1 /* $Id: rset.h,v 1.63 2007-01-17 15:35:47 adam Exp $
2 Copyright (C) 1995-2007
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 this program; if not, write to the Free Software
19 Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
26 #include <yaz/yaz-util.h>
27 /* unfortunately we need the isam includes here, for the arguments for */
29 #include <idzebra/isamb.h>
30 #include <idzebra/isamc.h>
31 #include <idzebra/isams.h>
35 typedef struct rsfd *RSFD;
36 typedef struct rset *RSET;
40 struct ord_list *next;
43 struct ord_list *ord_list_create(NMEM nmem);
44 struct ord_list *ord_list_append(NMEM nmem, struct ord_list *list, int ord);
45 struct ord_list *ord_list_dup(NMEM nmem, struct ord_list *list);
48 * rset_term is all we need to know of a term to do ranking etc.
49 * As far as the rsets are concerned, it is just a dummy pointer to
53 char *name; /** the term itself in internal encoding (UTF-8/raw) */
54 char *flags; /** flags for rank method */
55 int type; /** Term_type from RPN Query. Actually this
56 is Z_Term_general, Z_Term_numeric,
57 Z_Term_characterString, ..
58 This info is used to return encoded term back for
61 int reg_type; /** register type */
62 RSET rset; /** the rset corresponding to this term */
63 void *rankpriv;/** private stuff for the ranking algorithm */
64 zint hits_limit;/** limit for hits if > 0 */
65 char *ref_id; /** reference for this term */
69 typedef struct rset_term *TERMID;
70 TERMID rset_term_create (const char *name, int length, const char *flags,
71 int type, NMEM nmem, struct ord_list *ol,
72 int reg_type, zint hits_limit, const char *ref_id);
74 /** rsfd is a "file descriptor" for reading from a rset */
75 struct rsfd { /* the stuff common to all rsfd's. */
76 RSET rset; /* ptr to the rset this FD is opened to */
77 void *priv; /* private parameters for this type */
78 RSFD next; /* to keep lists of used/free rsfd's */
85 * rset_control has function pointers to all the important functions
86 * of a rset. Each type of rset will have its own control block, pointing
87 * to the functions for that type. They all have their own create function
88 * which is not part of the control block, as it takes different args for
93 /** text description of set type (for debugging) */
95 /* RSET rs_something_create(const struct rset_control *sel, ...); */
96 void (*f_delete)(RSET ct);
98 /** recursively fills the terms array with terms. call with curterm=0 */
99 /* always counts them all into cur, but of course won't touch the term */
100 /* array past max. You can use this to count, set max=0 */
101 void (*f_getterms)(RSET ct, TERMID *terms, int maxterms, int *curterm);
103 RSFD (*f_open)(RSET ct, int wflag);
104 void (*f_close)(RSFD rfd);
105 /** forward behaves like a read, but it skips some items first */
106 int (*f_forward)(RSFD rfd, void *buf, TERMID *term, const void *untilbuf);
107 void (*f_pos)(RSFD rfd, double *current, double *total);
108 /* returns -1,-1 if pos function not implemented for this type */
109 int (*f_read)(RSFD rfd, void *buf, TERMID *term);
110 int (*f_write)(RSFD rfd, const void *buf);
113 /** rset_default_forward implements a generic forward with a read-loop */
114 int rset_default_forward(RSFD rfd, void *buf, TERMID *term,
115 const void *untilbuf);
117 /** rset_default_read implements a generic read */
118 int rset_default_read(RSFD rfd, void *buf, TERMID *term);
120 void rset_get_one_term(RSET ct,TERMID *terms,int maxterms,int *curterm);
123 * key_control contains all there is to know about the keys stored in
124 * an isam, and therefore operated by the rsets. Other than this info,
125 * all we assume is that all keys are the same size, and they can be
128 struct rset_key_control {
131 int scope; /* default for what level we operate (book/chapter/verse) on*/
132 /* usual sysno/seqno is 2 */
133 int (*cmp)(const void *p1, const void *p2);
134 void (*key_logdump_txt) (int logmask, const void *p, const char *txt);
135 zint (*getseq)(const void *p);
136 zint (*get_segment)(const void *p);
137 int (*filter_func)(const void *p, void *data);
139 void (*inc)(struct rset_key_control *kc);
140 void (*dec)(struct rset_key_control *kc);
141 /* FIXME - Should not need a getseq, it won't make much sense with */
142 /* higher-order keys. Use a (generalized) cmp instead, or something */
143 /* FIXME - decode and encode, and lots of other stuff */
147 * A rset is an ordered sequence of keys, either directly from an underlaying
148 * isam, or from one of the higher-level operator rsets (and, or, ...).
149 * Actually, it is "virtual base class", no pure rsets exist in the system,
150 * they all are of some derived type.
154 const struct rset_control *control;
155 struct rset_key_control *keycontrol;
156 int refcount; /* reference count */
157 void *priv; /* stuff private to the given type of rset */
158 NMEM nmem; /* nibble memory for various allocs */
159 RSFD free_list; /* all rfd's allocated but not currently in use */
160 RSFD use_list; /* all rfd's in use */
161 int scope; /* On what level do we count hits and compare them? */
162 TERMID term; /* the term thing for ranking etc */
170 /* rset is a "virtual base class", which will never exist on its own
171 * all instances are rsets of some specific type, like rsisamb, or rsbool
172 * They keep their own stuff behind the priv pointer. */
174 /* On the old sysno-seqno type isams, the scope was hard-coded to be 2.
175 * This means that we count hits on the sysno level, and when matching an
176 * 'and', we consider it a match if both term occur within the same sysno.
177 * In more complex isams we can specify on what level we wish to do the
178 * matching and counting of hits. For example, we can have book / chapter /
179 * verse, and a seqno. Scope 2 means then "give me all verses that match",
180 * 3 would be chapters, 4 books.
181 * The resolution tells how much of the occurences we need to return. If we
182 * are doing some sort of proximity, we need to get the seqnos of all
183 * occurences, whereas if we are only counting hits, we do not need anything
184 * below the scope. Again 1 is seqnos, 2 sysnos (or verses), 3 books, etc.
187 RSFD rfd_create_base(RSET rs);
188 int rfd_is_last(RSFD rfd);
190 RSET rset_create_base(const struct rset_control *sel,
192 struct rset_key_control *kcontrol,
195 int no_children, RSET *children);
197 void rset_delete(RSET rs);
198 RSET rset_dup (RSET rs);
199 void rset_close(RSFD rfd);
202 #define RSETF_WRITE 1
203 /* RSFD rset_open(RSET rs, int wflag); */
204 #define rset_open(rs, wflag) (*(rs)->control->f_open)((rs), (wflag))
206 /* int rset_forward(RSFD rfd, void *buf, TERMID term, void *untilbuf); */
207 #define rset_forward(rfd, buf, term, untilbuf) \
208 rset_default_forward((rfd), (buf), (term), (untilbuf))
210 /* void rset_getterms(RSET ct, TERMID *terms, int maxterms, int *curterm); */
211 #define rset_getterms(ct, terms, maxterms, curterm) \
212 (*(ct)->control->f_getterms)((ct),(terms),(maxterms),(curterm))
214 /* int rset_pos(RSFD fd, double *current, double *total); */
215 #define rset_pos(rfd,cur,tot) \
216 (*(rfd)->rset->control->f_pos)((rfd),(cur),(tot))
218 /* int rset_read(RSFD rfd, void *buf, TERMID term); */
219 #define rset_read(rfd, buf, term) rset_default_read((rfd), (buf), (term))
221 /* int rset_write(RSFD rfd, const void *buf); */
222 #define rset_write(rfd, buf) (*(rfd)->rset->control->f_write)((rfd), (buf))
224 /* int rset_type (RSET) */
225 #define rset_type(rs) ((rs)->control->desc)
227 /** rset_count counts or estimates the keys in it*/
228 zint rset_count(RSET rs);
230 RSET rset_create_temp(NMEM nmem, struct rset_key_control *kcontrol,
231 int scope, const char *temp_path, TERMID term);
233 RSET rset_create_null(NMEM nmem, struct rset_key_control *kcontrol, TERMID term);
235 RSET rset_create_not(NMEM nmem, struct rset_key_control *kcontrol,
236 int scope, RSET rset_l, RSET rset_r);
238 RSET rset_create_between(NMEM nmem, struct rset_key_control *kcontrol,
239 int scope, RSET rset_l, RSET rset_m, RSET rset_r,
242 RSET rset_create_or(NMEM nmem, struct rset_key_control *kcontrol,
243 int scope, TERMID termid, int no_rsets, RSET* rsets);
245 RSET rset_create_and(NMEM nmem, struct rset_key_control *kcontrol,
246 int scope, int no_rsets, RSET* rsets);
248 RSET rset_create_prox(NMEM nmem, struct rset_key_control *kcontrol,
249 int scope, int rset_no, RSET *rset,
250 int ordered, int exclusion, int relation, int distance);
252 RSET rsisamb_create(NMEM nmem, struct rset_key_control *kcontrol,
253 int scope, ISAMB is, ISAM_P pos, TERMID term);
255 RSET rsisamc_create(NMEM nmem, struct rset_key_control *kcontrol,
256 int scope, ISAMC is, ISAM_P pos, TERMID term);
258 RSET rsisams_create(NMEM nmem, struct rset_key_control *kcontrol,
259 int scope, ISAMS is, ISAM_P pos, TERMID term);
261 void rset_visit(RSET rset, int level);
263 void rset_set_hits_limit(RSET rs, zint l);
271 * indent-tabs-mode: nil
273 * vim: shiftwidth=4 tabstop=8 expandtab