X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fextract.c;h=be000c2dd8ae2b11732e94979cbc380bf5d5cc0a;hp=24e5b9ef5eb480307b3b33c47b4aae294aba273d;hb=4097bac1dce126e79ce1550fad8b2034a311ff80;hpb=d82c0efad7971d102220a8824e1ea674db5b7fe2 diff --git a/index/extract.c b/index/extract.c index 24e5b9e..be000c2 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,4 +1,4 @@ -/* $Id: extract.c,v 1.263 2007-10-29 09:25:40 adam Exp $ +/* $Id: extract.c,v 1.275 2007-12-20 11:15:42 adam Exp $ Copyright (C) 1995-2007 Index Data ApS @@ -20,6 +20,10 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ +/** \file + \brief indexes records and extract tokens for indexing and sorting +*/ + #include #include #include @@ -31,10 +35,12 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA #endif #include + #include "index.h" #include "orddict.h" #include #include +#include static int log_level_extract = 0; static int log_level_details = 0; @@ -61,9 +67,6 @@ static void zebra_init_log_level(void) } } -static void extract_flush_record_keys(ZebraHandle zh, zint sysno, - int cmd, zebra_rec_keys_t reckeys, - zint staticrank); static void extract_flush_sort_keys(ZebraHandle zh, zint sysno, int cmd, zebra_rec_keys_t skp); static void extract_schema_add(struct recExtractCtrl *p, Odr_oid *oid); @@ -93,14 +96,6 @@ static void logRecord(ZebraHandle zh) static void init_extractCtrl(ZebraHandle zh, struct recExtractCtrl *ctrl) { - int i; - for (i = 0; i<256; i++) - { - if (zebra_maps_is_positioned(zh->reg->zebra_maps, i)) - ctrl->seqno[i] = 1; - else - ctrl->seqno[i] = 0; - } ctrl->flagShowRecords = !zh->m_flag_rw; } @@ -115,7 +110,7 @@ static void extract_init(struct recExtractCtrl *p, RecWord *w) { w->seqno = 1; w->index_name = "any"; - w->index_type = 'w'; + w->index_type = "w"; w->extractCtrl = p; w->record_id = 0; w->section_id = 0; @@ -128,10 +123,10 @@ struct snip_rec_info { }; -static void snippet_add_complete_field(RecWord *p, int ord) +static void snippet_add_complete_field(RecWord *p, int ord, + zebra_map_t zm) { struct snip_rec_info *h = p->extractCtrl->handle; - ZebraHandle zh = h->zh; const char *b = p->term_buf; char buf[IT_MAX_WORD+1]; @@ -141,7 +136,7 @@ static void snippet_add_complete_field(RecWord *p, int ord) const char *last = 0; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 1); + map = zebra_maps_input(zm, &b, remain, 1); while (remain > 0 && i < IT_MAX_WORD) { @@ -155,8 +150,7 @@ static void snippet_add_complete_field(RecWord *p, int ord) { int first = i ? 0 : 1; /* first position */ - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, - &b, remain, first); + map = zebra_maps_input(zm, &b, remain, first); } else map = 0; @@ -185,8 +179,7 @@ static void snippet_add_complete_field(RecWord *p, int ord) remain = p->term_len - (b - p->term_buf); if (remain > 0) { - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, - remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); } else map = 0; @@ -199,10 +192,9 @@ static void snippet_add_complete_field(RecWord *p, int ord) start, last - start); } -static void snippet_add_incomplete_field(RecWord *p, int ord) +static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) { struct snip_rec_info *h = p->extractCtrl->handle; - ZebraHandle zh = h->zh; const char *b = p->term_buf; int remain = p->term_len; int first = 1; @@ -211,7 +203,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord) const char *last = b; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); while (map) { @@ -224,8 +216,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord) remain = p->term_len - (b - p->term_buf); last = b; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, - remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); else map = 0; } @@ -249,7 +240,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord) remain = p->term_len - (b - p->term_buf); last = b; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); else map = 0; } @@ -259,7 +250,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord) if (first) { first = 0; - if (zebra_maps_is_first_in_field(zh->reg->zebra_maps, p->index_type)) + if (zebra_maps_is_first_in_field(zm)) { /* first in field marker */ p->seqno++; @@ -274,21 +265,47 @@ static void snippet_add_incomplete_field(RecWord *p, int ord) } +static void snippet_add_icu(RecWord *p, int ord, zebra_map_t zm) +{ + struct snip_rec_info *h = p->extractCtrl->handle; + + const char *res_buf = 0; + size_t res_len = 0; + + const char *display_buf = 0; + size_t display_len = 0; + + zebra_map_tokenize_start(zm, p->term_buf, p->term_len); + while (zebra_map_tokenize_next(zm, &res_buf, &res_len, + &display_buf, &display_len)) + { + zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, + display_buf, display_len); + p->seqno++; + } +} + static void snippet_token_add(RecWord *p) { struct snip_rec_info *h = p->extractCtrl->handle; ZebraHandle zh = h->zh; + zebra_map_t zm = zebra_map_get(zh->reg->zebra_maps, p->index_type); - if (zebra_maps_is_index(zh->reg->zebra_maps, p->index_type)) + if (zm && zebra_maps_is_index(zm)) { ZebraExplainInfo zei = zh->reg->zei; int ch = zebraExplain_lookup_attr_str( zei, zinfo_index_category_index, p->index_type, p->index_name); - if(zebra_maps_is_complete (h->zh->reg->zebra_maps, p->index_type)) - snippet_add_complete_field(p, ch); + if (zebra_maps_is_icu(zm)) + snippet_add_icu(p, ch, zm); else - snippet_add_incomplete_field(p, ch); + { + if (zebra_maps_is_complete(zm)) + snippet_add_complete_field(p, ch, zm); + else + snippet_add_incomplete_field(p, ch, zm); + } } } @@ -344,11 +361,11 @@ static void searchRecordKey(ZebraHandle zh, ws[i] = NULL; if (ch < 0) - ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, '0', index_name); + ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, "0", index_name); if (ch < 0) - ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, 'p', index_name); + ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, "p", index_name); if (ch < 0) - ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, 'w', index_name); + ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, "w", index_name); if (ch < 0) return ; @@ -528,13 +545,29 @@ struct recordLogInfo { struct recordGroup *rGroup; }; -static void all_matches_add(struct recExtractCtrl *ctrl) +/** \brief add the always-matches index entry and map to real record ID + \param ctrl record control + \param record_id custom record ID + \param sysno system record ID + + This function serves two purposes.. It adds the always matches + entry and makes a pointer from the custom record ID (if defined) + back to the system record ID (sysno) + See zebra_recid_to_sysno . + */ +static void all_matches_add(struct recExtractCtrl *ctrl, zint record_id, + zint sysno) { RecWord word; extract_init(ctrl, &word); + word.record_id = record_id; + /* we use the seqno as placeholder for a way to get back to + record database from _ALLRECORDS.. This is used if a custom + RECORD was defined */ + word.seqno = sysno; word.index_name = "_ALLRECORDS"; - word.index_type = 'w'; - word.seqno = 1; + word.index_type = "w"; + extract_add_index_string(&word, zinfo_index_category_alwaysmatches, "", 0); } @@ -552,7 +585,7 @@ ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, - int deleteFlag) + enum zebra_recctrl_action_t action) { ZEBRA_RES r = ZEBRA_OK; int i, fd; @@ -622,7 +655,7 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, default: yaz_log(YLOG_WARN, "Bad filter version: %s", zh->m_record_type); } - if (sysno && deleteFlag) + if (sysno && (action == action_delete || action == action_a_delete)) { streamp = 0; fi = 0; @@ -650,8 +683,7 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, zebra_create_stream_fd(streamp, fd, 0); } r = zebra_extract_records_stream(zh, streamp, - deleteFlag ? - action_delete : action_update, + action, 0, /* tst_mode */ zh->m_record_type, sysno, @@ -882,8 +914,6 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, else end_offset = stream->tellf(stream); - all_matches_add(&extractCtrl); - if (extractCtrl.match_criteria[0]) match_criteria = extractCtrl.match_criteria; } @@ -927,6 +957,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, } } } + if (zebra_rec_keys_empty(zh->reg->keys)) { /* the extraction process returned no information - the record @@ -937,14 +968,21 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, if (! *sysno) { - /* new record */ + /* new record AKA does not exist already */ if (action == action_delete) { - yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, - pr_fname, (zint) start_offset); + yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, + pr_fname, (zint) start_offset); yaz_log(YLOG_WARN, "cannot delete record above (seems new)"); return ZEBRA_FAIL; } + else if (action == action_a_delete) + { + if (show_progress) + yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, + pr_fname, (zint) start_offset); + return ZEBRA_OK; + } else if (action == action_replace) { yaz_log(YLOG_LOG, "update %s %s " ZINT_FORMAT, recordType, @@ -959,6 +997,15 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, *sysno = rec->sysno; + + if (stream) + { + all_matches_add(&extractCtrl, + zebra_rec_keys_get_custom_record_id(zh->reg->keys), + *sysno); + } + + recordAttr = rec_init_attr(zh->reg->zei, rec); if (extractCtrl.staticrank < 0) { @@ -1000,6 +1047,13 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, rec = rec_get(zh->reg->records, *sysno); assert(rec); + + if (stream) + { + all_matches_add(&extractCtrl, + zebra_rec_keys_get_custom_record_id(zh->reg->keys), + *sysno); + } recordAttr = rec_init_attr(zh->reg->zei, rec); @@ -1021,7 +1075,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, extract_flush_record_keys(zh, *sysno, 0, delkeys, recordAttr->staticrank); #endif - if (action == action_delete) + if (action == action_delete || action == action_a_delete) { /* record going to be deleted */ #if FLUSH2 @@ -1258,7 +1312,8 @@ void extract_rec_keys_log(ZebraHandle zh, int is_insert, char keystr[200]; /* room for zints to print */ char *dst_term = 0; int ord = CAST_ZINT_TO_INT(key.mem[0]); - int index_type, i; + const char *index_type; + int i; const char *string_index; zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, @@ -1292,12 +1347,12 @@ void extract_rec_keys_log(ZebraHandle zh, int is_insert, { sprintf(dst_buf + strlen(dst_buf), " %d", str[i] & 0xff); } - yaz_log(level, "%s%c %s %s", keystr, index_type, + yaz_log(level, "%s%s %s %s", keystr, index_type, string_index, dst_buf); } else - yaz_log(level, "%s%c %s \"%s\"", keystr, index_type, + yaz_log(level, "%s%s %s \"%s\"", keystr, index_type, string_index, dst_term); nmem_reset(nmem); @@ -1431,49 +1486,8 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, yaz_log(log_level_extract, "normal=%d optimized=%d", normal, optimized); } -void extract_flush_record_keys(ZebraHandle zh, zint sysno, int cmd, - zebra_rec_keys_t reckeys, - zint staticrank) -{ - ZebraExplainInfo zei = zh->reg->zei; - extract_rec_keys_adjust(zh, cmd, reckeys); - - if (log_level_details) - { - yaz_log(log_level_details, "Keys for record " ZINT_FORMAT " %s", - sysno, cmd ? "insert" : "delete"); - extract_rec_keys_log(zh, cmd, reckeys, log_level_details); - } - - if (!zh->reg->key_block) - { - int mem = 1024*1024 * atoi( res_get_def( zh->res, "memmax", "8")); - const char *key_tmp_dir = res_get_def(zh->res, "keyTmpDir", "."); - int use_threads = atoi(res_get_def(zh->res, "threads", "1")); - zh->reg->key_block = key_block_create(mem, key_tmp_dir, use_threads); - } - zebraExplain_recordCountIncrement(zei, cmd ? 1 : -1); - -#if 0 - yaz_log(YLOG_LOG, "sysno=" ZINT_FORMAT " cmd=%d", sysno, cmd); - print_rec_keys(zh, reckeys); -#endif - if (zebra_rec_keys_rewind(reckeys)) - { - size_t slen; - const char *str; - struct it_key key_in; - while(zebra_rec_keys_read(reckeys, &str, &slen, &key_in)) - { - key_block_write(zh->reg->key_block, sysno, - &key_in, cmd, str, slen, - staticrank, zh->m_staticrank); - } - } -} - -ZEBRA_RES zebra_rec_keys_to_snippets(ZebraHandle zh, +ZEBRA_RES zebra_rec_keys_to_snippets1(ZebraHandle zh, zebra_rec_keys_t reckeys, zebra_snippets *snippets) { @@ -1488,7 +1502,7 @@ ZEBRA_RES zebra_rec_keys_to_snippets(ZebraHandle zh, char *dst_term = 0; int ord; zint seqno; - int index_type; + const char *index_type; assert(key.len <= IT_KEY_LEVEL_MAX && key.len > 2); seqno = key.mem[key.len-1]; @@ -1519,7 +1533,7 @@ void print_rec_keys(ZebraHandle zh, zebra_rec_keys_t reckeys) { char dst_buf[IT_MAX_WORD]; zint seqno; - int index_type; + const char *index_type; int ord = CAST_ZINT_TO_INT(key.mem[0]); const char *db = 0; assert(key.len <= IT_KEY_LEVEL_MAX && key.len > 2); @@ -1593,19 +1607,19 @@ static void extract_add_staticrank_string(RecWord *p, ctrl->staticrank = atozint(valz); } -static void extract_add_string(RecWord *p, const char *string, int length) +static void extract_add_string(RecWord *p, zebra_map_t zm, + const char *string, int length) { - ZebraHandle zh = p->extractCtrl->handle; assert(length > 0); if (!p->index_name) return; - if (zebra_maps_is_index(zh->reg->zebra_maps, p->index_type)) + if (zebra_maps_is_index(zm)) { extract_add_index_string(p, zinfo_index_category_index, string, length); - if (zebra_maps_is_alwaysmatches(zh->reg->zebra_maps, p->index_type)) + if (zebra_maps_is_alwaysmatches(zm)) { RecWord word; memcpy(&word, p, sizeof(word)); @@ -1615,26 +1629,25 @@ static void extract_add_string(RecWord *p, const char *string, int length) &word, zinfo_index_category_alwaysmatches, "", 0); } } - else if (zebra_maps_is_sort(zh->reg->zebra_maps, p->index_type)) + else if (zebra_maps_is_sort(zm)) { extract_add_sort_string(p, string, length); } - else if (zebra_maps_is_staticrank(zh->reg->zebra_maps, p->index_type)) + else if (zebra_maps_is_staticrank(zm)) { extract_add_staticrank_string(p, string, length); } } -static void extract_add_incomplete_field(RecWord *p) +static void extract_add_incomplete_field(RecWord *p, zebra_map_t zm) { - ZebraHandle zh = p->extractCtrl->handle; const char *b = p->term_buf; int remain = p->term_len; int first = 1; const char **map = 0; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); while (map) { @@ -1646,8 +1659,7 @@ static void extract_add_incomplete_field(RecWord *p) { remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, - remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); else map = 0; } @@ -1662,7 +1674,7 @@ static void extract_add_incomplete_field(RecWord *p) buf[i++] = *(cp++); remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); else map = 0; } @@ -1672,28 +1684,27 @@ static void extract_add_incomplete_field(RecWord *p) if (first) { first = 0; - if (zebra_maps_is_first_in_field(zh->reg->zebra_maps, p->index_type)) + if (zebra_maps_is_first_in_field(zm)) { /* first in field marker */ - extract_add_string(p, FIRST_IN_FIELD_STR, FIRST_IN_FIELD_LEN); + extract_add_string(p, zm, FIRST_IN_FIELD_STR, FIRST_IN_FIELD_LEN); p->seqno++; } } - extract_add_string(p, buf, i); + extract_add_string(p, zm, buf, i); p->seqno++; } } -static void extract_add_complete_field(RecWord *p) +static void extract_add_complete_field(RecWord *p, zebra_map_t zm) { - ZebraHandle zh = p->extractCtrl->handle; const char *b = p->term_buf; char buf[IT_MAX_WORD+1]; const char **map = 0; int i = 0, remain = p->term_len; if (remain > 0) - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, 1); + map = zebra_maps_input(zm, &b, remain, 1); while (remain > 0 && i < IT_MAX_WORD) { @@ -1704,7 +1715,7 @@ static void extract_add_complete_field(RecWord *p) if (remain > 0) { int first = i ? 0 : 1; /* first position */ - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, remain, first); + map = zebra_maps_input(zm, &b, remain, first); } else map = 0; @@ -1732,8 +1743,7 @@ static void extract_add_complete_field(RecWord *p) remain = p->term_len - (b - p->term_buf); if (remain > 0) { - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &b, - remain, 0); + map = zebra_maps_input(zm, &b, remain, 0); } else map = 0; @@ -1741,31 +1751,67 @@ static void extract_add_complete_field(RecWord *p) } if (!i) return; - extract_add_string(p, buf, i); + extract_add_string(p, zm, buf, i); +} + +static void extract_add_icu(RecWord *p, zebra_map_t zm) +{ + const char *res_buf = 0; + size_t res_len = 0; + + zebra_map_tokenize_start(zm, p->term_buf, p->term_len); + while (zebra_map_tokenize_next(zm, &res_buf, &res_len, 0, 0)) + { + extract_add_string(p, zm, res_buf, res_len); + p->seqno++; + } } + +/** \brief top-level indexing handler for recctrl system + \param p token data to be indexed + + Call sequence: + extract_token_add + extract_add_{in}_complete / extract_add_icu + extract_add_string + + extract_add_index_string + or + extract_add_sort_string + or + extract_add_staticrank_string + +*/ static void extract_token_add(RecWord *p) { ZebraHandle zh = p->extractCtrl->handle; + zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, p->index_type); WRBUF wrbuf; if (log_level_details) { yaz_log(log_level_details, "extract_token_add " - "type=%c index=%s seqno=" ZINT_FORMAT " s=%.*s", + "type=%s index=%s seqno=" ZINT_FORMAT " s=%.*s", p->index_type, p->index_name, p->seqno, p->term_len, p->term_buf); } - if ((wrbuf = zebra_replace(zh->reg->zebra_maps, p->index_type, 0, - p->term_buf, p->term_len))) + if ((wrbuf = zebra_replace(zm, 0, p->term_buf, p->term_len))) { - p->term_buf = wrbuf_buf(wrbuf); - p->term_len = wrbuf_len(wrbuf); + p->term_buf = wrbuf_buf(wrbuf); + p->term_len = wrbuf_len(wrbuf); + } + if (zebra_maps_is_icu(zm)) + { + extract_add_icu(p, zm); } - if (zebra_maps_is_complete(zh->reg->zebra_maps, p->index_type)) - extract_add_complete_field(p); else - extract_add_incomplete_field(p); + { + if (zebra_maps_is_complete(zm)) + extract_add_complete_field(p, zm); + else + extract_add_incomplete_field(p, zm); + } } static void extract_set_store_data_cb(struct recExtractCtrl *p,