X-Git-Url: http://mj.ucw.cz/gitweb/?a=blobdiff_plain;ds=sidebyside;f=lib%2Findex.h;h=3aed7b97ee7c7b780182ed529c3bd23c58224bb3;hb=5815f75d01bb508be9e8a9aacf954e2db7c983bb;hp=0fb8a147bba9b1b091e8311cb82f793f08bdb8fd;hpb=f91be42d158f7663d636169d8a09d986c1595a67;p=libucw.git diff --git a/lib/index.h b/lib/index.h index 0fb8a147..3aed7b97 100644 --- a/lib/index.h +++ b/lib/index.h @@ -1,18 +1,18 @@ /* * Sherlock: Data structures used in indices * - * (c) 2001--2002 Martin Mares + * (c) 2001--2004 Martin Mares */ #ifndef _SHERLOCK_INDEX_H #define _SHERLOCK_INDEX_H -#include "lib/fastbuf.h" -#include SHERLOCK_CUSTOM -#include "charset/unistream.h" +#include "custom/lib/custom.h" + +#define INDEX_VERSION (0x32260200+sizeof(struct card_attr)) /* Increase with each incompatible change in index format */ /* - * Words and word complexes + * Words * * MAX_WORD_LEN is the maximum length (measured in UTF-8 characters, excluding * the terminating zero byte if there's any) of any word which may appear in the @@ -21,20 +21,38 @@ * * Caveat: If you are upcasing/downcasing the word, the UTF-8 encoding can * expand, although at most twice, so you need to reserve 2*MAX_WORD_LEN bytes. - * - * MAX_COMPLEX_LEN is the upper bound on number of words in any word complex. */ #define MAX_WORD_LEN 64 /* a multiple of 4 */ -#define MAX_COMPLEX_LEN 10 /* Word and string types are defined in lib/custom.h */ -/* Global index parameters */ - -struct index_params { - sh_time_t ref_time; /* Reference time (for document ages etc.) */ -}; +/* Types used for storing contexts */ + +#ifdef CONFIG_CONTEXTS +#if CONFIG_MAX_CONTEXTS == 32768 +typedef u16 context_t; +#define bget_context bgetw +#define bput_context bputw +#define GET_CONTEXT GET_U16 +#define PUT_CONTEXT PUT_U16 +#elif CONFIG_MAX_CONTEXTS == 256 +typedef byte context_t; +#define bget_context bgetc +#define bput_context bputc +#define GET_CONTEXT GET_U8 +#define PUT_CONTEXT PUT_U8 +#else +#error CONFIG_MAX_CONTEXTS set to an invalid value. +#endif +#else +struct fastbuf; +typedef struct { } context_t; +static inline uns bget_context(struct fastbuf *b UNUSED) { return 0; } +static inline void bput_context(struct fastbuf *b UNUSED, uns context UNUSED) { } +#define GET_CONTEXT(p) 0 +#define PUT_CONTEXT(p,x) do {} while(0) +#endif /* Index card attributes */ @@ -60,22 +78,84 @@ enum card_flag { CARD_FLAG_DUP = 4, /* Removed as a duplicate [merger] */ CARD_FLAG_MERGED = 8, /* Destination of a merge [merger] */ CARD_FLAG_IMAGE = 16, /* Is an image object [scanner] */ + CARD_FLAG_FRAMESET = 32, /* Contains a frameset to be ignored [scanner] */ + CARD_FLAG_OVERRIDEN = 64, /* Overriden by another index [sherlockd] */ }; -#define CARD_POS_SHIFT 5 /* Card positions are shifted this # of bytes to the right */ +#define CARD_POS_SHIFT 5 /* Card positions are shifted this # of bits to the right */ /* * We store document type and several other properties in card_attr->type_flags. - * Here we define only the basic structure, the details and also how to match the - * types are defined in custom.h. + * Here we define only the basic structure, the details are defined in custom.h + * (the list of type names custom_file_type_names[] and also setting of the file + * types in custom_create_attrs()). * * bits 7--5 file type: (0-3: text types, 4-7: other types, defined by custom.h) * bits 4--0 type-dependent information, for text types it's document language code */ +#ifdef CONFIG_FILETYPE #define CA_GET_FILE_TYPE(a) ((a)->type_flags >> 5) #define CA_GET_FILE_INFO(a) ((a)->type_flags & 0x1f) #define CA_GET_FILE_LANG(a) ((a)->type_flags & 0x80 ? 0 : CA_GET_FILE_INFO(a)) +#define MAX_FILE_TYPES 8 +#define FILETYPE_IS_TEXT(f) ((f) < 4) +byte *ext_ft_parse(u32 *dest, byte *value, uns intval); +extern byte *custom_file_type_names[MAX_FILE_TYPES]; +#define FILETYPE_STAT_VARS uns matching_per_type[MAX_FILE_TYPES]; +#define FILETYPE_SHOW_STATS(q,f) ext_ft_show(q,f) +#define FILETYPE_INIT_STATS(q) bzero(q->matching_per_type, sizeof(q->matching_per_type)) +#ifdef CONFIG_COUNT_ALL_FILETYPES +#define FILETYPE_ATTRS LATE_SMALL_SET_ATTR(ftype, FILETYPE, CA_GET_FILE_TYPE, ext_ft_parse) +#define FILETYPE_EARLY_STATS(q,a) q->matching_per_type[CA_GET_FILE_TYPE(a)]++ +#define FILETYPE_LATE_STATS(q,a) +#else +#define FILETYPE_ATTRS SMALL_SET_ATTR(ftype, FILETYPE, CA_GET_FILE_TYPE, ext_ft_parse) +#define FILETYPE_EARLY_STATS(q,a) +#define FILETYPE_LATE_STATS(q,a) q->matching_per_type[CA_GET_FILE_TYPE(a)]++ +#endif +#else +#define FILETYPE_ATTRS +#define FILETYPE_STAT_VARS +#define FILETYPE_INIT_STATS(q) +#define FILETYPE_EARLY_STATS(q,a) +#define FILETYPE_LATE_STATS(q,a) +#define FILETYPE_SHOW_STATS(q,f) +#endif + +#ifdef CONFIG_LANG +/* You can use language matching without CONFIG_FILETYPE, but you have to define CA_GET_FILE_LANG yourself. */ +#define LANG_ATTRS SMALL_SET_ATTR(lang, LANG, CA_GET_FILE_LANG, ext_lang_parse) +byte *ext_lang_parse(u32 *dest, byte *value, uns intval); +#else +#define LANG_ATTRS +#endif + +/* + * A list of all extended attributes: custom attributes and also some + * built-in attributes treated in the same way. + */ + +#define EXTENDED_ATTRS CUSTOM_ATTRS FILETYPE_ATTRS LANG_ATTRS + +/* + * A list of all statistics collectors, also composed of custom parts + * and built-in parts. + */ + +#ifndef CUSTOM_STAT_VARS +#define CUSTOM_STAT_VARS +#define CUSTOM_INIT_STATS(q) +#define CUSTOM_EARLY_STATS(q,a) +#define CUSTOM_LATE_STATS(q,a) +#define CUSTOM_SHOW_STATS(q,f) +#endif + +#define EXTENDED_STAT_VARS CUSTOM_STAT_VARS FILETYPE_STAT_VARS +#define EXTENDED_INIT_STATS(q) CUSTOM_INIT_STATS(q) FILETYPE_INIT_STATS(q) +#define EXTENDED_EARLY_STATS(q,a) CUSTOM_EARLY_STATS(q,a) FILETYPE_EARLY_STATS(q,a) +#define EXTENDED_LATE_STATS(q,a) CUSTOM_LATE_STATS(q,a) FILETYPE_LATE_STATS(q,a) +#define EXTENDED_SHOW_STATS(q,f) CUSTOM_SHOW_STATS(q,f) FILETYPE_SHOW_STATS(q,f) /* String fingerprints */ @@ -91,72 +171,19 @@ fp_hash(struct fingerprint *fp) return (fp->hash[0] << 24) | (fp->hash[1] << 16) | (fp->hash[2] << 8) | fp->hash[3]; } -/* Reading of tagged text (Unicode values, tags mapped to 0x80000000 and higher) */ - -#define GET_TAGGED_CHAR(p,u) do { \ - u = *p; \ - if (u >= 0xc0) \ - GET_UTF8_CHAR(p,u); \ - else if (u >= 0x80) \ - { \ - p++; \ - if (u >= 0xb0) \ - { \ - ASSERT(u == 0xb0); \ - u += 0x80020000; \ - } \ - else if (u >= 0xa0) \ - { \ - ASSERT(*p >= 0x80 && *p <= 0xbf); \ - u = 0x80010000 + ((u & 0x0f) << 6) + (*p++ & 0x3f); \ - } \ - else \ - u += 0x80000000; \ - } \ - else \ - p++; \ -} while (0) - -#define SKIP_TAGGED_CHAR(p) do { \ - if (*p >= 0x80 && *p < 0xc0) \ - { \ - uns u = *p++; \ - if (u >= 0xa0 && u < 0xb0 && *p >= 0x80 && *p < 0xc0) \ - p++; \ - } \ - else \ - UTF8_SKIP(p); \ -} while (0) - -static inline uns -bget_tagged_char(struct fastbuf *f) -{ - uns u = bgetc(f); - if ((int)u < 0x80) - ; - else if (u < 0xc0) - { - if (u >= 0xb0) - { - ASSERT(u == 0xb0); - u += 0x80020000; - } - else if (u >= 0xa0) - { - uns v = bgetc(f); - ASSERT(v >= 0x80 && v <= 0xbf); - u = 0x80010000 + ((u & 0x0f) << 6) + (v & 0x3f); - } - else - u += 0x80000000; - } - else - { - bungetc(f); - u = bget_utf8(f); - } - return u; -} +/* The card fingerprints */ + +struct card_print { + struct fingerprint fp; + u32 cardid; +}; + +/* URL keys */ + +#define URL_KEY_BUF_SIZE (3*MAX_URL_SIZE) +byte *url_key(byte *url, byte *buf); +void url_fingerprint(byte *url, struct fingerprint *fp); +void url_key_init(void); /* Conversion of document age from seconds to our internal units */