X-Git-Url: http://mj.ucw.cz/gitweb/?a=blobdiff_plain;f=lib%2Findex.h;h=0585141245e440dd4b03e0dc75068956ee00abea;hb=8ffdaed1c5f5bf4ad0914b8d17f14af51ce037a5;hp=0fe46f6bb6a9f45e4a676e0198e58a9894d11722;hpb=c7b8fa1cf1dae26c8a852f6e3e67e96da7209741;p=libucw.git diff --git a/lib/index.h b/lib/index.h index 0fe46f6b..05851412 100644 --- a/lib/index.h +++ b/lib/index.h @@ -1,83 +1,33 @@ /* - * Sherlock Gatherer: Data structures used in indices + * Sherlock: Data structures used in indices * - * (c) 2001 Martin Mares + * (c) 2001--2002 Martin Mares */ /* Words */ #define MAX_WORD_LEN 64 -/* Word types */ - -enum word_type { - WT_RESERVED, /* Reserved word type */ - WT_TEXT, /* Ordinary text */ - WT_EMPH, /* Emphasized text */ - WT_SMALL, /* Small font */ - WT_TITLE, /* Document title */ - WT_SMALL_HEADING, /* Heading */ - WT_BIG_HEADING, /* Larger heading */ - WT_KEYWORD, /* Explicitly marked keyword */ - WT_META, /* Various meta-information */ - WT_ALT, /* Alternate texts for graphical elements */ - WT_MAX -}; - -/* Descriptive names used for user output */ -#define WORD_TYPE_USER_NAMES \ - "reserved", "text", "emph", "small", "title", "hdr1", "hdr2", "keywd", \ - "meta", "alt", "type10", "type11", "type12", "type13", "type14", "type15" - -/* Keywords for word type names */ -#define WORD_TYPE_NAMES \ - T(WORD, ~0) \ - T(TEXT, 1 << WT_TEXT) \ - T(EMPH, 1 << WT_EMPH) \ - T(SMALL, 1 << WT_SMALL) \ - T(TITLE, 1 << WT_TITLE) \ - T(HDR, (1 << WT_SMALL_HEADING) | (1 << WT_BIG_HEADING)) \ - T(HDR1, 1 << WT_SMALL_HEADING) \ - T(HDR2, 1 << WT_BIG_HEADING) \ - T(KEYWD, 1 << WT_KEYWORD) \ - T(META, 1 << WT_META) \ - T(ALT, 1 << WT_ALT) - -/* String types */ - -enum string_type { - ST_RESERVED, /* Reserved string type */ - ST_URL, /* URL of the document */ - ST_HOST, /* Host name */ - ST_DOMAIN, /* Domain name */ - ST_REF, /* URL reference */ - ST_BACKREF, /* Back-reference (frame or redirect source) */ - ST_MAX -}; +/* Word and string types are defined in lib/custom.h */ -#define STRING_TYPE_USER_NAMES \ - "URL", "host", "domain", "ref", "backref", "type5", "type6", "type7", \ - "type8", "type9", "type10", "type11", "type12", "type13", "type14", "type15" +/* Global index parameters */ -#define STRING_TYPE_NAMES \ - T(URL, 1 << ST_URL) \ - T(HOST, 1 << ST_HOST) \ - T(DOMAIN, 1 << ST_DOMAIN) \ - T(REF, 1 << ST_REF) \ - T(BACKREF, 1 << ST_BACKREF) - -#define STRING_TYPES_URL ((1 << ST_URL) | (1 << ST_REF) | (1 << ST_BACKREF)) -/* These must be indexed in lowercase form */ -#define STRING_TYPES_CASE_INSENSITIVE ((1 << ST_HOST) | (1 << ST_DOMAIN)) +struct index_params { + sh_time_t ref_time; /* Reference time (for document ages etc.) */ +}; /* Index card attributes */ struct card_attr { u32 card; /* Reference to card description (either oid or filepos) */ u32 site_id; +#define INT_ATTR(t,i,o,k,g,p) t i; + CUSTOM_ATTRS /* Include all custom attributes */ +#undef INT_ATTR byte weight; byte flags; - byte rfu[2]; + byte age; /* Document age in pseudo-logarithmic units wrt. reference time */ + // byte rfu[1]; /* If no custom attributes are defined */ }; enum card_flag { @@ -108,7 +58,7 @@ fp_hash(struct fingerprint *fp) #define GET_TAGGED_CHAR(p,u) do { \ u = *p; \ if (u >= 0xc0) \ - GET_UTF8(p,u); \ + GET_UTF8_CHAR(p,u); \ else if (u >= 0x80) \ { \ p++; \ @@ -129,3 +79,26 @@ fp_hash(struct fingerprint *fp) else \ p++; \ } while (0) + +/* Conversion of document age from seconds to our internal units */ + +static inline int +convert_age(sh_time_t lastmod, sh_time_t reftime) +{ + sh_time_t age; + if (reftime < lastmod) /* past times */ + return -1; + age = (reftime - lastmod) / 3600; + if (age < 48) /* last 2 days: 1 hour resolution */ + return age; + age = (age-48) / 24; + if (age < 64) /* next 64 days: 1 day resolution */ + return 48 + age; + age = (age-64) / 7; + if (age < 135) /* next 135 weeks: 1 week resolution */ + return 112 + age; + age = (age-135) / 52; + if (age < 8) /* next 8 years: 1 year resolution */ + return 247 + age; + return 255; /* then just "infinite future" */ +}