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;
45 // mode flag: Default includes both the index and "naive" text
46 enum index_mode_t { index_mode_default, index_mode_text_only }; // index_mode_index_only
51 * New samplerate can be given, otherwise will use the one specified in the save file!
53 * Throws an exception if std::fread() fails.
56 static TextCollection* Load(FILE *, index_mode_t = index_mode_default, unsigned samplerate = 0);
59 * Save data structure into a file
61 * Throws an exception if std::fwrite() fails.
63 virtual void Save(FILE *) const = 0;
68 virtual ~TextCollection() { };
71 * Tests if the string pointed to by DocId is empty
73 virtual bool EmptyText(DocId) const = 0;
78 * Returns the i'th text in the collection.
79 * The numbering starts from 0.
81 * Call DeleteText() for each pointer returned by GetText()
82 * to avoid possible memory leaks.
84 virtual uchar* GetText(DocId) const = 0;
85 virtual void DeleteText(uchar *text) const = 0;
88 * Returns a pointer to the beginning of texts i, i+1, ..., j.
89 * Texts are separated by a '\0' byte.
91 * Call DeleteText() for each pointer returned by GetText()
92 * to avoid possible memory leaks.
94 virtual uchar * GetText(DocId i, DocId j) const = 0;
97 * Returns substring [i, j] of k'th text
99 * Note: Parameters i and j are text positions inside the k'th text.
101 // virtual uchar* GetText(DocId, TextPosition, TextPosition) const = 0;
103 * Returns backwards (reverse) iterator to the end of i'th text
105 * Note: Do we need this?
106 * Forward iterator would be really in-efficient compared to
109 * TODO Define and implement const_reverse_iterator.
111 //const_reverse_iterator rend(DocId) const;
114 * Existential queries
116 // Is there a text prefixed by given string?
117 virtual bool IsPrefix(uchar const *) const = 0;
118 // Is there a text having given string as a suffix?
119 virtual bool IsSuffix(uchar const *) const = 0;
120 // Is there a text that equals given string?
121 virtual bool IsEqual(uchar const *) const = 0;
122 // Does a text contain given string?
123 virtual bool IsContains(uchar const *) const = 0;
124 // Is there a text that is lexicographically less than given string?
125 virtual bool IsLessThan(uchar const *) const = 0;
128 * Existential queries for given DocId interval.
130 virtual bool IsPrefix(uchar const *, DocId, DocId) const = 0;
131 virtual bool IsSuffix(uchar const *, DocId, DocId) const = 0;
132 virtual bool IsEqual(uchar const *, DocId, DocId) const = 0;
133 virtual bool IsContains(uchar const *, DocId, DocId) const = 0;
134 virtual bool IsLessThan(uchar const *, DocId, DocId) const = 0;
138 * Result is the number of occurrences.
140 virtual ulong Count(uchar const *) const = 0;
142 * More counting queries
143 * Result is the number of documents.
145 virtual unsigned CountPrefix(uchar const *) const = 0;
146 virtual unsigned CountSuffix(uchar const *) const = 0;
147 virtual unsigned CountEqual(uchar const *) const = 0;
148 virtual unsigned CountContains(uchar const *) const = 0;
149 virtual unsigned CountLessThan(uchar const *) const = 0;
152 * Counting queries for given DocId interval
154 virtual unsigned CountPrefix(uchar const *, DocId, DocId) const = 0;
155 virtual unsigned CountSuffix(uchar const *, DocId, DocId) const = 0;
156 virtual unsigned CountEqual(uchar const *, DocId, DocId) const = 0;
157 virtual unsigned CountContains(uchar const *, DocId, DocId) const = 0;
158 virtual unsigned CountLessThan(uchar const *, DocId, DocId) const = 0;
161 * Document reporting queries
163 * Result is a vector of document id's in some undefined order.
165 // Data type for results
166 typedef std::vector<DocId> document_result;
167 virtual document_result Prefix(uchar const *) const = 0;
168 virtual document_result Suffix(uchar const *) const = 0;
169 virtual document_result Equal(uchar const *) const = 0;
170 virtual document_result Contains(uchar const *) const = 0;
171 virtual document_result LessThan(uchar const *) const = 0;
172 virtual document_result KMismaches(uchar const *, unsigned) const = 0;
173 virtual document_result KErrors(uchar const *, unsigned) const = 0;
176 * Document reporting queries for given DocId interval.
178 virtual document_result Prefix(uchar const *, DocId, DocId) const = 0;
179 virtual document_result Suffix(uchar const *, DocId, DocId) const = 0;
180 virtual document_result Equal(uchar const *, DocId, DocId) const = 0;
181 virtual document_result Contains(uchar const *, DocId, DocId) const = 0;
182 virtual document_result LessThan(uchar const *, DocId, DocId) const = 0;
185 * Full reporting queries
187 * Result is a vector of pairs <doc id, offset> in some undefined order.
189 // Data type for results
190 typedef std::vector<std::pair<DocId, TextPosition> > full_result;
191 virtual full_result FullContains(uchar const *) const = 0;
192 // Full reporting query for given DocId interval
193 virtual full_result FullContains(uchar const *, DocId, DocId) const = 0;
195 virtual full_result FullKMismatches(uchar const *, unsigned) const = 0;
196 virtual full_result FullKErrors(uchar const *, unsigned) const = 0;
199 // Protected constructor; use TextCollectionBuilder
200 TextCollection() { };
202 // index_mode_t indexMode;
204 // No copy constructor or assignment
205 TextCollection(TextCollection const&);
206 TextCollection& operator = (TextCollection const&);