X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fextract.c;h=5071fd74fbfae9a8c5f7a25d83fe43babda50017;hp=d9ed681bbb83945145099c1275a50696f6888c24;hb=e2e073b5c947e996304ed7d577497af5e9a879ee;hpb=5623222cf968d493c2bc9506c42b8344de60c4e5 diff --git a/index/extract.c b/index/extract.c index d9ed681..5071fd7 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,8 +1,5 @@ -/* $Id: extract.c,v 1.260 2007-08-21 13:27:04 adam Exp $ - Copyright (C) 1995-2007 - Index Data ApS - -This file is part of the Zebra server. +/* This file is part of the Zebra server. + 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 @@ -20,6 +17,13 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ +/** \file + \brief indexes records and extract tokens for indexing and sorting +*/ + +#if HAVE_CONFIG_H +#include +#endif #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; @@ -44,11 +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) { @@ -61,13 +74,25 @@ 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 WRBUF wrbuf_hex_str(const char *cstr) +{ + size_t i; + WRBUF w = wrbuf_alloc(); + for (i = 0; cstr[i]; i++) + { + if (cstr[i] < ' ' || cstr[i] > 126) + wrbuf_printf(w, "\\%02X", cstr[i] & 0xff); + else + wrbuf_putc(w, cstr[i]); + } + return w; +} + + 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); -static void extract_token_add (RecWord *p); +static void extract_schema_add(struct recExtractCtrl *p, Odr_oid *oid); +static void extract_token_add(RecWord *p); static void check_log_limit(ZebraHandle zh) { @@ -78,34 +103,26 @@ static void check_log_limit(ZebraHandle zh) } } -static void logRecord (ZebraHandle zh) +static void logRecord(ZebraHandle zh) { check_log_limit(zh); ++zh->records_processed; 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); } } 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; } -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); @@ -115,7 +132,7 @@ 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; @@ -127,36 +144,25 @@ struct snip_rec_info { zebra_snippets *snippets; }; - -static void snippet_add_complete_field(RecWord *p, int ord) +static int parse_complete_field(RecWord *p, zebra_map_t zm, + char *buf) { - struct snip_rec_info *h = p->extractCtrl->handle; - ZebraHandle zh = h->zh; - 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 (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) { 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(zh->reg->zebra_maps, p->index_type, - &b, remain, first); + map = zebra_maps_input(zm, &b, remain, first); } else map = 0; @@ -181,42 +187,50 @@ 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) { - 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; } } + 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) + 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) +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; const char **map = 0; const char *start = b; - + 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) { - char buf[IT_MAX_WORD+1]; - const char *last = b; - int i, remain; + int remain; /* Skip spaces */ while (map && *map && **map == *CHR_SPACE) @@ -224,48 +238,40 @@ 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; } if (!map) break; - if (start != last) + if (start != last && zebra_maps_is_index(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) - 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; } - if (!i) - return; + if (start == last) + return ; 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++; } } - if (start != last) + if (start != last && zebra_maps_is_index(zm)) zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, start, last - start); start = last; @@ -274,21 +280,48 @@ 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)) + { + if (zebra_maps_is_index(zm)) + 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_or_add(zh->reg->zebra_maps, p->index_type); - if (zebra_maps_is_index(zh->reg->zebra_maps, p->index_type)) + if (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); + } } } @@ -304,7 +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; @@ -315,18 +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); - r = (*rt->extract)(recTypeClientData, &extractCtrl); + extractCtrl.setStoreData = 0; + (*rt->extract)(recTypeClientData, &extractCtrl); } static void searchRecordKey(ZebraHandle zh, @@ -342,11 +375,11 @@ 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 ; @@ -363,11 +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; @@ -400,15 +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 != ',') @@ -417,23 +450,34 @@ 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; attname_str[i] = '\0'; } - - searchRecordKey (zh, reckeys, attname_str, ws, 32); - if (*s != ')') { - yaz_log (YLOG_WARN, "Missing ) in match criteria %s in group %s", + yaz_log(YLOG_WARN, "Missing ) in match criteria %s in group %s", spec, zh->m_group ? zh->m_group : "none"); return NULL; } s++; + searchRecordKey(zh, reckeys, attname_str, ws, 32); + if (0) /* for debugging */ + { + for (i = 0; i<32; i++) + { + if (ws[i]) + { + WRBUF w = wrbuf_hex_str(ws[i]); + yaz_log(YLOG_LOG, "ws[%d] = %s", i, wrbuf_cstr(w)); + wrbuf_destroy(w); + } + } + } + for (i = 0; i<32; i++) if (ws[i]) { @@ -442,12 +486,12 @@ static char *get_match_from_spec(ZebraHandle zh, *dst++ = ' '; first = 0; } - strcpy (dst, ws[i]); + strcpy(dst, ws[i]); dst += strlen(ws[i]); } if (first) { - yaz_log (YLOG_WARN, "Record didn't contain match" + yaz_log(YLOG_WARN, "Record didn't contain match" " fields in (%s,%s)", attset_str, attname_str); return NULL; } @@ -464,25 +508,25 @@ static char *get_match_from_spec(ZebraHandle zh, spec_len = s1 - s; if (spec_len > sizeof(special)-1) spec_len = sizeof(special)-1; - memcpy (special, s, spec_len); + memcpy(special, s, spec_len); special[spec_len] = '\0'; s = s1; - if (!strcmp (special, "group")) + if (!strcmp(special, "group")) spec_src = zh->m_group; - else if (!strcmp (special, "database")) + else if (!strcmp(special, "database")) spec_src = zh->basenames[0]; - else if (!strcmp (special, "filename")) { + else if (!strcmp(special, "filename")) { spec_src = fname; } - else if (!strcmp (special, "type")) + else if (!strcmp(special, "type")) spec_src = zh->m_record_type; - else + else spec_src = NULL; if (spec_src) { - strcpy (dst, spec_src); - dst += strlen (spec_src); + strcpy(dst, spec_src); + dst += strlen(spec_src); } } else if (*s == '\"' || *s == '\'') @@ -499,12 +543,12 @@ static char *get_match_from_spec(ZebraHandle zh, if (*s) s++; tmpString[i] = '\0'; - strcpy (dst, tmpString); - dst += strlen (tmpString); + strcpy(dst, tmpString); + dst += strlen(tmpString); } else { - yaz_log (YLOG_WARN, "Syntax error in match criteria %s in group %s", + yaz_log(YLOG_WARN, "Syntax error in match criteria %s in group %s", spec, zh->m_group ? zh->m_group : "none"); return NULL; } @@ -512,11 +556,19 @@ static char *get_match_from_spec(ZebraHandle zh, } if (dst == dstBuf) { - yaz_log (YLOG_WARN, "No match criteria for record %s in group %s", + yaz_log(YLOG_WARN, "No match criteria for record %s in group %s", fname, zh->m_group ? zh->m_group : "none"); return NULL; } *dst = '\0'; + + if (0) /* for debugging */ + { + WRBUF w = wrbuf_hex_str(dstBuf); + yaz_log(YLOG_LOG, "get_match_from_spec %s", wrbuf_cstr(w)); + wrbuf_destroy(w); + } + return dstBuf; } @@ -526,21 +578,37 @@ 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; - extract_add_index_string (&word, zinfo_index_category_alwaysmatches, + word.index_type = "w"; + + extract_add_index_string(&word, zinfo_index_category_alwaysmatches, "", 0); } -ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, +/* forward declaration */ +ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, struct ZebraRecStream *stream, enum zebra_recctrl_action_t action, - int test_mode, const char *recordType, zint *sysno, const char *match_criteria, @@ -549,15 +617,14 @@ ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, void *recTypeClientData); -ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, - int deleteFlag) +ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, + enum zebra_recctrl_action_t action) { ZEBRA_RES r = ZEBRA_OK; int i, fd; 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; @@ -568,8 +635,8 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, if (!zh->m_group || !*zh->m_group) *gprefix = '\0'; else - sprintf (gprefix, "%s.", zh->m_group); - + sprintf(gprefix, "%s.", zh->m_group); + yaz_log(log_level_extract, "zebra_extract_file %s", fname); /* determine file extension */ @@ -579,34 +646,34 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, break; else if (fname[i] == '.') { - strcpy (ext, fname+i+1); + strcpy(ext, fname+i+1); break; } /* determine file type - depending on extension */ original_record_type = zh->m_record_type; if (!zh->m_record_type) { - sprintf (ext_res, "%srecordType.%s", gprefix, ext); - zh->m_record_type = res_get (zh->res, ext_res); + sprintf(ext_res, "%srecordType.%s", gprefix, ext); + zh->m_record_type = res_get(zh->res, ext_res); } if (!zh->m_record_type) { check_log_limit(zh); if (zh->records_processed + zh->records_skipped < zh->m_file_verbose_limit) - yaz_log (YLOG_LOG, "? %s", fname); + yaz_log(YLOG_LOG, "? %s", fname); zh->records_skipped++; return 0; } /* determine match criteria */ if (!zh->m_record_id) { - sprintf (ext_res, "%srecordId.%s", gprefix, ext); - zh->m_record_id = res_get (zh->res, ext_res); + sprintf(ext_res, "%srecordId.%s", gprefix, ext); + zh->m_record_id = res_get(zh->res, ext_res); } if (!(recType = - recType_byName (zh->reg->recTypes, zh->res, zh->m_record_type, + recType_byName(zh->reg->recTypes, zh->res, zh->m_record_type, &recTypeClientData))) { yaz_log(YLOG_WARN, "No such record type: %s", zh->m_record_type); @@ -620,27 +687,26 @@ 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; } else { char full_rep[1024]; - if (zh->path_reg && !yaz_is_abspath (fname)) + if (zh->path_reg && !yaz_is_abspath(fname)) { - strcpy (full_rep, zh->path_reg); - strcat (full_rep, "/"); - strcat (full_rep, fname); + strcpy(full_rep, zh->path_reg); + strcat(full_rep, "/"); + strcat(full_rep, fname); } else - strcpy (full_rep, fname); - - if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1) + strcpy(full_rep, fname); + + if ((fd = open(full_rep, O_BINARY|O_RDONLY)) == -1) { - yaz_log (YLOG_WARN|YLOG_ERRNO, "open %s", full_rep); + yaz_log(YLOG_WARN|YLOG_ERRNO, "open %s", full_rep); zh->m_record_type = original_record_type; return ZEBRA_FAIL; } @@ -648,9 +714,7 @@ 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, - 0, /* tst_mode */ + action, zh->m_record_type, sysno, 0, /*match_criteria */ @@ -666,13 +730,12 @@ 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, - int test_mode, const char *recordType, zint *sysno, const char *match_criteria, @@ -687,26 +750,26 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, { yaz_log(log_level_extract, "Record type explicitly specified: %s", recordType); - recType = recType_byName (zh->reg->recTypes, zh->res, recordType, + recType = recType_byName(zh->reg->recTypes, zh->res, recordType, &clientData); - } + } else { if (!(zh->m_record_type)) { - yaz_log (YLOG_WARN, "No such record type defined"); + yaz_log(YLOG_WARN, "No such record type defined"); return ZEBRA_FAIL; } yaz_log(log_level_extract, "Get record type from rgroup: %s", zh->m_record_type); - recType = recType_byName (zh->reg->recTypes, zh->res, + recType = recType_byName(zh->reg->recTypes, zh->res, zh->m_record_type, &clientData); recordType = zh->m_record_type; } - + if (!recType) { - yaz_log (YLOG_WARN, "No such record type: %s", recordType); + yaz_log(YLOG_WARN, "No such record type: %s", recordType); return ZEBRA_FAIL; } @@ -714,7 +777,6 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, res = zebra_extract_records_stream(zh, &stream, action, - test_mode, recordType, sysno, match_criteria, @@ -724,68 +786,16 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, return res; } -ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, - struct ZebraRecStream *stream, - enum zebra_recctrl_action_t action, - int test_mode, - const char *recordType, - zint *sysno, - const char *match_criteria, - const char *fname, - RecType recType, - void *recTypeClientData) -{ - ZEBRA_RES res = ZEBRA_OK; - while (1) - { - int more = 0; - res = zebra_extract_record_stream(zh, stream, - action, - test_mode, - recordType, - sysno, - match_criteria, - fname, - recType, recTypeClientData, &more); - if (!more) - { - res = ZEBRA_OK; - break; - } - if (res != ZEBRA_OK) - break; - if (sysno) - break; - } - return res; -} - - -static WRBUF wrbuf_hex_str(const char *cstr) -{ - size_t i; - WRBUF w = wrbuf_alloc(); - for (i = 0; cstr[i]; i++) - { - if (cstr[i] < ' ' || cstr[i] > 126) - wrbuf_printf(w, "\\%02X", cstr[i] & 0xff); - else - wrbuf_putc(w, cstr[i]); - } - return w; -} - -ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, - struct ZebraRecStream *stream, - enum zebra_recctrl_action_t action, - int test_mode, - const char *recordType, - zint *sysno, - const char *match_criteria, - const char *fname, - RecType recType, - void *recTypeClientData, - int *more) +static ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, + struct ZebraRecStream *stream, + enum zebra_recctrl_action_t action, + const char *recordType, + zint *sysno, + const char *match_criteria, + const char *fname, + RecType recType, + void *recTypeClientData, + int *more) { zint sysno0 = 0; @@ -796,7 +806,7 @@ 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(); @@ -807,9 +817,9 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, zebra_rec_keys_reset(zh->reg->keys); zebra_rec_keys_reset(zh->reg->sortKeys); - if (zebraExplain_curDatabase (zh->reg->zei, zh->basenames[0])) + 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; } @@ -822,7 +832,7 @@ 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; @@ -837,32 +847,32 @@ 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: return ZEBRA_FAIL; case RECCTRL_EXTRACT_ERROR_GENERIC: /* error occured during extraction ... */ - yaz_log (YLOG_WARN, "extract error: generic"); + yaz_log(YLOG_WARN, "extract error: generic"); return ZEBRA_FAIL; case RECCTRL_EXTRACT_ERROR_NO_SUCH_FILTER: /* error occured during extraction ... */ - yaz_log (YLOG_WARN, "extract error: no such filter"); + yaz_log(YLOG_WARN, "extract error: no such filter"); return ZEBRA_FAIL; case RECCTRL_EXTRACT_SKIP: if (show_progress) - yaz_log (YLOG_LOG, "skip %s %s " ZINT_FORMAT, + 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); @@ -871,7 +881,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, case RECCTRL_EXTRACT_OK: break; default: - yaz_log (YLOG_WARN, "extract error: unknown error: %d", r); + yaz_log(YLOG_WARN, "extract error: unknown error: %d", r); return ZEBRA_FAIL; } end_offset = stream->endf(stream, 0); @@ -880,38 +890,58 @@ 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; } *more = 1; + + if (zh->m_flag_rw == 0) + { + yaz_log(YLOG_LOG, "test %s %s " ZINT_FORMAT, recordType, + pr_fname, (zint) start_offset); + /* test mode .. Do not perform match */ + return ZEBRA_OK; + } + if (!sysno) { sysno = &sysno0; - if (match_criteria && *match_criteria) { + 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, + else + { + if (zh->m_record_id && *zh->m_record_id) + { + matchStr = get_match_from_spec(zh, zh->reg->keys, pr_fname, zh->m_record_id); if (!matchStr) { - yaz_log (YLOG_LOG, "error %s %s " ZINT_FORMAT, recordType, + yaz_log(YLOG_LOG, "error %s %s " ZINT_FORMAT, recordType, pr_fname, (zint) start_offset); return ZEBRA_FAIL; } + if (0 && matchStr) + { + WRBUF w = wrbuf_alloc(); + size_t i; + for (i = 0; i < strlen(matchStr); i++) + { + wrbuf_printf(w, "%02X", matchStr[i] & 0xff); + } + yaz_log(YLOG_LOG, "Got match %s", wrbuf_cstr(w)); + wrbuf_destroy(w); + } } } - 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); @@ -921,43 +951,52 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, if (rinfo) { assert(*rinfo == sizeof(*sysno)); - memcpy (sysno, rinfo+1, sizeof(*sysno)); + memcpy(sysno, rinfo+1, sizeof(*sysno)); } } } - if (zebra_rec_keys_empty(zh->reg->keys)) - { - /* the extraction process returned no information - the record - is probably empty - unless flagShowRecords is in use */ - if (test_mode) - return ZEBRA_OK; - } 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_WARN, "cannot delete record above (seems new)"); + 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, "adelete %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, + yaz_log(YLOG_LOG, "update %s %s " ZINT_FORMAT, recordType, pr_fname, (zint) start_offset); - yaz_log (YLOG_WARN, "cannot update record above (seems new)"); + yaz_log(YLOG_WARN, "cannot update record above (seems new)"); return ZEBRA_FAIL; } if (show_progress) - yaz_log (YLOG_LOG, "add %s %s " ZINT_FORMAT, recordType, pr_fname, + yaz_log(YLOG_LOG, "add %s %s " ZINT_FORMAT, recordType, pr_fname, (zint) start_offset); - rec = rec_new (zh->reg->records); + rec = rec_new(zh->reg->records); *sysno = rec->sysno; - recordAttr = rec_init_attr (zh->reg->zei, 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); if (extractCtrl.staticrank < 0) { yaz_log(YLOG_WARN, "Negative staticrank for record. Set to 0"); @@ -982,7 +1021,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, #endif recordAttr->staticrank = extractCtrl.staticrank; zh->records_inserted++; - } + } else { /* record already exists */ @@ -990,19 +1029,26 @@ 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; } - rec = rec_get (zh->reg->records, *sysno); - assert (rec); - - recordAttr = rec_init_attr (zh->reg->zei, rec); + 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); /* decrease total size */ - zebraExplain_recordBytesIncrement (zh->reg->zei, + zebraExplain_recordBytesIncrement(zh->reg->zei, - recordAttr->recordSize); zebra_rec_keys_set_buf(delkeys, @@ -1019,13 +1065,13 @@ 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 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, @@ -1044,7 +1090,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, int db_ord = zebraExplain_get_database_ord(zh->reg->zei); dict_delete_ord(zh->reg->matchDict, db_ord, matchStr); } - rec_del (zh->reg->records, &rec); + rec_del(zh->reg->records, &rec); } zebra_rec_keys_close(delkeys); zebra_rec_keys_close(sortKeys); @@ -1064,7 +1110,7 @@ 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; @@ -1074,17 +1120,17 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, zebra_rec_keys_close(sortKeys); } /* update file type */ - xfree (rec->info[recInfo_fileType]); + xfree(rec->info[recInfo_fileType]); rec->info[recInfo_fileType] = - rec_strdup (recordType, &rec->size[recInfo_fileType]); + rec_strdup(recordType, &rec->size[recInfo_fileType]); /* update filename */ - xfree (rec->info[recInfo_filename]); + xfree(rec->info[recInfo_filename]); rec->info[recInfo_filename] = - rec_strdup (fname, &rec->size[recInfo_filename]); + rec_strdup(fname, &rec->size[recInfo_filename]); /* update delete keys */ - xfree (rec->info[recInfo_delKeys]); + xfree(rec->info[recInfo_delKeys]); if (!zebra_rec_keys_empty(zh->reg->keys) && zh->m_store_keys == 1) { zebra_rec_keys_get_buf(zh->reg->keys, @@ -1097,7 +1143,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, rec->size[recInfo_delKeys] = 0; } /* update sort keys */ - xfree (rec->info[recInfo_sortKeys]); + xfree(rec->info[recInfo_sortKeys]); zebra_rec_keys_get_buf(zh->reg->sortKeys, &rec->info[recInfo_sortKeys], @@ -1112,10 +1158,10 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, /* set run-number for this record */ recordAttr->runNumber = - zebraExplain_runNumberIncrement (zh->reg->zei, 0); + zebraExplain_runNumberIncrement(zh->reg->zei, 0); /* update store data */ - xfree (rec->info[recInfo_storeData]); + xfree(rec->info[recInfo_storeData]); /* update store data */ if (zh->store_data_buf) @@ -1131,7 +1177,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, rec->size[recInfo_storeData] = recordAttr->recordSize; rec->info[recInfo_storeData] = (char *) - xmalloc (recordAttr->recordSize); + xmalloc(recordAttr->recordSize); stream->seekf(stream, start_offset); stream->readf(stream, rec->info[recInfo_storeData], recordAttr->recordSize); @@ -1143,31 +1189,77 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, rec->size[recInfo_storeData] = 0; } /* update database name */ - xfree (rec->info[recInfo_databaseName]); + 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); + rec_put(zh->reg->records, &rec); logRecord(zh); return ZEBRA_OK; } +/** \brief extracts records from stream + \param zh Zebra Handle + \param stream stream that we read from + \param action (action_insert, action_replace, action_delete, ..) + \param recordType Record filter type "grs.xml", etc. + \param sysno pointer to sysno if already known; NULL otherwise + \param match_criteria (NULL if not already given) + \param fname filename that we read from (for logging purposes only) + \param recType record type + \param recTypeClientData client data for record type + \returns ZEBRA_OK for success; ZEBRA_FAIL for failure +*/ +ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, + struct ZebraRecStream *stream, + enum zebra_recctrl_action_t action, + const char *recordType, + zint *sysno, + const char *match_criteria, + const char *fname, + RecType recType, + void *recTypeClientData) +{ + ZEBRA_RES res = ZEBRA_OK; + while (1) + { + int more = 0; + res = zebra_extract_record_stream(zh, stream, + action, + recordType, + sysno, + match_criteria, + fname, + recType, recTypeClientData, &more); + if (!more) + { + res = ZEBRA_OK; + break; + } + if (res != ZEBRA_OK) + break; + if (sysno) + break; + } + return res; +} + ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n) { ZebraHandle zh = (ZebraHandle) handle; struct recExtractCtrl extractCtrl; - if (zebraExplain_curDatabase (zh->reg->zei, + if (zebraExplain_curDatabase(zh->reg->zei, rec->info[recInfo_databaseName])) { abort(); - if (zebraExplain_newDatabase (zh->reg->zei, + if (zebraExplain_newDatabase(zh->reg->zei, rec->info[recInfo_databaseName], 0)) - abort (); + abort(); } zebra_rec_keys_reset(zh->reg->keys); @@ -1187,7 +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) @@ -1196,14 +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); @@ -1223,23 +1315,82 @@ 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]); + + 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]); + xfree(rec->info[recInfo_sortKeys]); zebra_rec_keys_get_buf(zh->reg->sortKeys, &rec->info[recInfo_sortKeys], &rec->size[recInfo_sortKeys]); return ZEBRA_OK; } +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 */ + 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); + *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)) + strcpy(dst_buf, "firstinfield"); + else if (!strcmp(str, CHR_UNKNOWN)) + strcpy(dst_buf, "unknown"); + else if (!strcmp(str, CHR_SPACE)) + strcpy(dst_buf, "space"); + + for (i = 0; ireg->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; ireg->zei; int normal = 0; @@ -1367,8 +1476,8 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, 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")); + 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); } @@ -1376,14 +1485,14 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, { extract_rec_keys_adjust(zh, 1, ins_keys); if (!del_keys) - zebraExplain_recordCountIncrement (zei, 1); + zebraExplain_recordCountIncrement(zei, 1); zebra_rec_keys_rewind(ins_keys); } if (del_keys) { extract_rec_keys_adjust(zh, 0, del_keys); if (!ins_keys) - zebraExplain_recordCountIncrement (zei, -1); + zebraExplain_recordCountIncrement(zei, -1); zebra_rec_keys_rewind(del_keys); } @@ -1407,7 +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++; @@ -1415,23 +1524,24 @@ 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); } - -void extract_flush_record_keys(ZebraHandle zh, zint sysno, int cmd, - zebra_rec_keys_t reckeys, - zint staticrank) +#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; @@ -1446,12 +1556,12 @@ void extract_flush_record_keys(ZebraHandle zh, zint sysno, int cmd, 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")); + 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); + zebraExplain_recordCountIncrement(zei, cmd ? 1 : -1); #if 0 yaz_log(YLOG_LOG, "sysno=" ZINT_FORMAT " cmd=%d", sysno, cmd); @@ -1459,24 +1569,25 @@ void extract_flush_record_keys(ZebraHandle zh, zint sysno, int cmd, #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, + 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_snippets(ZebraHandle zh, +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; @@ -1486,12 +1597,12 @@ 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]; 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); @@ -1517,19 +1628,19 @@ 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); 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); } } } @@ -1570,9 +1681,10 @@ static void extract_add_sort_string(RecWord *p, const char *str, int length) ch = zebraExplain_lookup_attr_str(zei, cat, p->index_type, p->index_name); if (ch < 0) ch = zebraExplain_add_attr_str(zei, cat, p->index_type, p->index_name); - key.len = 2; + key.len = 3; key.mem[0] = ch; key.mem[1] = p->record_id; + key.mem[2] = p->section_id; zebra_rec_keys_write(zh->reg->sortKeys, str, length, &key); } @@ -1591,19 +1703,27 @@ 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); + assert(length > 0); if (!p->index_name) return; + if (log_level_details) + { - if (zebra_maps_is_index(zh->reg->zebra_maps, p->index_type)) + 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); + } + 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)); @@ -1613,26 +1733,25 @@ 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) { @@ -1644,8 +1763,7 @@ 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; } @@ -1660,7 +1778,7 @@ 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; } @@ -1668,102 +1786,87 @@ static void extract_add_incomplete_field(RecWord *p) return; 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; + int i = parse_complete_field(p, zm, buf); + if (!i) + return; + extract_add_string(p, zm, buf, i); + p->seqno++; +} - if (remain > 0) - map = zebra_maps_input (zh->reg->zebra_maps, p->index_type, &b, remain, 1); +static void extract_add_icu(RecWord *p, zebra_map_t zm) +{ + const char *res_buf = 0; + size_t res_len = 0; - while (remain > 0 && i < IT_MAX_WORD) + zebra_map_tokenize_start(zm, p->term_buf, p->term_len); + while (zebra_map_tokenize_next(zm, &res_buf, &res_len, 0, 0)) { - while (map && *map && **map == *CHR_SPACE) - { - remain = p->term_len - (b - p->term_buf); + 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++; + } +} - if (remain > 0) - { - int first = i ? 0 : 1; /* first position */ - map = zebra_maps_input(zh->reg->zebra_maps, p->index_type, &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; +/** \brief top-level indexing handler for recctrl system + \param p token data to be indexed - 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 (zh->reg->zebra_maps, p->index_type, &b, - remain, 0); - } - else - map = 0; - } - } - if (!i) - return; - extract_add_string (p, buf, i); -} + 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; - WRBUF wrbuf; + zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, p->index_type); if (log_level_details) { yaz_log(log_level_details, "extract_token_add " - "type=%c index=%s seqno=" ZINT_FORMAT " s=%.*s", - p->index_type, p->index_name, + "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 (zebra_maps_is_icu(zm)) { - p->term_buf = wrbuf_buf(wrbuf); - p->term_len = wrbuf_len(wrbuf); + 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, @@ -1794,7 +1897,7 @@ static void extract_set_store_data_prepare(struct recExtractCtrl *p) static void extract_schema_add(struct recExtractCtrl *p, Odr_oid *oid) { ZebraHandle zh = (ZebraHandle) p->handle; - zebraExplain_addSchema (zh->reg->zei, oid); + zebraExplain_addSchema(zh->reg->zei, oid); } void extract_flush_sort_keys(ZebraHandle zh, zint sysno, @@ -1813,24 +1916,68 @@ void extract_flush_sort_keys(ZebraHandle zh, zint sysno, const char *str; struct it_key key_in; - zebra_sort_sysno(si, sysno); + NMEM nmem = nmem_create(); + struct sort_add_ent { + int ord; + int cmd; + struct sort_add_ent *next; + WRBUF wrbuf; + zint sysno; + zint section_id; + }; + struct sort_add_ent *sort_ent_list = 0; while (zebra_rec_keys_read(reckeys, &str, &slen, &key_in)) { int ord = CAST_ZINT_TO_INT(key_in.mem[0]); - - zebra_sort_type(si, ord); - if (cmd == 1) - zebra_sort_add(si, str, slen); - else - zebra_sort_delete(si); + zint filter_sysno = key_in.mem[1]; + zint section_id = key_in.mem[2]; + + struct sort_add_ent **e = &sort_ent_list; + for (; *e; e = &(*e)->next) + if ((*e)->ord == ord && section_id == (*e)->section_id) + break; + if (!*e) + { + *e = nmem_malloc(nmem, sizeof(**e)); + (*e)->next = 0; + (*e)->wrbuf = wrbuf_alloc(); + (*e)->ord = ord; + (*e)->cmd = cmd; + (*e)->sysno = filter_sysno ? filter_sysno : sysno; + (*e)->section_id = section_id; + } + + wrbuf_write((*e)->wrbuf, str, slen); + wrbuf_putc((*e)->wrbuf, '\0'); } + if (sort_ent_list) + { + zint last_sysno = 0; + struct sort_add_ent *e = sort_ent_list; + for (; e; e = e->next) + { + if (last_sysno != e->sysno) + { + zebra_sort_sysno(si, e->sysno); + last_sysno = e->sysno; + } + zebra_sort_type(si, e->ord); + if (e->cmd == 1) + zebra_sort_add(si, e->section_id, e->wrbuf); + else + zebra_sort_delete(si, e->section_id); + wrbuf_destroy(e->wrbuf); + } + } + nmem_destroy(nmem); } } /* * Local variables: * c-basic-offset: 4 + * c-file-style: "Stroustrup" * indent-tabs-mode: nil * End: * vim: shiftwidth=4 tabstop=8 expandtab