X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fextract.c;h=5071fd74fbfae9a8c5f7a25d83fe43babda50017;hp=23f0b6ed2102c96a8a6cb1fa326b83d5d15fe681;hb=e2e073b5c947e996304ed7d577497af5e9a879ee;hpb=a030c87bc444608639905eca95e29f84a4f1d991 diff --git a/index/extract.c b/index/extract.c index 23f0b6e..5071fd7 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,5 +1,5 @@ /* This file is part of the Zebra server. - Copyright (C) 1994-2011 Index Data + Copyright (C) Index Data Zebra is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free @@ -21,6 +21,9 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA \brief indexes records and extract tokens for indexing and sorting */ +#if HAVE_CONFIG_H +#include +#endif #include #include #include @@ -47,11 +50,18 @@ static int log_level_initialized = 0; /* eventually this the 0-case code will be removed */ #define FLUSH2 1 -void extract_flush_record_keys2(ZebraHandle zh, zint sysno, - zebra_rec_keys_t ins_keys, - zint ins_rank, - zebra_rec_keys_t del_keys, - zint del_rank); +#if FLUSH2 +static void extract_flush_record_keys2(ZebraHandle zh, zint sysno, + zebra_rec_keys_t ins_keys, + zint ins_rank, + zebra_rec_keys_t del_keys, + zint del_rank); +#else +static void extract_flush_record_keys(ZebraHandle zh, zint sysno, + int cmd, + zebra_rec_keys_t reckeys, + zint staticrank); +#endif static void zebra_init_log_level(void) { @@ -100,8 +110,8 @@ static void logRecord(ZebraHandle zh) if (!(zh->records_processed % 1000)) { yaz_log(YLOG_LOG, "Records: "ZINT_FORMAT" i/u/d " - ZINT_FORMAT"/"ZINT_FORMAT"/"ZINT_FORMAT, - zh->records_processed, zh->records_inserted, + ZINT_FORMAT"/"ZINT_FORMAT"/"ZINT_FORMAT, + zh->records_processed, zh->records_inserted, zh->records_updated, zh->records_deleted); } } @@ -112,7 +122,7 @@ static void init_extractCtrl(ZebraHandle zh, struct recExtractCtrl *ctrl) } -static void extract_add_index_string(RecWord *p, +static void extract_add_index_string(RecWord *p, zinfo_index_category_t cat, const char *str, int length); @@ -134,34 +144,24 @@ struct snip_rec_info { zebra_snippets *snippets; }; - -static void snippet_add_complete_field(RecWord *p, int ord, - zebra_map_t zm) +static int parse_complete_field(RecWord *p, zebra_map_t zm, + char *buf) { - struct snip_rec_info *h = 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; - const char *start = b; - const char *last = 0; if (remain > 0) map = zebra_maps_input(zm, &b, remain, 1); - while (remain > 0 && i < IT_MAX_WORD) { while (map && *map && **map == *CHR_SPACE) { remain = p->term_len - (b - p->term_buf); - if (i == 0) - start = b; /* set to first non-ws area */ if (remain > 0) { int first = i ? 0 : 1; /* first position */ - map = zebra_maps_input(zm, &b, remain, first); } else @@ -187,7 +187,6 @@ static void snippet_add_complete_field(RecWord *p, int ord, while (i < IT_MAX_WORD && *cp) buf[i++] = *(cp++); } - last = b; remain = p->term_len - (b - p->term_buf); if (remain > 0) { @@ -197,11 +196,23 @@ static void snippet_add_complete_field(RecWord *p, int ord, map = 0; } } + return i; +} + +static void snippet_add_complete_field(RecWord *p, int ord, + zebra_map_t zm) +{ + struct snip_rec_info *h = p->extractCtrl->handle; + char buf[IT_MAX_WORD+1]; + int i = parse_complete_field(p, zm, buf); + if (!i) - return; - if (last && start != last && zebra_maps_is_index(zm)) + return; + + if (p->term_len && p->term_buf && zebra_maps_is_index(zm)) zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, - start, last - start); + p->term_buf, p->term_len); + p->seqno++; } static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) @@ -219,8 +230,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) while (map) { - char buf[IT_MAX_WORD+1]; - int i, remain; + int remain; /* Skip spaces */ while (map && *map && **map == *CHR_SPACE) @@ -238,17 +248,10 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) { zebra_snippets_appendn(h->snippets, p->seqno, 1, ord, start, last - start); - } start = last; - - i = 0; while (map && *map && **map != *CHR_SPACE) { - const char *cp = *map; - - while (i < IT_MAX_WORD && *cp) - buf[i++] = *(cp++); remain = p->term_len - (b - p->term_buf); last = b; if (remain > 0) @@ -256,11 +259,11 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) else map = 0; } - if (!i) - return; + if (start == last) + return ; if (first) - { + { first = 0; if (zebra_maps_is_first_in_field(zm)) { @@ -334,7 +337,6 @@ void extract_snippet(ZebraHandle zh, zebra_snippets *sn, { struct recExtractCtrl extractCtrl; struct snip_rec_info info; - int r; extractCtrl.stream = stream; extractCtrl.first_record = 1; @@ -345,20 +347,19 @@ void extract_snippet(ZebraHandle zh, zebra_snippets *sn, assert(zh->reg->dh); extractCtrl.dh = zh->reg->dh; - + info.zh = zh; info.snippets = sn; extractCtrl.handle = &info; extractCtrl.match_criteria[0] = '\0'; extractCtrl.staticrank = 0; extractCtrl.action = action_insert; - + init_extractCtrl(zh, &extractCtrl); extractCtrl.setStoreData = 0; - r = (*rt->extract)(recTypeClientData, &extractCtrl); - + (*rt->extract)(recTypeClientData, &extractCtrl); } static void searchRecordKey(ZebraHandle zh, @@ -395,11 +396,11 @@ static void searchRecordKey(ZebraHandle zh, assert(key.len <= IT_KEY_LEVEL_MAX && key.len > 2); seqno = key.mem[key.len-1]; - + if (key.mem[0] == ch) { zint woff; - + if (startSeq == -1) startSeq = seqno; woff = seqno - startSeq; @@ -432,15 +433,15 @@ static char *get_match_from_spec(ZebraHandle zh, char attset_str[64], attname_str[64]; int i; int first = 1; - + for (s++; strchr(FILE_MATCH_BLANK, *s); s++) ; - for (i = 0; *s && *s != ',' && *s != ')' && + for (i = 0; *s && *s != ',' && *s != ')' && !strchr(FILE_MATCH_BLANK, *s); s++) if (i+1 < sizeof(attset_str)) attset_str[i++] = *s; attset_str[i] = '\0'; - + for (; strchr(FILE_MATCH_BLANK, *s); s++) ; if (*s != ',') @@ -449,7 +450,7 @@ static char *get_match_from_spec(ZebraHandle zh, { for (s++; strchr(FILE_MATCH_BLANK, *s); s++) ; - for (i = 0; *s && *s != ')' && + for (i = 0; *s && *s != ')' && !strchr(FILE_MATCH_BLANK, *s); s++) if (i+1 < sizeof(attname_str)) attname_str[i++] = *s; @@ -465,7 +466,7 @@ static char *get_match_from_spec(ZebraHandle zh, searchRecordKey(zh, reckeys, attname_str, ws, 32); if (0) /* for debugging */ - { + { for (i = 0; i<32; i++) { if (ws[i]) @@ -520,7 +521,7 @@ static char *get_match_from_spec(ZebraHandle zh, } else if (!strcmp(special, "type")) spec_src = zh->m_record_type; - else + else spec_src = NULL; if (spec_src) { @@ -581,7 +582,7 @@ struct recordLogInfo { \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) @@ -605,7 +606,7 @@ static void all_matches_add(struct recExtractCtrl *ctrl, zint record_id, } /* forward declaration */ -ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, +ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, struct ZebraRecStream *stream, enum zebra_recctrl_action_t action, const char *recordType, @@ -616,7 +617,7 @@ ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, void *recTypeClientData); -ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, +ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, enum zebra_recctrl_action_t action) { ZEBRA_RES r = ZEBRA_OK; @@ -624,7 +625,6 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, char gprefix[128]; char ext[128]; char ext_res[128]; - struct file_read_info *fi = 0; const char *original_record_type = 0; RecType recType; void *recTypeClientData; @@ -636,7 +636,7 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, *gprefix = '\0'; else sprintf(gprefix, "%s.", zh->m_group); - + yaz_log(log_level_extract, "zebra_extract_file %s", fname); /* determine file extension */ @@ -690,7 +690,6 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, if (sysno && (action == action_delete || action == action_a_delete)) { streamp = 0; - fi = 0; } else { @@ -704,7 +703,7 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, } else strcpy(full_rep, fname); - + if ((fd = open(full_rep, O_BINARY|O_RDONLY)) == -1) { yaz_log(YLOG_WARN|YLOG_ERRNO, "open %s", full_rep); @@ -731,10 +730,10 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, If sysno is provided, then it's used to identify the reocord. If not, and match_criteria is provided, then sysno is guessed If not, and a record is provided, then sysno is got from there - + */ -ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, +ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, const char *buf, size_t buf_size, enum zebra_recctrl_action_t action, const char *recordType, @@ -753,7 +752,7 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, "Record type explicitly specified: %s", recordType); recType = recType_byName(zh->reg->recTypes, zh->res, recordType, &clientData); - } + } else { if (!(zh->m_record_type)) @@ -767,7 +766,7 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, zh->m_record_type, &clientData); recordType = zh->m_record_type; } - + if (!recType) { yaz_log(YLOG_WARN, "No such record type: %s", recordType); @@ -787,7 +786,7 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, return res; } -static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, +static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, struct ZebraRecStream *stream, enum zebra_recctrl_action_t action, const char *recordType, @@ -797,7 +796,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, RecType recType, void *recTypeClientData, int *more) - + { zint sysno0 = 0; RecordAttr *recordAttr; @@ -807,7 +806,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, Record rec; off_t start_offset = 0, end_offset = 0; const char *pr_fname = fname; /* filename to print .. */ - int show_progress = zh->records_processed + zh->records_skipped + int show_progress = zh->records_processed + zh->records_skipped < zh->m_file_verbose_limit ? 1:0; zebra_init_log_level(); @@ -820,7 +819,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, if (zebraExplain_curDatabase(zh->reg->zei, zh->basenames[0])) { - if (zebraExplain_newDatabase(zh->reg->zei, zh->basenames[0], + if (zebraExplain_newDatabase(zh->reg->zei, zh->basenames[0], zh->m_explain_database)) return ZEBRA_FAIL; } @@ -833,7 +832,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, start_offset = stream->tellf(stream); extractCtrl.first_record = start_offset ? 0 : 1; - + stream->endf(stream, &null_offset);; extractCtrl.init = extract_init; @@ -848,14 +847,14 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, init_extractCtrl(zh, &extractCtrl); extract_set_store_data_prepare(&extractCtrl); - + r = (*recType->extract)(recTypeClientData, &extractCtrl); if (action == action_update) { action = extractCtrl.action; } - + switch (r) { case RECCTRL_EXTRACT_EOF: @@ -873,7 +872,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, yaz_log(YLOG_LOG, "skip %s %s " ZINT_FORMAT, recordType, pr_fname, (zint) start_offset); *more = 1; - + end_offset = stream->endf(stream, 0); if (end_offset) stream->seekf(stream, end_offset); @@ -904,18 +903,18 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, /* test mode .. Do not perform match */ return ZEBRA_OK; } - + if (!sysno) { sysno = &sysno0; - + if (match_criteria && *match_criteria) matchStr = match_criteria; else { if (zh->m_record_id && *zh->m_record_id) { - matchStr = get_match_from_spec(zh, zh->reg->keys, pr_fname, + matchStr = get_match_from_spec(zh, zh->reg->keys, pr_fname, zh->m_record_id); if (!matchStr) { @@ -936,13 +935,13 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, } } } - if (matchStr) + if (matchStr) { int db_ord = zebraExplain_get_database_ord(zh->reg->zei); char *rinfo = dict_lookup_ord(zh->reg->matchDict, db_ord, matchStr); - + if (log_level_extract) { WRBUF w = wrbuf_hex_str(matchStr); @@ -1022,7 +1021,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, #endif recordAttr->staticrank = extractCtrl.staticrank; zh->records_inserted++; - } + } else { /* record already exists */ @@ -1030,7 +1029,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, zebra_rec_keys_t sortKeys = zebra_rec_keys_open(); if (action == action_insert) { - yaz_log(YLOG_LOG, "skipped %s %s " ZINT_FORMAT, + yaz_log(YLOG_LOG, "skipped %s %s " ZINT_FORMAT, recordType, pr_fname, (zint) start_offset); logRecord(zh); return ZEBRA_FAIL; @@ -1045,7 +1044,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, zebra_rec_keys_get_custom_record_id(zh->reg->keys), *sysno); } - + recordAttr = rec_init_attr(zh->reg->zei, rec); /* decrease total size */ @@ -1072,7 +1071,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, #if FLUSH2 extract_flush_record_keys2(zh, *sysno, 0, recordAttr->staticrank, delkeys, recordAttr->staticrank); -#endif +#endif if (zebra_rec_keys_empty(delkeys)) { yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, @@ -1111,7 +1110,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, zh->reg->keys, extractCtrl.staticrank, delkeys, recordAttr->staticrank); #else - extract_flush_record_keys(zh, *sysno, 1, + extract_flush_record_keys(zh, *sysno, 1, zh->reg->keys, extractCtrl.staticrank); #endif recordAttr->staticrank = extractCtrl.staticrank; @@ -1192,11 +1191,11 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, /* update database name */ xfree(rec->info[recInfo_databaseName]); rec->info[recInfo_databaseName] = - rec_strdup(zh->basenames[0], &rec->size[recInfo_databaseName]); + rec_strdup(zh->basenames[0], &rec->size[recInfo_databaseName]); /* update offset */ recordAttr->recordOffset = start_offset; - + /* commit this record */ rec_put(zh->reg->records, &rec); logRecord(zh); @@ -1215,7 +1214,7 @@ static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, \param recTypeClientData client data for record type \returns ZEBRA_OK for success; ZEBRA_FAIL for failure */ -ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, +ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, struct ZebraRecStream *stream, enum zebra_recctrl_action_t action, const char *recordType, @@ -1280,7 +1279,7 @@ ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n) extractCtrl.handle = handle; extractCtrl.first_record = 1; - + extract_set_store_data_prepare(&extractCtrl); if (n) @@ -1289,14 +1288,14 @@ ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n) if (rec->size[recInfo_delKeys]) { zebra_rec_keys_t delkeys = zebra_rec_keys_open(); - + zebra_rec_keys_t sortkeys = zebra_rec_keys_open(); zebra_rec_keys_set_buf(delkeys, rec->info[recInfo_delKeys], rec->size[recInfo_delKeys], 0); #if FLUSH2 - extract_flush_record_keys2(zh, rec->sysno, + extract_flush_record_keys2(zh, rec->sysno, zh->reg->keys, 0, delkeys, 0); #else extract_flush_record_keys(zh, rec->sysno, 0, delkeys, 0); @@ -1316,14 +1315,14 @@ ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n) #if FLUSH2 extract_flush_record_keys2(zh, rec->sysno, zh->reg->keys, 0, 0, 0); #else - extract_flush_record_keys(zh, rec->sysno, 1, zh->reg->keys, 0); + extract_flush_record_keys(zh, rec->sysno, 1, zh->reg->keys, 0); #endif } extract_flush_sort_keys(zh, rec->sysno, 1, zh->reg->sortKeys); - + xfree(rec->info[recInfo_delKeys]); zebra_rec_keys_get_buf(zh->reg->keys, - &rec->info[recInfo_delKeys], + &rec->info[recInfo_delKeys], &rec->size[recInfo_delKeys]); xfree(rec->info[recInfo_sortKeys]); @@ -1337,30 +1336,27 @@ void zebra_it_key_str_dump(ZebraHandle zh, struct it_key *key, const char *str, size_t slen, NMEM nmem, int level) { char keystr[200]; /* room for zints to print */ - char *dst_term = 0; int ord = CAST_ZINT_TO_INT(key->mem[0]); const char *index_type; int i; const char *string_index; - + zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, 0/* db */, &string_index); assert(index_type); - zebra_term_untrans_iconv(zh, nmem, index_type, - &dst_term, str); *keystr = '\0'; for (i = 0; i < key->len; i++) { sprintf(keystr + strlen(keystr), ZINT_FORMAT " ", key->mem[i]); } - + if (*str < CHR_BASE_CHAR) { int i; char dst_buf[200]; /* room for special chars */ - + strcpy(dst_buf , "?"); - + if (!strcmp(str, "")) strcpy(dst_buf, "alwaysmatches"); if (!strcmp(str, FIRST_IN_FIELD_STR)) @@ -1369,18 +1365,30 @@ void zebra_it_key_str_dump(ZebraHandle zh, struct it_key *key, strcpy(dst_buf, "unknown"); else if (!strcmp(str, CHR_SPACE)) strcpy(dst_buf, "space"); - + for (i = 0; ireg->zei; int normal = 0; @@ -1506,7 +1516,7 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, &ins_key_in); if (del && ins && ins_rank == del_rank - && !key_compare(&del_key_in, &ins_key_in) + && !key_compare(&del_key_in, &ins_key_in) && ins_slen == del_slen && !memcmp(del_str, ins_str, del_slen)) { optimized++; @@ -1514,27 +1524,70 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, } if (!del && !ins) break; - + normal++; if (del) - key_block_write(zh->reg->key_block, sysno, + key_block_write(zh->reg->key_block, sysno, &del_key_in, 0, del_str, del_slen, del_rank, zh->m_staticrank); if (ins) - key_block_write(zh->reg->key_block, sysno, + key_block_write(zh->reg->key_block, sysno, &ins_key_in, 1, ins_str, ins_slen, ins_rank, zh->m_staticrank); } yaz_log(log_level_extract, "normal=%d optimized=%d", normal, optimized); } +#else +static 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); + } + } +} +#endif ZEBRA_RES zebra_rec_keys_to_snippets1(ZebraHandle zh, zebra_rec_keys_t reckeys, zebra_snippets *snippets) { NMEM nmem = nmem_create(); - if (zebra_rec_keys_rewind(reckeys)) + if (zebra_rec_keys_rewind(reckeys)) { const char *str; size_t slen; @@ -1549,7 +1602,7 @@ ZEBRA_RES zebra_rec_keys_to_snippets1(ZebraHandle zh, assert(key.len <= IT_KEY_LEVEL_MAX && key.len > 2); seqno = key.mem[key.len-1]; ord = CAST_ZINT_TO_INT(key.mem[0]); - + zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, 0/* db */, 0 /* string_index */); assert(index_type); @@ -1581,13 +1634,13 @@ void print_rec_keys(ZebraHandle zh, zebra_rec_keys_t reckeys) assert(key.len <= IT_KEY_LEVEL_MAX && key.len > 2); zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, &db, 0); - + seqno = key.mem[key.len-1]; - + zebra_term_untrans(zh, index_type, dst_buf, str); - - yaz_log(YLOG_LOG, "ord=%d seqno=" ZINT_FORMAT - " term=%s", ord, seqno, dst_buf); + + yaz_log(YLOG_LOG, "ord=%d seqno=" ZINT_FORMAT + " term=%s", ord, seqno, dst_buf); } } } @@ -1661,7 +1714,7 @@ static void extract_add_string(RecWord *p, zebra_map_t zm, { WRBUF w = wrbuf_alloc(); - + wrbuf_write_escaped(w, string, length); yaz_log(log_level_details, "extract_add_string: %s", wrbuf_cstr(w)); wrbuf_destroy(w); @@ -1696,7 +1749,7 @@ static void extract_add_incomplete_field(RecWord *p, zebra_map_t zm) int remain = p->term_len; int first = 1; const char **map = 0; - + if (remain > 0) map = zebra_maps_input(zm, &b, remain, 0); @@ -1733,7 +1786,7 @@ static void extract_add_incomplete_field(RecWord *p, zebra_map_t zm) return; if (first) - { + { first = 0; if (zebra_maps_is_first_in_field(zm)) { @@ -1749,60 +1802,12 @@ static void extract_add_incomplete_field(RecWord *p, zebra_map_t zm) static void extract_add_complete_field(RecWord *p, zebra_map_t zm) { - 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(zm, &b, remain, 1); - - while (remain > 0 && i < IT_MAX_WORD) - { - while (map && *map && **map == *CHR_SPACE) - { - remain = p->term_len - (b - p->term_buf); - - if (remain > 0) - { - int first = i ? 0 : 1; /* first position */ - map = zebra_maps_input(zm, &b, remain, first); - } - else - map = 0; - } - if (!map) - break; - - if (i && i < IT_MAX_WORD) - buf[i++] = *CHR_SPACE; - while (map && *map && **map != *CHR_SPACE) - { - const char *cp = *map; - - if (**map == *CHR_CUT) - { - i = 0; - } - else - { - if (i >= IT_MAX_WORD) - break; - while (i < IT_MAX_WORD && *cp) - buf[i++] = *(cp++); - } - remain = p->term_len - (b - p->term_buf); - if (remain > 0) - { - map = zebra_maps_input(zm, &b, remain, 0); - } - else - map = 0; - } - } + int i = parse_complete_field(p, zm, buf); if (!i) return; extract_add_string(p, zm, buf, i); + p->seqno++; } static void extract_add_icu(RecWord *p, zebra_map_t zm) @@ -1813,6 +1818,11 @@ static void extract_add_icu(RecWord *p, zebra_map_t zm) zebra_map_tokenize_start(zm, p->term_buf, p->term_len); while (zebra_map_tokenize_next(zm, &res_buf, &res_len, 0, 0)) { + if (res_len > IT_MAX_WORD) + { + yaz_log(YLOG_LOG, "Truncating long term %ld", (long) res_len); + res_len = IT_MAX_WORD; + } extract_add_string(p, zm, res_buf, res_len); p->seqno++; } @@ -1826,32 +1836,26 @@ static void extract_add_icu(RecWord *p, zebra_map_t zm) 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=%s index=%s seqno=" ZINT_FORMAT " s=%.*s", - p->index_type, p->index_name, + p->index_type, p->index_name, p->seqno, p->term_len, p->term_buf); } - if ((wrbuf = zebra_replace(zm, 0, p->term_buf, p->term_len))) - { - p->term_buf = wrbuf_buf(wrbuf); - p->term_len = wrbuf_len(wrbuf); - } if (zebra_maps_is_icu(zm)) { extract_add_icu(p, zm); @@ -1943,7 +1947,7 @@ void extract_flush_sort_keys(ZebraHandle zh, zint sysno, (*e)->sysno = filter_sysno ? filter_sysno : sysno; (*e)->section_id = section_id; } - + wrbuf_write((*e)->wrbuf, str, slen); wrbuf_putc((*e)->wrbuf, '\0'); }