1 /* $Id: rset.h,v 1.47 2005-01-17 01:21:43 adam Exp $
2 Copyright (C) 1995-2005
5 This file is part of the Zebra server.
7 Zebra is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with Zebra; see the file LICENSE.zebra. If not, write to the
19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
26 /* unfortunately we need the isam includes here, for the arguments for */
28 #include <idzebra/isamb.h>
29 #include <idzebra/isamc.h>
30 #include <idzebra/isams.h>
34 typedef struct rsfd *RSFD; /* Rset "file descriptor" */
35 typedef struct rset *RSET; /* Result set */
39 * rset_term is all we need to know of a term to do ranking etc.
40 * As far as the rsets are concerned, it is just a dummy pointer to
45 /** the term itself */
49 /** the rset corresponding to this term */
51 /** private stuff for the ranking algorithm */
55 typedef struct rset_term *TERMID;
56 TERMID rset_term_create (const char *name, int length, const char *flags,
61 /** rsfd is a "file descriptor" for reading from a rset */
62 struct rsfd { /* the stuff common to all rsfd's. */
63 RSET rset; /* ptr to the rset this FD is opened to */
64 void *priv; /* private parameters for this type */
65 RSFD next; /* to keep lists of used/free rsfd's */
70 * rset_control has function pointers to all the important functions
71 * of a rset. Each type of rset will have its own control block, pointing
72 * to the functions for that type. They all have their own create function
73 * which is not part of the control block, as it takes different args for
78 /** text description of set type (for debugging) */
80 /* RSET rs_something_create(const struct rset_control *sel, ...); */
81 void (*f_delete)(RSET ct);
83 /** recursively fills the terms array with terms. call with curterm=0 */
84 /* always counts them all into cur, but of course won't touch the term */
85 /* array past max. You can use this to count, set max=0 */
86 void (*f_getterms)(RSET ct, TERMID *terms, int maxterms, int *curterm);
88 RSFD (*f_open)(RSET ct, int wflag);
89 void (*f_close)(RSFD rfd);
90 /** forward behaves like a read, but it skips some items first */
91 int (*f_forward)(RSFD rfd, void *buf, TERMID *term, const void *untilbuf);
92 void (*f_pos)(RSFD rfd, double *current, double *total);
93 /* returns -1,-1 if pos function not implemented for this type */
94 int (*f_read)(RSFD rfd, void *buf, TERMID *term);
95 int (*f_write)(RSFD rfd, const void *buf);
98 /** rset_default_forward implements a generic forward with a read-loop */
99 int rset_default_forward(RSFD rfd, void *buf, TERMID *term,
100 const void *untilbuf);
102 /** rset_get_no_terms is a getterms function for those that don't have any */
103 void rset_get_no_terms(RSET ct, TERMID *terms, int maxterms, int *curterm);
106 * rset_get_one_term is a getterms function for those rsets that have
107 * exactly one term, like all rsisamX types.
109 void rset_get_one_term(RSET ct,TERMID *terms,int maxterms,int *curterm);
112 * key_control contains all there is to know about the keys stored in
113 * an isam, and therefore operated by the rsets. Other than this info,
114 * all we assume is that all keys are the same size, and they can be
119 int scope; /* default for what level we operate (book/chapter/verse) on*/
120 /* usual sysno/seqno is 2 */
121 int (*cmp) (const void *p1, const void *p2);
122 void (*key_logdump_txt) (int logmask, const void *p, const char *txt);
123 zint (*getseq)(const void *p);
124 /* FIXME - Should not need a getseq, it won't make much sense with */
125 /* higher-order keys. Use a (generalized) cmp instead, or something */
126 /* FIXME - decode and encode, and lots of other stuff */
130 * A rset is an ordered sequence of keys, either directly from an underlaying
131 * isam, or from one of the higher-level operator rsets (and, or, ...).
132 * Actually, it is "virtual base class", no pure rsets exist in the system,
133 * they all are of some derived type.
137 const struct rset_control *control;
138 const struct key_control *keycontrol;
139 int count; /* reference count */
140 void *priv; /* stuff private to the given type of rset */
141 NMEM nmem; /* nibble memory for various allocs */
142 char my_nmem; /* Should the nmem be destroyed with the rset? */
143 /* 1 if created with it, 0 if passed from above */
144 RSFD free_list; /* all rfd's allocated but not currently in use */
145 RSFD use_list; /* all rfd's in use */
146 int scope; /* On what level do we count hits and compare them? */
147 TERMID term; /* the term thing for ranking etc */
149 /* rset is a "virtual base class", which will never exist on its own
150 * all instances are rsets of some specific type, like rsisamb, or rsbool
151 * They keep their own stuff behind the priv pointer. */
153 /* On the old sysno-seqno type isams, the scope was hard-coded to be 2.
154 * This means that we count hits on the sysno level, and when matching an
155 * 'and', we consider it a match if both term occur within the same sysno.
156 * In more complex isams we can specify on what level we wish to do the
157 * matching and counting of hits. For example, we can have book / chapter /
158 * verse, and a seqno. Scope 2 means then "give me all verses that match",
159 * 3 would be chapters, 4 books.
160 * The resolution tells how much of the occurences we need to return. If we
161 * are doing some sort of proximity, we need to get the seqnos of all
162 * occurences, whereas if we are only counting hits, we do not need anything
163 * below the scope. Again 1 is seqnos, 2 sysnos (or verses), 3 books, etc.
166 RSFD rfd_create_base(RSET rs);
167 void rfd_delete_base(RSFD rfd);
168 int rfd_is_last(RSFD rfd);
170 RSET rset_create_base(const struct rset_control *sel,
172 const struct key_control *kcontrol,
176 void rset_delete(RSET rs);
177 RSET rset_dup (RSET rs);
181 #define RSETF_WRITE 1
182 /* RSFD rset_open(RSET rs, int wflag); */
183 #define rset_open(rs, wflag) (*(rs)->control->f_open)((rs), (wflag))
185 /* void rset_close(RSFD rfd); */
186 #define rset_close(rfd) (*(rfd)->rset->control->f_close)(rfd)
188 /* int rset_forward(RSFD rfd, void *buf, TERMID term, void *untilbuf); */
189 #define rset_forward(rfd, buf, term, untilbuf) \
190 (*(rfd)->rset->control->f_forward)((rfd),(buf),(term),(untilbuf))
192 /* void rset_getterms(RSET ct, TERMID *terms, int maxterms, int *curterm); */
193 #define rset_getterms(ct, terms, maxterms, curterm) \
194 (*(ct)->control->f_getterms)((ct),(terms),(maxterms),(curterm))
196 /* int rset_pos(RSFD fd, double *current, double *total); */
197 #define rset_pos(rfd,cur,tot) \
198 (*(rfd)->rset->control->f_pos)((rfd),(cur),(tot))
200 /* int rset_read(RSFD rfd, void *buf, TERMID term); */
201 #define rset_read(rfd, buf, term) \
202 (*(rfd)->rset->control->f_read)((rfd), (buf), (term))
204 /* int rset_write(RSFD rfd, const void *buf); */
205 #define rset_write(rfd, buf) (*(rfd)->rset->control->f_write)((rfd), (buf))
207 /* int rset_type (RSET) */
208 #define rset_type(rs) ((rs)->control->desc)
210 /** rset_count counts or estimates the keys in it*/
211 zint rset_count(RSET rs);
213 RSET rstemp_create(NMEM nmem, const struct key_control *kcontrol,
215 const char *temp_path, TERMID term);
217 RSET rsnull_create(NMEM nmem, const struct key_control *kcontrol);
219 RSET rsbool_create_and(NMEM nmem, const struct key_control *kcontrol,
220 int scope, RSET rset_l, RSET rset_r);
222 RSET rsbool_create_or (NMEM nmem, const struct key_control *kcontrol,
223 int scope, RSET rset_l, RSET rset_r);
225 RSET rsbool_create_not(NMEM nmem, const struct key_control *kcontrol,
226 int scope, RSET rset_l, RSET rset_r);
228 RSET rsbetween_create(NMEM nmem, const struct key_control *kcontrol,
229 int scope, RSET rset_l, RSET rset_m, RSET rset_r,
232 RSET rsmulti_or_create(NMEM nmem, const struct key_control *kcontrol,
233 int scope, int no_rsets, RSET* rsets);
235 RSET rsmulti_and_create(NMEM nmem, const struct key_control *kcontrol,
236 int scope, int no_rsets, RSET* rsets);
238 RSET rsprox_create(NMEM nmem, const struct key_control *kcontrol,
239 int scope, int rset_no, RSET *rset,
240 int ordered, int exclusion, int relation, int distance);
242 RSET rsisamb_create(NMEM nmem, const struct key_control *kcontrol,
243 int scope, ISAMB is, ISAMB_P pos, TERMID term);
245 RSET rsisamc_create(NMEM nmem, const struct key_control *kcontrol,
246 int scope, ISAMC is, ISAMC_P pos, TERMID term);
248 RSET rsisams_create(NMEM nmem, const struct key_control *kcontrol,
249 int scope, ISAMS is, ISAMS_P pos, TERMID term);