X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fextract.c;h=dc951f389c543978130446eabdfe3fd4d39ffc54;hb=527dab66d9847bb7f8a931c558306a070064bf25;hp=66eaff1f8ee1bfe9cf9271a4a33d41399858eb12;hpb=6f7dfe3c3f09f7104c1ae7616c9d207edeab308d;p=idzebra-moved-to-github.git diff --git a/index/extract.c b/index/extract.c index 66eaff1..dc951f3 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,4 +1,4 @@ -/* $Id: extract.c,v 1.266 2007-10-30 19:17:15 adam Exp $ +/* $Id: extract.c,v 1.276 2007-12-20 11:21:29 adam Exp $ Copyright (C) 1995-2007 Index Data ApS @@ -67,14 +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_token_add2(RecWord *p); static void check_log_limit(ZebraHandle zh) { @@ -100,15 +96,6 @@ static void logRecord(ZebraHandle zh) static void init_extractCtrl(ZebraHandle zh, struct recExtractCtrl *ctrl) { - int i; - for (i = 0; i<256; i++) - { - zebra_map_t zm = zebra_map_get(zh->reg->zebra_maps, i); - if (zebra_maps_is_positioned(zm)) - ctrl->seqno[i] = 1; - else - ctrl->seqno[i] = 0; - } ctrl->flagShowRecords = !zh->m_flag_rw; } @@ -278,11 +265,31 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) } +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); + zebra_map_t zm = zebra_map_get(zh->reg->zebra_maps, p->index_type); if (zm && zebra_maps_is_index(zm)) { @@ -290,10 +297,15 @@ static void snippet_token_add(RecWord *p) int ch = zebraExplain_lookup_attr_str( zei, zinfo_index_category_index, p->index_type, p->index_name); - if (zebra_maps_is_complete(zm)) - snippet_add_complete_field(p, ch, zm); + if (zebra_maps_is_icu(zm)) + snippet_add_icu(p, ch, zm); else - snippet_add_incomplete_field(p, ch, zm); + { + if (zebra_maps_is_complete(zm)) + snippet_add_complete_field(p, ch, zm); + else + snippet_add_incomplete_field(p, ch, zm); + } } } @@ -533,13 +545,29 @@ struct recordLogInfo { struct recordGroup *rGroup; }; -static void all_matches_add(struct recExtractCtrl *ctrl) +/** \brief add the always-matches index entry and map to real record ID + \param ctrl record control + \param record_id custom record ID + \param sysno system record ID + + This function serves two purposes.. It adds the always matches + entry and makes a pointer from the custom record ID (if defined) + back to the system record ID (sysno) + See zebra_recid_to_sysno . + */ +static void all_matches_add(struct recExtractCtrl *ctrl, zint record_id, + zint sysno) { RecWord word; extract_init(ctrl, &word); + word.record_id = record_id; + /* we use the seqno as placeholder for a way to get back to + record database from _ALLRECORDS.. This is used if a custom + RECORD was defined */ + word.seqno = sysno; word.index_name = "_ALLRECORDS"; word.index_type = "w"; - word.seqno = 1; + extract_add_index_string(&word, zinfo_index_category_alwaysmatches, "", 0); } @@ -557,7 +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; @@ -627,7 +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; @@ -655,8 +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, @@ -833,14 +860,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, stream->endf(stream, &null_offset);; extractCtrl.init = extract_init; - if (zh->reg->index_types) - { - extractCtrl.tokenAdd = extract_token_add2; - } - else - { - extractCtrl.tokenAdd = extract_token_add; - } + extractCtrl.tokenAdd = extract_token_add; extractCtrl.schemaAdd = extract_schema_add; extractCtrl.dh = zh->reg->dh; extractCtrl.handle = zh; @@ -894,8 +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; } @@ -939,6 +957,7 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, } } } + if (zebra_rec_keys_empty(zh->reg->keys)) { /* the extraction process returned no information - the record @@ -949,14 +968,21 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, if (! *sysno) { - /* new record */ + /* new record AKA does not exist already */ if (action == action_delete) { - yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, - pr_fname, (zint) start_offset); + yaz_log(YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType, + pr_fname, (zint) start_offset); yaz_log(YLOG_WARN, "cannot delete record above (seems new)"); return ZEBRA_FAIL; } + else if (action == action_a_delete) + { + if (show_progress) + yaz_log(YLOG_LOG, "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, @@ -971,6 +997,15 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, *sysno = rec->sysno; + + if (stream) + { + all_matches_add(&extractCtrl, + zebra_rec_keys_get_custom_record_id(zh->reg->keys), + *sysno); + } + + recordAttr = rec_init_attr(zh->reg->zei, rec); if (extractCtrl.staticrank < 0) { @@ -1012,6 +1047,13 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, rec = rec_get(zh->reg->records, *sysno); assert(rec); + + if (stream) + { + all_matches_add(&extractCtrl, + zebra_rec_keys_get_custom_record_id(zh->reg->keys), + *sysno); + } recordAttr = rec_init_attr(zh->reg->zei, rec); @@ -1033,7 +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 @@ -1277,7 +1319,7 @@ void extract_rec_keys_log(ZebraHandle zh, int is_insert, zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, 0/* db */, &string_index); assert(index_type); - zebra_term_untrans_iconv(zh, nmem, *index_type, + zebra_term_untrans_iconv(zh, nmem, index_type, &dst_term, str); *keystr = '\0'; for (i = 0; ireg->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) { @@ -1510,7 +1511,7 @@ ZEBRA_RES zebra_rec_keys_to_snippets(ZebraHandle zh, zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, 0/* db */, 0 /* string_index */); assert(index_type); - zebra_term_untrans_iconv(zh, nmem, *index_type, + zebra_term_untrans_iconv(zh, nmem, index_type, &dst_term, str); zebra_snippets_append(snippets, seqno, 0, ord, dst_term); nmem_reset(nmem); @@ -1541,7 +1542,7 @@ void print_rec_keys(ZebraHandle zh, zebra_rec_keys_t reckeys) seqno = key.mem[key.len-1]; - zebra_term_untrans(zh, *index_type, dst_buf, str); + zebra_term_untrans(zh, index_type, dst_buf, str); yaz_log(YLOG_LOG, "ord=%d seqno=" ZINT_FORMAT " term=%s", ord, seqno, dst_buf); @@ -1753,62 +1754,26 @@ static void extract_add_complete_field(RecWord *p, zebra_map_t zm) extract_add_string(p, zm, buf, i); } -static void extract_token_add2_index(ZebraHandle zh, zebra_index_type_t type, - RecWord *p) +static void extract_add_icu(RecWord *p, zebra_map_t zm) { - struct it_key key; const char *res_buf = 0; size_t res_len = 0; - int r = zebra_index_type_tokenize(type, p->term_buf, p->term_len, - &res_buf, &res_len); - int cat = zinfo_index_category_index; - int ch = zebraExplain_lookup_attr_str(zh->reg->zei, cat, p->index_type, p->index_name); - if (ch < 0) - ch = zebraExplain_add_attr_str(zh->reg->zei, cat, p->index_type, p->index_name); - while (r) - { - int i = 0; - key.mem[i++] = ch; - key.mem[i++] = p->record_id; - key.mem[i++] = p->section_id; - - if (zh->m_segment_indexing) - key.mem[i++] = p->segment; - key.mem[i++] = p->seqno; - key.len = i; - yaz_log(YLOG_LOG, "keys_write %.*s", (int) res_len, res_buf); - zebra_rec_keys_write(zh->reg->keys, res_buf, res_len, &key); - + 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++; - r = zebra_index_type_tokenize(type, 0, 0, &res_buf, &res_len); } } -static void extract_token_add2(RecWord *p) -{ - ZebraHandle zh = p->extractCtrl->handle; - zebra_index_type_t type = zebra_index_type_get(zh->reg->index_types, p->index_type); - if (type) - { - if (zebra_index_type_is_index(type)) - { - extract_token_add2_index(zh, type, p); - } - else if (zebra_index_type_is_sort(type)) - { - ; - - } - } -} /** \brief top-level indexing handler for recctrl system \param p token data to be indexed Call sequence: - extract_token - zebra_add_{in}_complete + extract_token_add + extract_add_{in}_complete / extract_add_icu extract_add_string extract_add_index_string @@ -1821,7 +1786,7 @@ static void extract_token_add2(RecWord *p) 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); + zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, p->index_type); WRBUF wrbuf; if (log_level_details) @@ -1833,13 +1798,20 @@ static void extract_token_add(RecWord *p) } if ((wrbuf = zebra_replace(zm, 0, p->term_buf, p->term_len))) { - p->term_buf = wrbuf_buf(wrbuf); - p->term_len = wrbuf_len(wrbuf); + p->term_buf = wrbuf_buf(wrbuf); + p->term_len = wrbuf_len(wrbuf); + } + if (zebra_maps_is_icu(zm)) + { + extract_add_icu(p, zm); } - if (zebra_maps_is_complete(zm)) - extract_add_complete_field(p, zm); else - extract_add_incomplete_field(p, zm); + { + 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,