X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fextract.c;h=69b449ca5d13ce0d886831907cd148e4dc56be23;hb=6a0f9234f945bc4956e2bcef75f715661a9eba9a;hp=48dd978f7b8e3077caa25f10bb1075dc09b7fa6c;hpb=896b30853daabb6294afe8b0a2f74fa6d6e397d8;p=idzebra-moved-to-github.git diff --git a/index/extract.c b/index/extract.c index 48dd978..69b449c 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,8 +1,5 @@ -/* $Id: extract.c,v 1.271 2007-12-07 14:09:09 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) 1994-2009 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 @@ -67,6 +64,21 @@ static void zebra_init_log_level(void) } } +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); @@ -187,7 +199,7 @@ static void snippet_add_complete_field(RecWord *p, int ord, } if (!i) return; - if (last && start != last) + if (last && start != last && zebra_maps_is_index(zm)) zebra_snippets_appendn(h->snippets, p->seqno, 0, ord, start, last - start); } @@ -222,7 +234,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) } 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); @@ -256,7 +268,7 @@ static void snippet_add_incomplete_field(RecWord *p, int ord, zebra_map_t zm) 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; @@ -265,22 +277,48 @@ 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)) + { + 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(zh->reg->zebra_maps, p->index_type); - if (zm && zebra_maps_is_index(zm)) + 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(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); + } } } @@ -417,9 +455,6 @@ static char *get_match_from_spec(ZebraHandle zh, 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", @@ -428,6 +463,20 @@ static char *get_match_from_spec(ZebraHandle zh, } 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]) { @@ -511,6 +560,14 @@ static char *get_match_from_spec(ZebraHandle zh, 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; } @@ -547,10 +604,10 @@ static void all_matches_add(struct recExtractCtrl *ctrl, zint record_id, "", 0); } +/* 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, @@ -560,7 +617,7 @@ 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; @@ -630,7 +687,7 @@ 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; @@ -658,9 +715,7 @@ 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 */ @@ -682,7 +737,6 @@ ZEBRA_RES zebra_extract_file(ZebraHandle zh, zint *sysno, const char *fname, 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, @@ -724,7 +778,6 @@ ZEBRA_RES zebra_buffer_extract_record(ZebraHandle zh, res = zebra_extract_records_stream(zh, &stream, action, - test_mode, recordType, sysno, match_criteria, @@ -734,69 +787,17 @@ 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; RecordAttr *recordAttr; @@ -895,14 +896,25 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, } *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) { + 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) @@ -911,6 +923,17 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, 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) @@ -934,24 +957,23 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle zh, } } - 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_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, @@ -1044,7 +1066,7 @@ 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 @@ -1181,6 +1203,52 @@ ZEBRA_RES zebra_extract_record_stream(ZebraHandle 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; @@ -1456,7 +1524,7 @@ void extract_flush_record_keys2(ZebraHandle zh, zint sysno, } -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) { @@ -1555,9 +1623,10 @@ 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); } @@ -1583,7 +1652,15 @@ static void extract_add_string(RecWord *p, zebra_map_t zm, if (!p->index_name) return; + if (log_level_details) + { + 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, @@ -1725,32 +1802,13 @@ static void extract_add_complete_field(RecWord *p, zebra_map_t zm) 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; - ZebraHandle zh = p->extractCtrl->handle; - - 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); + zebra_map_tokenize_start(zm, p->term_buf, p->term_len); - while (zebra_map_tokenize_next(zm, &res_buf, &res_len)) + while (zebra_map_tokenize_next(zm, &res_buf, &res_len, 0, 0)) { - 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; - - zebra_rec_keys_write(zh->reg->keys, res_buf, res_len, &key); - + extract_add_string(p, zm, res_buf, res_len); p->seqno++; } } @@ -1760,8 +1818,8 @@ static void extract_add_icu(RecWord *p, zebra_map_t zm) \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 @@ -1849,24 +1907,68 @@ 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]); + 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; + } - zebra_sort_type(si, ord); - if (cmd == 1) - zebra_sort_add(si, str, slen); - else - zebra_sort_delete(si); + 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