1 /******************************************************************************
2 * Copyright (C) 2008 by Niko Valimaki <nvalimak@cs.helsinki.fi> *
3 * Text collection interface for an in-memory XQuery/XPath engine *
5 * This program is free software; you can redistribute it and/or modify *
6 * it under the terms of the GNU Lesser General Public License as published *
7 * by the Free Software Foundation; either version 2 of the License, or *
8 * (at your option) any later version. *
10 * This program is distributed in the hope that it will be useful, *
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of *
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
13 * GNU Lesser General Public License for more details. *
15 * You should have received a copy of the GNU Lesser General Public License *
16 * along with this program; if not, write to the *
17 * Free Software Foundation, Inc., *
18 * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. *
19 ******************************************************************************/
21 #ifndef _SXSI_TextCollection_h_
22 #define _SXSI_TextCollection_h_
24 #include "Tools.h" // Defines ulong and uchar.
26 #include <utility> // Defines std::pair.
32 * General interface for a text collection
34 * Class is virtual, make objects by calling
35 * the static method InitTextCollection().
40 // Type of document identifier
42 // Type for text position (FIXME ulong or long?)
43 typedef ulong TextPosition;
48 * New samplerate can be given, otherwise will use the one specified in the save file!
49 * Note: This is not a static method; call InitTextCollection() first to get the object handle.
51 * Throws an exception if std::fread() fails.
54 static TextCollection* Load(FILE *, unsigned samplerate = 0);
57 * Save data structure into a file
59 * Throws an exception if std::fwrite() fails.
61 virtual void Save(FILE *) const = 0;
66 virtual ~TextCollection() { };
69 * Tests if the string pointed to by DocId is empty
71 virtual bool EmptyText(DocId) const = 0;
76 * Returns the i'th text in the collection.
77 * The numbering starts from 0.
79 * Call DeleteText() for each pointer returned by GetText()
80 * to avoid possible memory leaks.
82 virtual uchar* GetText(DocId) const = 0;
83 virtual void DeleteText(uchar *text) const = 0;
86 * Returns a pointer to the beginning of texts i, i+1, ..., j.
87 * Texts are separated by a '\0' byte.
89 * Call DeleteText() for each pointer returned by GetText()
90 * to avoid possible memory leaks.
92 virtual uchar * GetText(DocId i, DocId j) const = 0;
95 * Returns substring [i, j] of k'th text
97 * Note: Parameters i and j are text positions inside the k'th text.
99 // virtual uchar* GetText(DocId, TextPosition, TextPosition) const = 0;
101 * Returns backwards (reverse) iterator to the end of i'th text
103 * Note: Do we need this?
104 * Forward iterator would be really in-efficient compared to
107 * TODO Define and implement const_reverse_iterator.
109 //const_reverse_iterator rend(DocId) const;
112 * Existential queries
114 // Is there a text prefixed by given string?
115 virtual bool IsPrefix(uchar const *) const = 0;
116 // Is there a text having given string as a suffix?
117 virtual bool IsSuffix(uchar const *) const = 0;
118 // Is there a text that equals given string?
119 virtual bool IsEqual(uchar const *) const = 0;
120 // Does a text contain given string?
121 virtual bool IsContains(uchar const *) const = 0;
122 // Is there a text that is lexicographically less than given string?
123 virtual bool IsLessThan(uchar const *) const = 0;
126 * Existential queries for given DocId interval.
128 virtual bool IsPrefix(uchar const *, DocId, DocId) const = 0;
129 virtual bool IsSuffix(uchar const *, DocId, DocId) const = 0;
130 virtual bool IsEqual(uchar const *, DocId, DocId) const = 0;
131 virtual bool IsContains(uchar const *, DocId, DocId) const = 0;
132 virtual bool IsLessThan(uchar const *, DocId, DocId) const = 0;
136 * Result is the number of occurrences.
138 virtual ulong Count(uchar const *) const = 0;
140 * More counting queries
141 * Result is the number of documents.
143 virtual unsigned CountPrefix(uchar const *) const = 0;
144 virtual unsigned CountSuffix(uchar const *) const = 0;
145 virtual unsigned CountEqual(uchar const *) const = 0;
146 virtual unsigned CountContains(uchar const *) const = 0;
147 virtual unsigned CountLessThan(uchar const *) const = 0;
150 * Counting queries for given DocId interval
152 virtual unsigned CountPrefix(uchar const *, DocId, DocId) const = 0;
153 virtual unsigned CountSuffix(uchar const *, DocId, DocId) const = 0;
154 virtual unsigned CountEqual(uchar const *, DocId, DocId) const = 0;
155 virtual unsigned CountContains(uchar const *, DocId, DocId) const = 0;
156 virtual unsigned CountLessThan(uchar const *, DocId, DocId) const = 0;
159 * Document reporting queries
161 * Result is a vector of document id's in some undefined order.
163 // Data type for results
164 typedef std::vector<DocId> document_result;
165 virtual document_result Prefix(uchar const *) const = 0;
166 virtual document_result Suffix(uchar const *) const = 0;
167 virtual document_result Equal(uchar const *) const = 0;
168 virtual document_result Contains(uchar const *) const = 0;
169 virtual document_result LessThan(uchar const *) const = 0;
170 virtual document_result KMismaches(uchar const *, unsigned) const = 0;
171 virtual document_result KErrors(uchar const *, unsigned) const = 0;
174 * Document reporting queries for given DocId interval.
176 virtual document_result Prefix(uchar const *, DocId, DocId) const = 0;
177 virtual document_result Suffix(uchar const *, DocId, DocId) const = 0;
178 virtual document_result Equal(uchar const *, DocId, DocId) const = 0;
179 virtual document_result Contains(uchar const *, DocId, DocId) const = 0;
180 virtual document_result LessThan(uchar const *, DocId, DocId) const = 0;
183 * Full reporting queries
185 * Result is a vector of pairs <doc id, offset> in some undefined order.
187 // Data type for results
188 typedef std::vector<std::pair<DocId, TextPosition> > full_result;
189 virtual full_result FullContains(uchar const *) const = 0;
190 // Full reporting query for given DocId interval
191 virtual full_result FullContains(uchar const *, DocId, DocId) const = 0;
193 virtual full_result FullKMismatches(uchar const *, unsigned) const = 0;
194 virtual full_result FullKErrors(uchar const *, unsigned) const = 0;
197 // Protected constructor; use TextCollectionBuilder
198 TextCollection() { };
200 // No copy constructor or assignment
201 TextCollection(TextCollection const&);
202 TextCollection& operator = (TextCollection const&);