2 * Sherlock: Data structures used in indices
4 * (c) 2001--2004 Martin Mares <mj@ucw.cz>
7 #ifndef _SHERLOCK_INDEX_H
8 #define _SHERLOCK_INDEX_H
10 #include "custom/lib/custom.h"
13 * Magic number which should help to avoid mixing incompatible indices.
14 * Syntax: <version><revision><custom-type><custom-version>
15 * Remember to increase with each change of index format.
18 #define INDEX_VERSION (0x34010000|((CUSTOM_INDEX_TYPE)<< 8)|(CUSTOM_INDEX_VERSION))
23 * MAX_WORD_LEN is the maximum length (measured in UTF-8 characters, excluding
24 * the terminating zero byte if there's any) of any word which may appear in the
25 * indices or in the bucket file. Naturally, the same constant also bounds
26 * the number of UCS-2 characters in a word.
28 * Caveat: If you are upcasing/downcasing the word, the UTF-8 encoding can
29 * expand, although at most twice, so you need to reserve 2*MAX_WORD_LEN bytes.
32 #define MAX_WORD_LEN 64 /* a multiple of 4 */
34 /* Word and string types are defined in custom/lib/custom.h */
36 /* Types used for storing contexts */
38 #ifdef CONFIG_CONTEXTS
39 #if CONFIG_MAX_CONTEXTS == 32768
40 typedef u16 context_t;
41 #define bget_context bgetw
42 #define bput_context bputw
43 #define GET_CONTEXT GET_U16
44 #define PUT_CONTEXT PUT_U16
45 #elif CONFIG_MAX_CONTEXTS == 256
46 typedef byte context_t;
47 #define bget_context bgetc
48 #define bput_context bputc
49 #define GET_CONTEXT GET_U8
50 #define PUT_CONTEXT PUT_U8
52 #error CONFIG_MAX_CONTEXTS set to an invalid value.
56 typedef struct { } context_t;
57 static inline uns bget_context(struct fastbuf *b UNUSED) { return 0; }
58 static inline void bput_context(struct fastbuf *b UNUSED, uns context UNUSED) { }
59 #define GET_CONTEXT(p) 0
60 #define PUT_CONTEXT(p,x) do {} while(0)
63 /* Index card attributes */
66 u32 card; /* Reference to card description (either oid or filepos) */
71 CUSTOM_CARD_ATTRS /* Include all custom attributes */
75 byte age; /* Document age in pseudo-logarithmic units wrt. reference time */
77 #ifdef CONFIG_FILETYPE
78 byte type_flags; /* File type flags (see below) */
83 CARD_FLAG_EMPTY = 1, /* Empty document (redirect, robot file etc.) [scanner] */
84 CARD_FLAG_ACCENTED = 2, /* Document contains accented characters [scanner] */
85 CARD_FLAG_DUP = 4, /* Removed as a duplicate [merger] */
86 CARD_FLAG_MERGED = 8, /* Destination of a merge [merger] */
87 CARD_FLAG_IMAGE = 16, /* Is an image object [scanner] */
88 CARD_FLAG_FRAMESET = 32, /* Contains a frameset to be ignored [scanner] */
89 CARD_FLAG_OVERRIDEN = 64, /* Overriden by another index [sherlockd] */
92 #define CARD_POS_SHIFT 5 /* Card positions are shifted this # of bits to the right */
95 * We store document type and several other properties in card_attr->type_flags.
96 * Here we define only the basic structure, the details are defined in custom.h
97 * (the list of type names custom_file_type_names[] and also setting of the file
98 * types in custom_create_attrs()).
100 * bits 7--5 file type: (0-3: text types, 4-7: other types, defined by custom.h)
101 * bits 4--0 type-dependent information, for text types it's document language code
104 #ifdef CONFIG_FILETYPE
105 #define CA_GET_FILE_TYPE(a) ((a)->type_flags >> 4)
106 #define CA_GET_FILE_INFO(a) ((a)->type_flags & 0x0f)
107 #define CA_GET_FILE_LANG(a) ((a)->type_flags & 0x80 ? 0 : CA_GET_FILE_INFO(a))
108 #define MAX_FILE_TYPES 16
109 #define FILETYPE_IS_TEXT(f) ((f) < 8)
110 byte *ext_ft_parse(u32 *dest, byte *value, uns intval);
111 extern byte *custom_file_type_names[MAX_FILE_TYPES];
112 #define FILETYPE_STAT_VARS uns matching_per_type[MAX_FILE_TYPES];
113 #define FILETYPE_SHOW_STATS(q,f) ext_ft_show(q,f)
114 #define FILETYPE_INIT_STATS(q) bzero(q->matching_per_type, sizeof(q->matching_per_type))
115 #ifdef CONFIG_COUNT_ALL_FILETYPES
116 #define FILETYPE_ATTRS LATE_SMALL_SET_ATTR(ftype, FILETYPE, CA_GET_FILE_TYPE, ext_ft_parse)
117 #define FILETYPE_EARLY_STATS(q,a) q->matching_per_type[CA_GET_FILE_TYPE(a)]++
118 #define FILETYPE_LATE_STATS(q,a)
120 #define FILETYPE_ATTRS SMALL_SET_ATTR(ftype, FILETYPE, CA_GET_FILE_TYPE, ext_ft_parse)
121 #define FILETYPE_EARLY_STATS(q,a)
122 #define FILETYPE_LATE_STATS(q,a) q->matching_per_type[CA_GET_FILE_TYPE(a)]++
125 #define FILETYPE_ATTRS
126 #define FILETYPE_STAT_VARS
127 #define FILETYPE_INIT_STATS(q)
128 #define FILETYPE_EARLY_STATS(q,a)
129 #define FILETYPE_LATE_STATS(q,a)
130 #define FILETYPE_SHOW_STATS(q,f)
134 /* You can use language matching without CONFIG_FILETYPE, but you have to define CA_GET_FILE_LANG yourself. */
135 #define LANG_ATTRS SMALL_SET_ATTR(lang, LANG, CA_GET_FILE_LANG, ext_lang_parse)
136 byte *ext_lang_parse(u32 *dest, byte *value, uns intval);
142 #define CA_GET_AREA(a) ((a)->area)
143 #define SPLIT_ATTRS INT_ATTR(area, AREA, CA_GET_AREA, ext_area_parse)
144 byte *ext_area_parse(u32 *dest, byte *value, uns intval);
150 * A list of all extended attributes: custom attributes and also some
151 * built-in attributes treated in the same way.
154 #define EXTENDED_ATTRS CUSTOM_ATTRS FILETYPE_ATTRS LANG_ATTRS SPLIT_ATTRS
157 * A list of all statistics collectors, also composed of custom parts
158 * and built-in parts.
161 #ifndef CUSTOM_STAT_VARS
162 #define CUSTOM_STAT_VARS
163 #define CUSTOM_INIT_STATS(q)
164 #define CUSTOM_EARLY_STATS(q,a)
165 #define CUSTOM_LATE_STATS(q,a)
166 #define CUSTOM_SHOW_STATS(q,f)
169 #define EXTENDED_STAT_VARS CUSTOM_STAT_VARS FILETYPE_STAT_VARS
170 #define EXTENDED_INIT_STATS(q) CUSTOM_INIT_STATS(q) FILETYPE_INIT_STATS(q)
171 #define EXTENDED_EARLY_STATS(q,a) CUSTOM_EARLY_STATS(q,a) FILETYPE_EARLY_STATS(q,a)
172 #define EXTENDED_LATE_STATS(q,a) CUSTOM_LATE_STATS(q,a) FILETYPE_LATE_STATS(q,a)
173 #define EXTENDED_SHOW_STATS(q,f) CUSTOM_SHOW_STATS(q,f) FILETYPE_SHOW_STATS(q,f)
175 /* String fingerprints */
181 void fingerprint(byte *string, struct fingerprint *fp);
184 fp_hash(struct fingerprint *fp)
186 return (fp->hash[0] << 24) | (fp->hash[1] << 16) | (fp->hash[2] << 8) | fp->hash[3];
189 /* The card fingerprints */
192 struct fingerprint fp;
198 #define URL_KEY_BUF_SIZE (3*MAX_URL_SIZE)
199 byte *url_key(byte *url, byte *buf);
200 void url_fingerprint(byte *url, struct fingerprint *fp);
201 void url_key_init(void);
203 /* Conversion of document age from seconds to our internal units */
206 convert_age(sh_time_t lastmod, sh_time_t reftime)
209 if (reftime < lastmod) /* past times */
211 age = (reftime - lastmod) / 3600;
212 if (age < 48) /* last 2 days: 1 hour resolution */
215 if (age < 64) /* next 64 days: 1 day resolution */
218 if (age < 135) /* next 135 weeks: 1 week resolution */
220 age = (age-135) / 52;
221 if (age < 8) /* next 8 years: 1 year resolution */
223 return 255; /* then just "infinite future" */