = new ArrayList() ;
83          int i = 0 ;
84          int j = 0 ;
85          boolean errorFound = false ;
86  
87       tupleLoop:
88          for (;;)
89          {
90              // Move to beginning of next item.
91              i = skipwhitespace(line, j) ;
92  
93              if ( i < 0 )
94                  break ;
95  
96              int iStart = -2 ;       // Points to the beginning of the item as found
97              int jStart = -2 ;       // Points to the item without quotes
98              int iFinish = -2 ;      // Points after the end of the item as found
99              int jFinish = -2 ;      // Points after the end of the item without quotes
100             int dtStart = -2 ;      // Points to start of datatype (after < quote)
101             int dtFinish = -2 ;     // Points to end of datatype
102             int type = TupleItem.UNKNOWN;
103 
104             switch (line.charAt(i))
105             {
106                 case COMMENTCHAR:
107                     break tupleLoop ;
108                 case '<':
109                     type = TupleItem.URI ;
110                     iStart = i ;
111                     jStart = i+1 ;
112                     int newPosn = parseURI(i, line) ;
113                     if (newPosn < 0)
114                     {
115                         errorFound = true;
116                         break tupleLoop;
117                     }
118                     j = newPosn ;
119                     
120                     iFinish = j+1 ;
121                     jFinish = j ;
122                     break ;
123                 case '"':
124                     type = TupleItem.STRING ;
125                     iStart = i ;
126                     jStart = i+1 ;
127                     boolean inEscape = false ;
128                     for ( j = i+1 ; j < line.length() ; j++ )
129                     {
130                         char ch = line.charAt(j) ;
131                         if ( inEscape )
132                         {
133                             // ToDo: escape
134                             inEscape = false ;
135                             continue ;
136                         }
137                         // Not an escape
138                         if ( ch == '"'  )
139                             break ;
140 
141                         if ( ch == '\\' )
142                             inEscape = true ;
143                         if ( ch == '\n' || ch == '\r' )
144                         {
145                             errorFound = true ;
146                             break tupleLoop;
147                             
148                         }
149                     }
150                     
151                     // Malformed
152                     if ( j == line.length() )
153                     {
154                         errorFound = true ;
155                         break tupleLoop;
156                     }
157                     
158                     iFinish = j+1 ;
159                     jFinish = j ;
160                     // RDF literals may be followed by their type.
161                          
162                     if ( j < line.length()-3 
163                          && line.charAt(j+1) == '^'
164                          && line.charAt(j+2) == '^'
165                          && line.charAt(j+3) == '<' )
166                     {
167                         dtFinish = parseURI(j+3, line) ;
168                         dtStart = j+4 ;
169                         if (dtFinish < 0)
170                         {
171                             errorFound = true;
172                             break tupleLoop;
173                         }
174                         j = dtFinish+1 ;
175                         //String dt = line.substring(dtStart, dtFinish) ;
176                         //System.out.println("I see a datatype:"+dt) ;
177                     }
178                     
179                     break ;
180                 case '_':
181                     type = TupleItem.ANON ;
182                     iStart = i ;
183                     for ( j = i+1 ; j < line.length() ; j++ )
184                     {
185                         char ch = line.charAt(j) ;
186                         if ( ch == ' '  || ch == '\t' || ch == '.' )
187                             break ;
188                         if ( ! Character.isLetterOrDigit(ch) && ! (ch == '_') && ! (ch == ':') )
189                         {
190                             errorFound = true ;
191                             break tupleLoop ;
192                         }
193                     }
194                     iFinish = j ;
195                     jStart = iStart ;
196                     jFinish = iFinish ;
197                     break ;
198                 case '.':
199                 case '\n':
200                 case '\r':
201                     return tuple ;
202                 default:
203                     type = TupleItem.UNQUOTED ;
204                     iStart = i ;
205                     jStart = i ;
206                     for ( j = i+1 ; j < line.length() ; j++ )
207                     {
208                         char ch = line.charAt(j) ;
209                         if ( ch == ' '  || ch == '\t' || ch == '.' )
210                             break ;
211 
212                         //if ( ! Character.isLetterOrDigit(line.charAt(i)) )
213                         //{
214                         //    errorFound = true ;
215                         //    break tupleLoop;
216                         //}
217                     }
218                     // Malformed
219                     if ( j == line.length()+1 )
220                     {
221                         errorFound = true ;
222                         break tupleLoop;
223                     }
224                     iFinish = j ;
225                     jFinish = j ;
226                     break ;
227             }
228             String   item = line.substring(jStart, jFinish) ;
229             String   literal = line.substring(iStart, iFinish) ;
230             String   dt = null ;
231             if ( dtStart > 0 )
232                 dt = line.substring(dtStart, dtFinish) ;
233             
234             tuple.add(new TupleItem(item, literal, type, dt)) ;
235             j++ ;
236             // End of item.
237         }
238         //End of this line.
239         if ( errorFound )
240         {
241             logger.error( "Error in TupleSet.tuple: " + line );
242             
243             String   s = "" ;
244             int k = 0 ;
245             for ( ; k < i ; k++ ) s = s+" " ;
246             s = s+"^" ;
247             for ( ; k < j-1 ; k++ ) s=s+" " ;
248             s = s+"^" ;
249             logger.error( s ) ;
250             return null ;
251         }
252 
253         if ( tuple.size() == 0 )
254         {
255             // Nothing found : loop by tail recursion
256             return tuple() ;
257         }
258         return tuple ;
259     }
260 
261     private int skipwhitespace(String   s, int i)
262     {
263         for ( ; i < s.length() ; i++ )
264         {
265             char ch = s.charAt(i) ;
266             // Horizonal whitespace
267             if ( ch != ' ' && ch != '\t' )
268                 return i ;
269         }
270         return -1 ;
271     }
272 
273     private int parseURI(int i, String   line)
274     {
275         int j;
276         for (j = i + 1; j < line.length(); j++)
277         {
278             char ch = line.charAt(j);
279             if (ch == '>')
280                 break;
281             if (ch == '\n' || ch == '\r')
282                 return -1;
283         }
284         // Malformed
285         if (j == line.length())
286             return -2;
287         return j ;
288     }
289 }
290 
291 /*
292  *  (c) Copyright 2001 Hewlett-Packard Development Company, LP
293  *  All rights reserved.
294  *
295  * Redistribution and use in source and binary forms, with or without
296  * modification, are permitted provided that the following conditions
297  * are met:
298  * 1. Redistributions of source code must retain the above copyright
299  *    notice, this list of conditions and the following disclaimer.
300  * 2. Redistributions in binary form must reproduce the above copyright
301  *    notice, this list of conditions and the following disclaimer in the
302  *    documentation and/or other materials provided with the distribution.
303  * 3. The name of the author may not be used to endorse or promote products
304  *    derived from this software without specific prior written permission.
305  *
306  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
307  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
308  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
309  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
310  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
311  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
312  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
313  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
314  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
315  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
316  */
317
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags