2 * Sherlock: Data structures used in indices
4 * (c) 2001--2002 Martin Mares <mj@ucw.cz>
7 #ifndef _SHERLOCK_INDEX_H
8 #define _SHERLOCK_INDEX_H
10 #include "lib/fastbuf.h"
11 #include "charset/unistream.h"
15 #define MAX_WORD_LEN 64
16 #define MAX_COMPLEX_LEN 10
18 /* Word and string types are defined in lib/custom.h */
20 /* Global index parameters */
23 sh_time_t ref_time; /* Reference time (for document ages etc.) */
26 /* Index card attributes */
29 u32 card; /* Reference to card description (either oid or filepos) */
31 #define INT_ATTR(t,i,o,k,g,p) t i;
32 CUSTOM_ATTRS /* Include all custom attributes */
36 byte age; /* Document age in pseudo-logarithmic units wrt. reference time */
37 // byte rfu[1]; /* If no custom attributes are defined */
41 CARD_FLAG_EMPTY = 1, /* Empty document (redirect, robot file etc.) [scanner] */
42 CARD_FLAG_ACCENTED = 2, /* Document contains accented characters [scanner] */
43 CARD_FLAG_DUP = 4, /* Removed as a duplicate [merger] */
44 CARD_FLAG_MERGED = 8, /* Destination of a merge [merger] */
45 CARD_FLAG_IMAGE = 16, /* Is an image object [scanner] */
48 #define CARD_POS_SHIFT 5 /* Card positions are shifted this # of bytes to the right */
50 /* String fingerprints */
56 void fingerprint(byte *string, struct fingerprint *fp);
59 fp_hash(struct fingerprint *fp)
61 return (fp->hash[0] << 24) | (fp->hash[1] << 16) | (fp->hash[2] << 8) | fp->hash[3];
64 /* Reading of tagged text (Unicode values, tags mapped to 0x80000000 and higher) */
66 #define GET_TAGGED_CHAR(p,u) do { \
80 ASSERT(*p >= 0x80 && *p <= 0xbf); \
81 u = 0x80010000 + ((u & 0x0f) << 6) + (*p++ & 0x3f); \
91 bget_tagged_char(struct fastbuf *f)
106 ASSERT(v >= 0x80 && v <= 0xbf);
107 u = 0x80010000 + ((u & 0x0f) << 6) + (v & 0x3f);
120 /* Conversion of document age from seconds to our internal units */
123 convert_age(sh_time_t lastmod, sh_time_t reftime)
126 if (reftime < lastmod) /* past times */
128 age = (reftime - lastmod) / 3600;
129 if (age < 48) /* last 2 days: 1 hour resolution */
132 if (age < 64) /* next 64 days: 1 day resolution */
135 if (age < 135) /* next 135 weeks: 1 week resolution */
137 age = (age-135) / 52;
138 if (age < 8) /* next 8 years: 1 year resolution */
140 return 255; /* then just "infinite future" */