X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fextract.c;h=be000c2dd8ae2b11732e94979cbc380bf5d5cc0a;hp=2a6829d9078556ca4c5b4cf5f03bd737e0e1c592;hb=4097bac1dce126e79ce1550fad8b2034a311ff80;hpb=4e67b823f05b521f193af096a40c0f64758c5fc1 diff --git a/index/extract.c b/index/extract.c index 2a6829d..be000c2 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,4 +1,4 @@ -/* $Id: extract.c,v 1.255 2007-04-07 22:26:27 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; @@ -61,13 +67,10 @@ 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); -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) { @@ -78,7 +81,7 @@ static void check_log_limit(ZebraHandle zh) } } -static void logRecord (ZebraHandle zh) +static void logRecord(ZebraHandle zh) { check_log_limit(zh); ++zh->records_processed; @@ -91,7 +94,13 @@ static void logRecord (ZebraHandle zh) } } -static void extract_add_index_string (RecWord *p, +static void init_extractCtrl(ZebraHandle zh, struct recExtractCtrl *ctrl) +{ + ctrl->flagShowRecords = !zh->m_flag_rw; +} + + +static void extract_add_index_string(RecWord *p, zinfo_index_category_t cat, const char *str, int length); @@ -101,13 +110,244 @@ 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; w->segment = 0; } +struct snip_rec_info { + ZebraHandle zh; + zebra_snippets *snippets; +}; + + +static void snippet_add_complete_field(RecWord *p, int ord, + zebra_map_t zm) +{ + 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 + 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++); + } + last = b; + remain = p->term_len - (b - p->term_buf); + if (remain > 0) + { + map = zebra_maps_input(zm, &b, remain, 0); + } + else + map = 0; + } + } + if (!i) + return; + if (last && start != last) + zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, + start, last - start); +} + +static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) +{ + struct snip_rec_info *h = p->extractCtrl->handle; + 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(zm, &b, remain, 0); + + while (map) + { + char buf[IT_MAX_WORD+1]; + int i, remain; + + /* Skip spaces */ + while (map && *map && **map == *CHR_SPACE) + { + remain = p->term_len - (b - p->term_buf); + last = b; + if (remain > 0) + map = zebra_maps_input(zm, &b, remain, 0); + else + map = 0; + } + if (!map) + break; + if (start != last) + { + 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(zm, &b, remain, 0); + else + map = 0; + } + if (!i) + return; + + if (first) + { + first = 0; + if (zebra_maps_is_first_in_field(zm)) + { + /* first in field marker */ + p->seqno++; + } + } + if (start != last) + zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, + start, last - start); + start = last; + p->seqno++; + } + +} + +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 (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_icu(zm)) + snippet_add_icu(p, ch, zm); + else + { + if (zebra_maps_is_complete(zm)) + snippet_add_complete_field(p, ch, zm); + else + snippet_add_incomplete_field(p, ch, zm); + } + } +} + +static void snippet_schema_add( + struct recExtractCtrl *p, Odr_oid *oid) +{ + +} + +void extract_snippet(ZebraHandle zh, zebra_snippets *sn, + struct ZebraRecStream *stream, + RecType rt, void *recTypeClientData) +{ + struct recExtractCtrl extractCtrl; + struct snip_rec_info info; + int r; + + extractCtrl.stream = stream; + extractCtrl.first_record = 1; + extractCtrl.init = extract_init; + extractCtrl.tokenAdd = snippet_token_add; + extractCtrl.schemaAdd = snippet_schema_add; + assert(zh->reg); + 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); + +} + static void searchRecordKey(ZebraHandle zh, zebra_rec_keys_t reckeys, const char *index_name, @@ -121,11 +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 ; @@ -203,11 +443,11 @@ static char *get_match_from_spec(ZebraHandle zh, attname_str[i] = '\0'; } - searchRecordKey (zh, reckeys, attname_str, ws, 32); + 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; } @@ -221,12 +461,12 @@ 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; } @@ -243,25 +483,25 @@ 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 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 == '\'') @@ -278,12 +518,12 @@ 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; } @@ -291,7 +531,7 @@ 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; } @@ -305,27 +545,30 @@ struct recordLogInfo { struct recordGroup *rGroup; }; -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 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); } @@ -342,7 +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; @@ -360,7 +603,7 @@ 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); @@ -371,34 +614,34 @@ 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); @@ -412,7 +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; @@ -421,18 +664,18 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, { 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); + strcpy(full_rep, fname); - if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1) + 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; } @@ -440,8 +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, @@ -479,26 +721,26 @@ 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; } @@ -553,6 +795,20 @@ ZEBRA_RES zebra_extract_records_stream(ZebraHandle zh, } +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, @@ -585,9 +841,9 @@ 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; } @@ -629,15 +885,15 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, 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; @@ -649,7 +905,7 @@ 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); @@ -658,8 +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; } @@ -677,7 +931,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, 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; } @@ -688,13 +942,22 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, 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); + yaz_log(log_level_extract, "matchStr: %s", wrbuf_cstr(w)); + wrbuf_destroy(w); + } 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 @@ -705,29 +968,45 @@ 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_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, "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, + 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"); @@ -760,19 +1039,26 @@ 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); + 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); + 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, @@ -789,7 +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 @@ -814,7 +1100,7 @@ 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); @@ -844,17 +1130,17 @@ 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, @@ -867,7 +1153,7 @@ 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], @@ -882,10 +1168,10 @@ 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) @@ -901,7 +1187,7 @@ 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); @@ -913,15 +1199,15 @@ 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; } @@ -931,13 +1217,13 @@ 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); @@ -998,12 +1284,12 @@ ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n) } 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->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]); @@ -1026,7 +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, @@ -1060,12 +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); @@ -1137,8 +1424,8 @@ 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); } @@ -1146,14 +1433,14 @@ 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); } @@ -1196,52 +1483,11 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, &ins_key_in, 1, ins_str, ins_slen, ins_rank, zh->m_staticrank); } - yaz_log(YLOG_LOG, "normal=%d optimized=%d", normal, optimized); + 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) { @@ -1256,7 +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]; @@ -1267,7 +1513,7 @@ ZEBRA_RES zebra_rec_keys_to_snippets(ZebraHandle zh, assert(index_type); zebra_term_untrans_iconv(zh, nmem, index_type, &dst_term, str); - zebra_snippets_append(snippets, seqno, ord, dst_term); + zebra_snippets_append(snippets, seqno, 0, ord, dst_term); nmem_reset(nmem); } } @@ -1287,7 +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); @@ -1361,19 +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); + 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)); @@ -1383,26 +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) { @@ -1414,8 +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; } @@ -1430,7 +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; } @@ -1440,28 +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) { @@ -1472,7 +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; @@ -1500,8 +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; @@ -1509,31 +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_extract) + if (log_level_details) { - yaz_log(log_level_extract, "extract_token_add " - "type=%c index=%s seqno=" ZINT_FORMAT " s=%.*s", + yaz_log(log_level_details, "extract_token_add " + "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); + } + 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, @@ -1564,7 +1842,7 @@ 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,