X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fextract.c;h=e7b79aaa802100ac9d0121fa4507a62c2f5ec5b3;hp=9561c6e330d02e6a713f42e84d2112b1720e71ec;hb=3e7372e15b55113d4d29582fc30e7e3a398161fc;hpb=2e4e9c6def27f1e1463dcb6f205fab6a98054f38 diff --git a/index/extract.c b/index/extract.c index 9561c6e..e7b79aa 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,6 +1,6 @@ -/* $Id: extract.c,v 1.158 2004-08-04 08:35:23 adam Exp $ - Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004 - Index Data Aps +/* $Id: extract.c,v 1.191 2005-08-19 11:03:53 adam Exp $ + Copyright (C) 1995-2005 + Index Data ApS This file is part of the Zebra server. @@ -25,7 +25,8 @@ Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA #include #ifdef WIN32 #include -#else +#endif +#if HAVE_UNISTD_H #include #endif #include @@ -72,60 +73,70 @@ static void logRecord (ZebraHandle zh) ++zh->records_processed; if (!(zh->records_processed % 1000)) { - logf (LOG_LOG, "Records: %7d i/u/d %d/%d/%d", + yaz_log (YLOG_LOG, "Records: "ZINT_FORMAT" i/u/d " + ZINT_FORMAT"/"ZINT_FORMAT"/"ZINT_FORMAT, zh->records_processed, zh->records_inserted, zh->records_updated, zh->records_deleted); } } +static void extract_set_store_data_prepare(struct recExtractCtrl *p); + static void extract_init (struct recExtractCtrl *p, RecWord *w) { w->zebra_maps = p->zebra_maps; w->seqno = 1; +#if NATTR +#else w->attrSet = VAL_BIB1; w->attrUse = 1016; - w->reg_type = 'w'; +#endif + w->index_name = 0; + w->index_type = 'w'; w->extractCtrl = p; + w->record_id = 0; + w->section_id = 0; } -static const char **searchRecordKey (ZebraHandle zh, - struct recKeys *reckeys, - int attrSetS, int attrUseS) +static void searchRecordKey(ZebraHandle zh, + const struct recKeys *reckeys, + int attrSetS, int attrUseS, + const char **ws, int ws_length) { -#if IT_KEY_NEW -/* #error searchRecordKey does not work yet in this mode.. */ - static const char *ws[32]; void *decode_handle = iscz1_start(); int off = 0; int startSeq = -1; int seqno = 0; int i; + int ch; - for (i = 0; i<32; i++) + for (i = 0; ireg->zei, + attrSetS, attrUseS); + if (ch < 0) + return ; + while (off < reckeys->buf_used) { const char *src = reckeys->buf + off; struct it_key key; char *dst = (char*) &key; - int attrSet, attrUse; iscz1_decode(decode_handle, &dst, &src); - assert(key.len < 4 && key.len > 2); + assert(key.len <= 4 && key.len > 2); - attrSet = key.mem[0]; - attrUse = key.mem[1]; - seqno = key.mem[2]; + seqno = (int) key.mem[key.len-1]; - if (attrUseS == attrUse && attrSetS == attrSet) + if (key.mem[0] == ch) { int woff; if (startSeq == -1) startSeq = seqno; woff = seqno - startSeq; - if (woff >= 0 && woff < 31) + if (woff >= 0 && woff < ws_length) ws[woff] = src; } @@ -135,87 +146,6 @@ static const char **searchRecordKey (ZebraHandle zh, } iscz1_stop(decode_handle); assert (off == reckeys->buf_used); - return ws; -#else - static const char *ws[32]; - int off = 0; - int startSeq = -1; - int i; - int seqno = 0; -#if SU_SCHEME - int chS, ch; -#else - short attrUse; - char attrSet; -#endif - - for (i = 0; i<32; i++) - ws[i] = NULL; - -#if SU_SCHEME - chS = zebraExplain_lookupSU (zh->reg->zei, attrSetS, attrUseS); - if (chS < 0) - return ws; -#endif - while (off < reckeys->buf_used) - { - - const char *src = reckeys->buf + off; - const char *wstart; - int lead; - - lead = *src++; -#if SU_SCHEME - if ((lead & 3)<3) - { - memcpy (&ch, src, sizeof(ch)); - src += sizeof(ch); - } -#else - if (!(lead & 1)) - { - memcpy (&attrSet, src, sizeof(attrSet)); - src += sizeof(attrSet); - } - if (!(lead & 2)) - { - memcpy (&attrUse, src, sizeof(attrUse)); - src += sizeof(attrUse); - } -#endif - wstart = src; - while (*src++) - ; - if (lead & 60) - seqno += ((lead>>2) & 15)-1; - else - { - memcpy (&seqno, src, sizeof(seqno)); - src += sizeof(seqno); - } - if ( -#if SU_SCHEME - ch == chS -#else - attrUseS == attrUse && attrSetS == attrSet -#endif - ) - { - int woff; - - - if (startSeq == -1) - startSeq = seqno; - woff = seqno - startSeq; - if (woff >= 0 && woff < 31) - ws[woff] = wstart; - } - - off = src - reckeys->buf; - } - assert (off == reckeys->buf_used); - return ws; -#endif } struct file_read_info { @@ -224,8 +154,6 @@ struct file_read_info { off_t file_moffset; /* offset of rec/rec boundary */ int file_more; int fd; - char *sdrbuf; - int sdrmax; }; static struct file_read_info *file_read_start (int fd) @@ -236,8 +164,8 @@ static struct file_read_info *file_read_start (int fd) fi->fd = fd; fi->file_max = 0; fi->file_moffset = 0; - fi->sdrbuf = 0; - fi->sdrmax = 0; + fi->file_offset = 0; + fi->file_more = 0; return fi; } @@ -250,8 +178,6 @@ static off_t file_seek (void *handle, off_t offset) { struct file_read_info *p = (struct file_read_info *) handle; p->file_offset = offset; - if (p->sdrbuf) - return offset; return lseek (p->fd, offset, SEEK_SET); } @@ -266,16 +192,7 @@ static int file_read (void *handle, char *buf, size_t count) struct file_read_info *p = (struct file_read_info *) handle; int fd = p->fd; int r; - if (p->sdrbuf) - { - r = count; - if (r > p->sdrmax - p->file_offset) - r = p->sdrmax - p->file_offset; - if (r) - memcpy (buf, p->sdrbuf + p->file_offset, r); - } - else - r = read (fd, buf, count); + r = read (fd, buf, count); if (r > 0) { p->file_offset += r; @@ -285,23 +202,15 @@ static int file_read (void *handle, char *buf, size_t count) return r; } -static void file_begin (void *handle) -{ - struct file_read_info *p = (struct file_read_info *) handle; - - p->file_offset = p->file_moffset; - if (!p->sdrbuf && p->file_moffset) - lseek (p->fd, p->file_moffset, SEEK_SET); - p->file_more = 0; -} - static void file_end (void *handle, off_t offset) { struct file_read_info *p = (struct file_read_info *) handle; - assert (p->file_more == 0); - p->file_more = 1; - p->file_moffset = offset; + if (offset != p->file_moffset) + { + p->file_moffset = offset; + p->file_more = 1; + } } static char *fileMatchStr (ZebraHandle zh, @@ -311,7 +220,6 @@ static char *fileMatchStr (ZebraHandle zh, static char dstBuf[2048]; /* static here ??? */ char *dst = dstBuf; const char *s = spec; - static const char **w; while (1) { @@ -321,6 +229,7 @@ static char *fileMatchStr (ZebraHandle zh, break; if (*s == '(') { + const char *ws[32]; char attset_str[64], attname_str[64]; data1_attset *attset; int i; @@ -353,8 +262,7 @@ static char *fileMatchStr (ZebraHandle zh, else attUse = atoi (attname_str); } - w = searchRecordKey (zh, reckeys, attSet, attUse); - assert (w); + searchRecordKey (zh, reckeys, attSet, attUse, ws, 32); if (*s == ')') { @@ -363,26 +271,26 @@ static char *fileMatchStr (ZebraHandle zh, } else { - logf (LOG_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++; for (i = 0; i<32; i++) - if (matchFlag[i] && w[i]) + if (matchFlag[i] && ws[i]) { if (first) { *dst++ = ' '; first = 0; } - strcpy (dst, w[i]); - dst += strlen(w[i]); + strcpy (dst, ws[i]); + dst += strlen(ws[i]); } if (first) { - logf (LOG_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; } @@ -439,7 +347,7 @@ static char *fileMatchStr (ZebraHandle zh, } else { - logf (LOG_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; } @@ -447,7 +355,7 @@ static char *fileMatchStr (ZebraHandle zh, } if (dst == dstBuf) { - logf (LOG_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; } @@ -464,39 +372,39 @@ struct recordLogInfo { void create_rec_keys_codec(struct recKeys *keys) { keys->buf_used = 0; -#if IT_KEY_NEW iscz1_reset(keys->codec_handle); -#else - keys->prevAttrUse = -1; - keys->prevAttrSet = -1; - keys->prevSeqNo = 0; -#endif } - + +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->zebra_maps = zh->reg->zebra_maps; + ctrl->flagShowRecords = !zh->m_flag_rw; +} + static int file_extract_record(ZebraHandle zh, SYSNO *sysno, const char *fname, int deleteFlag, struct file_read_info *fi, - int force_update) + int force_update, + RecType recType, + void *recTypeClientData) { RecordAttr *recordAttr; int r; - const char *matchStr; + const char *matchStr = 0; SYSNO sysnotmp; Record rec; off_t recordOffset = 0; - RecType recType; - void *clientData; - char subType[128]; + struct recExtractCtrl extractCtrl; - if (!(recType = - recType_byName (zh->reg->recTypes, zh->m_record_type, subType, - &clientData))) - { - logf (LOG_WARN, "No such record type: %s", zh->m_record_type); - return 0; - } - /* announce database */ if (zebraExplain_curDatabase (zh->reg->zei, zh->basenames[0])) { @@ -507,37 +415,33 @@ static int file_extract_record(ZebraHandle zh, if (fi->fd != -1) { - struct recExtractCtrl extractCtrl; - /* we are going to read from a file, so prepare the extraction */ - int i; - create_rec_keys_codec(&zh->reg->keys); - +#if NATTR + create_rec_keys_codec(&zh->reg->sortKeys); +#else zh->reg->sortKeys.buf_used = 0; - +#endif recordOffset = fi->file_moffset; + extractCtrl.handle = zh; extractCtrl.offset = fi->file_moffset; extractCtrl.readf = file_read; extractCtrl.seekf = file_seek; extractCtrl.tellf = file_tell; extractCtrl.endf = file_end; extractCtrl.fh = fi; - extractCtrl.subType = subType; extractCtrl.init = extract_init; extractCtrl.tokenAdd = extract_token_add; extractCtrl.schemaAdd = extract_schema_add; extractCtrl.dh = zh->reg->dh; - extractCtrl.handle = zh; - for (i = 0; i<256; i++) - { - if (zebra_maps_is_positioned(zh->reg->zebra_maps, i)) - extractCtrl.seqno[i] = 1; - else - extractCtrl.seqno[i] = 0; - } - extractCtrl.zebra_maps = zh->reg->zebra_maps; - extractCtrl.flagShowRecords = !zh->m_flag_rw; + extractCtrl.match_criteria[0] = '\0'; + extractCtrl.staticrank = 0; + + extractCtrl.first_record = fi->file_offset ? 0 : 1; + + extract_set_store_data_prepare(&extractCtrl); + + init_extractCtrl(zh, &extractCtrl); if (!zh->m_flag_rw) printf ("File: %s " PRINTF_OFF_T "\n", fname, recordOffset); @@ -548,7 +452,7 @@ static int file_extract_record(ZebraHandle zh, yaz_log_init_prefix2 (msg); } - r = (*recType->extract)(clientData, &extractCtrl); + r = (*recType->extract)(recTypeClientData, &extractCtrl); yaz_log_init_prefix2 (0); if (r == RECCTRL_EXTRACT_EOF) @@ -559,7 +463,7 @@ static int file_extract_record(ZebraHandle zh, if (zh->m_flag_rw && zh->records_processed < zh->m_file_verbose_limit) { - logf (LOG_WARN, "fail %s %s " PRINTF_OFF_T, zh->m_record_type, + yaz_log (YLOG_WARN, "fail %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); } return 0; @@ -570,53 +474,54 @@ static int file_extract_record(ZebraHandle zh, if (zh->m_flag_rw && zh->records_processed < zh->m_file_verbose_limit) { - logf (LOG_WARN, "no filter for %s %s " + yaz_log (YLOG_WARN, "no filter for %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); } return 0; } - if (zh->reg->keys.buf_used == 0) - { - /* the extraction process returned no information - the record - is probably empty - unless flagShowRecords is in use */ - if (!zh->m_flag_rw) - return 1; - - logf (LOG_WARN, "empty %s %s " PRINTF_OFF_T, zh->m_record_type, - fname, recordOffset); - return 1; - } + if (extractCtrl.match_criteria[0]) + matchStr = extractCtrl.match_criteria; } /* perform match if sysno not known and if match criteria is specified */ - - matchStr = NULL; if (!sysno) { sysnotmp = 0; sysno = &sysnotmp; - if (zh->m_record_id && *zh->m_record_id) + + if (matchStr == 0 && zh->m_record_id && *zh->m_record_id) { - char *rinfo; matchStr = fileMatchStr (zh, &zh->reg->keys, fname, zh->m_record_id); - if (matchStr) - { - rinfo = dict_lookup (zh->reg->matchDict, matchStr); - if (rinfo) - { - assert(*rinfo == sizeof(*sysno)); - memcpy (sysno, rinfo+1, sizeof(*sysno)); - } - } - else - { - logf (LOG_WARN, "Bad match criteria"); - return 0; - } - } + if (!matchStr) + { + yaz_log(YLOG_WARN, "Bad match criteria"); + return 0; + } + } + if (matchStr) + { + char *rinfo = dict_lookup (zh->reg->matchDict, matchStr); + if (rinfo) + { + assert(*rinfo == sizeof(*sysno)); + memcpy (sysno, rinfo+1, sizeof(*sysno)); + } + } + } + if (! *sysno && zh->reg->keys.buf_used == 0) + { + /* the extraction process returned no information - the record + is probably empty - unless flagShowRecords is in use */ + if (!zh->m_flag_rw) + return 1; + + if (zh->records_processed < zh->m_file_verbose_limit) + yaz_log (YLOG_WARN, "empty %s %s " PRINTF_OFF_T, zh->m_record_type, + fname, recordOffset); + return 1; } if (! *sysno) @@ -624,26 +529,28 @@ static int file_extract_record(ZebraHandle zh, /* new record */ if (deleteFlag) { - logf (LOG_LOG, "delete %s %s " PRINTF_OFF_T, zh->m_record_type, + yaz_log (YLOG_LOG, "delete %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); - logf (LOG_WARN, "cannot delete record above (seems new)"); + yaz_log (YLOG_WARN, "cannot delete record above (seems new)"); return 1; } if (zh->records_processed < zh->m_file_verbose_limit) - logf (LOG_LOG, "add %s %s " PRINTF_OFF_T, zh->m_record_type, + yaz_log (YLOG_LOG, "add %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); rec = rec_new (zh->reg->records); *sysno = rec->sysno; recordAttr = rec_init_attr (zh->reg->zei, rec); + recordAttr->staticrank = extractCtrl.staticrank; if (matchStr) { dict_insert (zh->reg->matchDict, matchStr, sizeof(*sysno), sysno); } extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys); - extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys); + extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys, + recordAttr->staticrank); zh->records_inserted++; } @@ -651,7 +558,11 @@ static int file_extract_record(ZebraHandle zh, { /* record already exists */ struct recKeys delkeys; +#if NATTR + struct recKeys sortKeys; +#else struct sortKeys sortKeys; +#endif rec = rec_get (zh->reg->records, *sysno); assert (rec); @@ -661,14 +572,16 @@ static int file_extract_record(ZebraHandle zh, if (!force_update && recordAttr->runNumber == zebraExplain_runNumberIncrement (zh->reg->zei, 0)) { - yaz_log (LOG_LOG, "run number = %d", recordAttr->runNumber); - yaz_log (LOG_LOG, "skipped %s %s " PRINTF_OFF_T, + yaz_log (YLOG_LOG, "run number = " ZINT_FORMAT, + recordAttr->runNumber); + yaz_log (YLOG_LOG, "skipped %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); extract_flushSortKeys (zh, *sysno, -1, &zh->reg->sortKeys); rec_rm (&rec); logRecord (zh); return 1; } + /* flush old keys for sort&search etc. */ delkeys.buf_used = rec->size[recInfo_delKeys]; delkeys.buf = rec->info[recInfo_delKeys]; @@ -676,20 +589,21 @@ static int file_extract_record(ZebraHandle zh, sortKeys.buf = rec->info[recInfo_sortKeys]; extract_flushSortKeys (zh, *sysno, 0, &sortKeys); - extract_flushRecordKeys (zh, *sysno, 0, &delkeys); + extract_flushRecordKeys (zh, *sysno, 0, &delkeys, + recordAttr->staticrank); /* old values */ if (deleteFlag) { /* record going to be deleted */ if (!delkeys.buf_used) { - logf (LOG_LOG, "delete %s %s " PRINTF_OFF_T, + yaz_log (YLOG_LOG, "delete %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); - logf (LOG_WARN, "cannot delete file above, storeKeys false"); + yaz_log (YLOG_WARN, "cannot delete file above, storeKeys false"); } else { if (zh->records_processed < zh->m_file_verbose_limit) - logf (LOG_LOG, "delete %s %s " PRINTF_OFF_T, + yaz_log (YLOG_LOG, "delete %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); zh->records_deleted++; if (matchStr) @@ -705,17 +619,20 @@ static int file_extract_record(ZebraHandle zh, /* record going to be updated */ if (!delkeys.buf_used) { - logf (LOG_LOG, "update %s %s " PRINTF_OFF_T, + yaz_log (YLOG_LOG, "update %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); - logf (LOG_WARN, "cannot update file above, storeKeys false"); + yaz_log (YLOG_WARN, "cannot update file above, storeKeys false"); } else { + /* flush new keys for sort&search etc */ if (zh->records_processed < zh->m_file_verbose_limit) - logf (LOG_LOG, "update %s %s " PRINTF_OFF_T, + yaz_log (YLOG_LOG, "update %s %s " PRINTF_OFF_T, zh->m_record_type, fname, recordOffset); + recordAttr->staticrank = extractCtrl.staticrank; extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys); - extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys); + extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys, + recordAttr->staticrank); zh->records_updated++; } } @@ -768,21 +685,27 @@ static int file_extract_record(ZebraHandle zh, /* update store data */ xfree (rec->info[recInfo_storeData]); - if (zh->m_store_data) + if (zh->store_data_buf) + { + rec->size[recInfo_storeData] = zh->store_data_size; + rec->info[recInfo_storeData] = zh->store_data_buf; + zh->store_data_buf = 0; + } + else if (zh->m_store_data) { rec->size[recInfo_storeData] = recordAttr->recordSize; rec->info[recInfo_storeData] = (char *) xmalloc (recordAttr->recordSize); if (lseek (fi->fd, recordOffset, SEEK_SET) < 0) { - logf (LOG_ERRNO|LOG_FATAL, "seek to " PRINTF_OFF_T " in %s", + yaz_log (YLOG_ERRNO|YLOG_FATAL, "seek to " PRINTF_OFF_T " in %s", recordOffset, fname); exit (1); } if (read (fi->fd, rec->info[recInfo_storeData], recordAttr->recordSize) < recordAttr->recordSize) { - logf (LOG_ERRNO|LOG_FATAL, "read %d bytes of %s", + yaz_log (YLOG_ERRNO|YLOG_FATAL, "read %d bytes of %s", recordAttr->recordSize, fname); exit (1); } @@ -815,13 +738,15 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, char ext_res[128]; struct file_read_info *fi; const char *original_record_type = 0; + RecType recType; + void *recTypeClientData; if (!zh->m_group || !*zh->m_group) *gprefix = '\0'; else sprintf (gprefix, "%s.", zh->m_group); - logf (LOG_DEBUG, "fileExtract %s", fname); + yaz_log (YLOG_DEBUG, "fileExtract %s", fname); /* determine file extension */ *ext = '\0'; @@ -843,7 +768,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, if (!zh->m_record_type) { if (zh->records_processed < zh->m_file_verbose_limit) - logf (LOG_LOG, "? %s", fname); + yaz_log (YLOG_LOG, "? %s", fname); return 0; } /* determine match criteria */ @@ -853,6 +778,21 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, zh->m_record_id = res_get (zh->res, ext_res); } + if (!(recType = + 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); + return 0; + } + + switch(recType->version) + { + case 0: + break; + default: + yaz_log(YLOG_WARN, "Bad filter version: %s", zh->m_record_type); + } if (sysno && deleteFlag) fd = -1; else @@ -871,7 +811,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1) { - logf (LOG_WARN|LOG_ERRNO, "open %s", full_rep); + yaz_log (YLOG_WARN|YLOG_ERRNO, "open %s", full_rep); zh->m_record_type = original_record_type; return 0; } @@ -879,9 +819,17 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, fi = file_read_start (fd); do { - file_begin (fi); - r = file_extract_record (zh, sysno, fname, deleteFlag, fi, 1); - } while (r && !sysno && fi->file_more); + fi->file_moffset = fi->file_offset; + fi->file_more = 0; /* file_end not called (yet) */ + r = file_extract_record (zh, sysno, fname, deleteFlag, fi, 1, + recType, recTypeClientData); + if (fi->file_more) + { /* file_end has been called so reset offset .. */ + fi->file_offset = fi->file_moffset; + lseek(fi->fd, fi->file_moffset, SEEK_SET); + } + } + while (r && !sysno); file_read_stop (fi); if (fd != -1) close (fd); @@ -895,28 +843,28 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, If not, and a record is provided, then sysno is got from there */ -int buffer_extract_record (ZebraHandle zh, - const char *buf, size_t buf_size, - int delete_flag, - int test_mode, - const char *recordType, - SYSNO *sysno, - const char *match_criteria, - const char *fname, - int force_update, - int allow_update) +ZEBRA_RES buffer_extract_record (ZebraHandle zh, + const char *buf, size_t buf_size, + int delete_flag, + int test_mode, + const char *recordType, + SYSNO *sysno, + const char *match_criteria, + const char *fname, + int force_update, + int allow_update) { RecordAttr *recordAttr; struct recExtractCtrl extractCtrl; - int i, r; + int r; const char *matchStr = 0; RecType recType = NULL; - char subType[1024]; void *clientData; Record rec; long recordOffset = 0; struct zebra_fetch_control fc; const char *pr_fname = fname; /* filename to print .. */ + int show_progress = zh->records_processed < zh->m_file_verbose_limit ? 1:0; if (!pr_fname) pr_fname = ""; /* make it printable if file is omitted */ @@ -933,84 +881,81 @@ int buffer_extract_record (ZebraHandle zh, extractCtrl.seekf = zebra_record_int_seek; extractCtrl.tellf = zebra_record_int_tell; extractCtrl.endf = zebra_record_int_end; + extractCtrl.first_record = 1; extractCtrl.fh = &fc; create_rec_keys_codec(&zh->reg->keys); - +#if NATTR + create_rec_keys_codec(&zh->reg->sortKeys); +#else zh->reg->sortKeys.buf_used = 0; - +#endif if (zebraExplain_curDatabase (zh->reg->zei, zh->basenames[0])) { if (zebraExplain_newDatabase (zh->reg->zei, zh->basenames[0], zh->m_explain_database)) - return 0; + return ZEBRA_FAIL; } - if (recordType && *recordType) { - logf (LOG_DEBUG, "Record type explicitly specified: %s", recordType); - recType = recType_byName (zh->reg->recTypes, recordType, subType, + if (recordType && *recordType) + { + yaz_log (YLOG_DEBUG, "Record type explicitly specified: %s", recordType); + recType = recType_byName (zh->reg->recTypes, zh->res, recordType, &clientData); - } else { - if (!(zh->m_record_type)) { - logf (LOG_WARN, "No such record type defined"); - return 0; + } + else + { + if (!(zh->m_record_type)) + { + yaz_log (YLOG_WARN, "No such record type defined"); + return ZEBRA_FAIL; } - logf (LOG_DEBUG, "Get record type from rgroup: %s",zh->m_record_type); - recType = recType_byName (zh->reg->recTypes, zh->m_record_type, subType, - &clientData); + yaz_log (YLOG_DEBUG, "Get record type from rgroup: %s",zh->m_record_type); + recType = recType_byName (zh->reg->recTypes, zh->res, + zh->m_record_type, &clientData); recordType = zh->m_record_type; } - if (!recType) { - logf (LOG_WARN, "No such record type: %s", zh->m_record_type); - return 0; + if (!recType) + { + yaz_log (YLOG_WARN, "No such record type: %s", zh->m_record_type); + return ZEBRA_FAIL; } - extractCtrl.subType = subType; extractCtrl.init = extract_init; extractCtrl.tokenAdd = extract_token_add; extractCtrl.schemaAdd = extract_schema_add; extractCtrl.dh = zh->reg->dh; extractCtrl.handle = zh; - extractCtrl.zebra_maps = zh->reg->zebra_maps; - extractCtrl.flagShowRecords = 0; - for (i = 0; i<256; i++) - { - if (zebra_maps_is_positioned(zh->reg->zebra_maps, i)) - extractCtrl.seqno[i] = 1; - else - extractCtrl.seqno[i] = 0; - } + extractCtrl.match_criteria[0] = '\0'; + extractCtrl.staticrank = 0; + + init_extractCtrl(zh, &extractCtrl); + + extract_set_store_data_prepare(&extractCtrl); r = (*recType->extract)(clientData, &extractCtrl); if (r == RECCTRL_EXTRACT_EOF) - return 0; + return ZEBRA_FAIL; else if (r == RECCTRL_EXTRACT_ERROR_GENERIC) { /* error occured during extraction ... */ - yaz_log (LOG_WARN, "extract error: generic"); - return 0; + yaz_log (YLOG_WARN, "extract error: generic"); + return ZEBRA_FAIL; } else if (r == RECCTRL_EXTRACT_ERROR_NO_SUCH_FILTER) { /* error occured during extraction ... */ - yaz_log (LOG_WARN, "extract error: no such filter"); - return 0; - } - if (zh->reg->keys.buf_used == 0) - { - /* the extraction process returned no information - the record - is probably empty - unless flagShowRecords is in use */ - if (test_mode) - return 1; - logf (LOG_WARN, "No keys generated for record"); - logf (LOG_WARN, " The file is probably empty"); - return 1; + yaz_log (YLOG_WARN, "extract error: no such filter"); + return ZEBRA_FAIL; } /* match criteria */ matchStr = NULL; + if (extractCtrl.match_criteria[0]) + match_criteria = extractCtrl.match_criteria; + if (! *sysno) { char *rinfo; if (match_criteria && *match_criteria) { @@ -1021,8 +966,8 @@ int buffer_extract_record (ZebraHandle zh, zh->m_record_id); if (!matchStr) { - logf (LOG_WARN, "Bad match criteria (recordID)"); - return 1; + yaz_log (YLOG_WARN, "Bad match criteria (recordID)"); + return ZEBRA_FAIL; } } } @@ -1035,24 +980,34 @@ int buffer_extract_record (ZebraHandle zh, } } } + if (zh->reg->keys.buf_used == 0) + { + /* 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 */ if (delete_flag) { - logf (LOG_LOG, "delete %s %s %ld", recordType, - pr_fname, (long) recordOffset); - logf (LOG_WARN, "cannot delete record above (seems new)"); - return 1; + if (show_progress) + yaz_log (YLOG_LOG, "delete %s %s %ld", recordType, + pr_fname, (long) recordOffset); + yaz_log (YLOG_WARN, "cannot delete record above (seems new)"); + return ZEBRA_FAIL; } - logf (LOG_LOG, "add %s %s %ld", recordType, pr_fname, - (long) recordOffset); + if (show_progress) + yaz_log (YLOG_LOG, "add %s %s %ld", recordType, pr_fname, + (long) recordOffset); rec = rec_new (zh->reg->records); *sysno = rec->sysno; recordAttr = rec_init_attr (zh->reg->zei, rec); + recordAttr->staticrank = extractCtrl.staticrank; if (matchStr) { @@ -1060,7 +1015,8 @@ int buffer_extract_record (ZebraHandle zh, sizeof(*sysno), sysno); } extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys); - extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys); + extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys, + recordAttr->staticrank); zh->records_inserted++; } @@ -1068,13 +1024,19 @@ int buffer_extract_record (ZebraHandle zh, { /* record already exists */ struct recKeys delkeys; +#if NATTR + struct recKeys sortKeys; +#else struct sortKeys sortKeys; +#endif - if (!allow_update) { - logf (LOG_LOG, "skipped %s %s %ld", - recordType, pr_fname, (long) recordOffset); - logRecord(zh); - return -1; + if (!allow_update) + { + if (show_progress) + yaz_log (YLOG_LOG, "skipped %s %s %ld", + recordType, pr_fname, (long) recordOffset); + logRecord(zh); + return ZEBRA_FAIL; } rec = rec_get (zh->reg->records, *sysno); @@ -1086,12 +1048,13 @@ int buffer_extract_record (ZebraHandle zh, if (recordAttr->runNumber == zebraExplain_runNumberIncrement (zh->reg->zei, 0)) { - logf (LOG_LOG, "skipped %s %s %ld", recordType, - pr_fname, (long) recordOffset); + if (show_progress) + yaz_log (YLOG_LOG, "skipped %s %s %ld", recordType, + pr_fname, (long) recordOffset); extract_flushSortKeys (zh, *sysno, -1, &zh->reg->sortKeys); rec_rm (&rec); logRecord(zh); - return -1; + return ZEBRA_FAIL; } } @@ -1102,20 +1065,26 @@ int buffer_extract_record (ZebraHandle zh, sortKeys.buf = rec->info[recInfo_sortKeys]; extract_flushSortKeys (zh, *sysno, 0, &sortKeys); - extract_flushRecordKeys (zh, *sysno, 0, &delkeys); + extract_flushRecordKeys (zh, *sysno, 0, &delkeys, + recordAttr->staticrank); if (delete_flag) { /* record going to be deleted */ if (!delkeys.buf_used) { - logf (LOG_LOG, "delete %s %s %ld", recordType, - pr_fname, (long) recordOffset); - logf (LOG_WARN, "cannot delete file above, storeKeys false"); - } + if (show_progress) + { + yaz_log (YLOG_LOG, "delete %s %s %ld", recordType, + pr_fname, (long) recordOffset); + yaz_log (YLOG_WARN, "cannot delete file above, " + "storeKeys false"); + } + } else { - logf (LOG_LOG, "delete %s %s %ld", recordType, - pr_fname, (long) recordOffset); + if (show_progress) + yaz_log (YLOG_LOG, "delete %s %s %ld", recordType, + pr_fname, (long) recordOffset); zh->records_deleted++; if (matchStr) dict_delete (zh->reg->matchDict, matchStr); @@ -1123,23 +1092,29 @@ int buffer_extract_record (ZebraHandle zh, } rec_rm (&rec); logRecord(zh); - return 0; + return ZEBRA_OK; } else { /* record going to be updated */ if (!delkeys.buf_used) { - logf (LOG_LOG, "update %s %s %ld", recordType, - pr_fname, (long) recordOffset); - logf (LOG_WARN, "cannot update file above, storeKeys false"); - } + if (show_progress) + { + yaz_log (YLOG_LOG, "update %s %s %ld", recordType, + pr_fname, (long) recordOffset); + yaz_log (YLOG_WARN, "cannot update file above, storeKeys false"); + } + } else { - logf (LOG_LOG, "update %s %s %ld", recordType, - pr_fname, (long) recordOffset); + if (show_progress) + yaz_log (YLOG_LOG, "update %s %s %ld", recordType, + pr_fname, (long) recordOffset); + recordAttr->staticrank = extractCtrl.staticrank; extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys); - extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys); + extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys, + recordAttr->staticrank); zh->records_updated++; } } @@ -1196,7 +1171,15 @@ int buffer_extract_record (ZebraHandle zh, /* update store data */ xfree (rec->info[recInfo_storeData]); - if (zh->m_store_data) + + /* update store data */ + if (zh->store_data_buf) + { + rec->size[recInfo_storeData] = zh->store_data_size; + rec->info[recInfo_storeData] = zh->store_data_buf; + zh->store_data_buf = 0; + } + else if (zh->m_store_data) { rec->size[recInfo_storeData] = recordAttr->recordSize; rec->info[recInfo_storeData] = (char *) @@ -1219,14 +1202,13 @@ int buffer_extract_record (ZebraHandle zh, /* commit this record */ rec_put (zh->reg->records, &rec); logRecord(zh); - return 0; + return ZEBRA_OK; } int explain_extract (void *handle, Record rec, data1_node *n) { ZebraHandle zh = (ZebraHandle) handle; struct recExtractCtrl extractCtrl; - int i; if (zebraExplain_curDatabase (zh->reg->zei, rec->info[recInfo_databaseName])) @@ -1238,18 +1220,25 @@ int explain_extract (void *handle, Record rec, data1_node *n) } create_rec_keys_codec(&zh->reg->keys); - +#if NATTR + create_rec_keys_codec(&zh->reg->sortKeys); +#else zh->reg->sortKeys.buf_used = 0; - +#endif extractCtrl.init = extract_init; extractCtrl.tokenAdd = extract_token_add; extractCtrl.schemaAdd = extract_schema_add; extractCtrl.dh = zh->reg->dh; - for (i = 0; i<256; i++) - extractCtrl.seqno[i] = 0; - extractCtrl.zebra_maps = zh->reg->zebra_maps; + + init_extractCtrl(zh, &extractCtrl); + extractCtrl.flagShowRecords = 0; + extractCtrl.match_criteria[0] = '\0'; + extractCtrl.staticrank = 0; extractCtrl.handle = handle; + extractCtrl.first_record = 1; + + extract_set_store_data_prepare(&extractCtrl); if (n) grs_extract_tree(&extractCtrl, n); @@ -1257,7 +1246,11 @@ int explain_extract (void *handle, Record rec, data1_node *n) if (rec->size[recInfo_delKeys]) { struct recKeys delkeys; +#if NATTR + struct recKeys sortkeys; +#else struct sortKeys sortkeys; +#endif delkeys.buf_used = rec->size[recInfo_delKeys]; delkeys.buf = rec->info[recInfo_delKeys]; @@ -1266,9 +1259,9 @@ int explain_extract (void *handle, Record rec, data1_node *n) sortkeys.buf = rec->info[recInfo_sortKeys]; extract_flushSortKeys (zh, rec->sysno, 0, &sortkeys); - extract_flushRecordKeys (zh, rec->sysno, 0, &delkeys); + extract_flushRecordKeys (zh, rec->sysno, 0, &delkeys, 0); } - extract_flushRecordKeys (zh, rec->sysno, 1, &zh->reg->keys); + extract_flushRecordKeys (zh, rec->sysno, 1, &zh->reg->keys, 0); extract_flushSortKeys (zh, rec->sysno, 1, &zh->reg->sortKeys); xfree (rec->info[recInfo_delKeys]); @@ -1287,18 +1280,10 @@ int explain_extract (void *handle, Record rec, data1_node *n) } void extract_flushRecordKeys (ZebraHandle zh, SYSNO sysno, - int cmd, struct recKeys *reckeys) + int cmd, struct recKeys *reckeys, + zint staticrank) { -#if IT_KEY_NEW void *decode_handle = iscz1_start(); -#else - int seqno = 0; -#if SU_SCHEME -#else - unsigned char attrSet = (unsigned char) -1; - unsigned short attrUse = (unsigned short) -1; -#endif -#endif int off = 0; int ch = 0; ZebraExplainInfo zei = zh->reg->zei; @@ -1308,7 +1293,7 @@ void extract_flushRecordKeys (ZebraHandle zh, SYSNO sysno, int mem= 1024*1024* atoi( res_get_def( zh->res, "memmax", "8")); if (mem <= 0) { - logf(LOG_WARN, "Invalid memory setting, using default 8 MB"); + yaz_log(YLOG_WARN, "Invalid memory setting, using default 8 MB"); mem= 1024*1024*8; } /* FIXME: That "8" should be in a default settings include */ @@ -1320,34 +1305,28 @@ void extract_flushRecordKeys (ZebraHandle zh, SYSNO sysno, zh->reg->key_file_no = 0; } zebraExplain_recordCountIncrement (zei, cmd ? 1 : -1); -#if IT_KEY_NEW + while (off < reckeys->buf_used) { const char *src = reckeys->buf + off; - struct it_key key; - char *dst = (char*) &key; - int attrSet, attrUse; + struct it_key key_in; + struct it_key key_out; + char *dst = (char*) &key_in; + zint *keyp = key_out.mem; iscz1_decode(decode_handle, &dst, &src); - assert(key.len < 4 && key.len > 2); - - attrSet = key.mem[0]; - attrUse = key.mem[1]; /* sequence in mem[2] */ + assert(key_in.len == 4); if (zh->reg->key_buf_used + 1024 > (zh->reg->ptr_top -zh->reg->ptr_i)*sizeof(char*)) extract_flushWriteKeys (zh,0); - assert(zh->reg->ptr_i >= 0); ++(zh->reg->ptr_i); assert(zh->reg->ptr_i > 0); (zh->reg->key_buf)[zh->reg->ptr_top - zh->reg->ptr_i] = (char*)zh->reg->key_buf + zh->reg->key_buf_used; - ch = zebraExplain_lookupSU (zei, attrSet, attrUse); - if (ch < 0) - ch = zebraExplain_addSU (zei, attrSet, attrUse); + ch = (int) key_in.mem[0]; /* ordinal for field/use/attribute */ - assert (ch > 0); zh->reg->key_buf_used += key_SU_encode (ch,((char*)zh->reg->key_buf) + zh->reg->key_buf_used); @@ -1357,85 +1336,28 @@ void extract_flushRecordKeys (ZebraHandle zh, SYSNO sysno, ((char*)(zh->reg->key_buf))[(zh->reg->key_buf_used)++] = '\0'; ((char*)(zh->reg->key_buf))[(zh->reg->key_buf_used)++] = cmd; - key.len = 2; - key.mem[0] = sysno; - key.mem[1] = key.mem[2]; /* sequence .. */ - - memcpy ((char*)zh->reg->key_buf + zh->reg->key_buf_used, - &key, sizeof(key)); - (zh->reg->key_buf_used) += sizeof(key); - off = src - reckeys->buf; - } -#else - while (off < reckeys->buf_used) - { - const char *src = reckeys->buf + off; - struct it_key key; - int lead; - - lead = *src++; - -#if SU_SCHEME - if ((lead & 3) < 3) + if (zh->m_staticrank) /* rank config enabled ? */ { - memcpy (&ch, src, sizeof(ch)); - src += sizeof(ch); + *keyp++ = staticrank; + key_out.len = 4; } -#else - if (!(lead & 1)) - { - memcpy (&attrSet, src, sizeof(attrSet)); - src += sizeof(attrSet); - } - if (!(lead & 2)) - { - memcpy (&attrUse, src, sizeof(attrUse)); - src += sizeof(attrUse); - } -#endif - if (zh->reg->key_buf_used + 1024 > - (zh->reg->ptr_top -zh->reg->ptr_i)*sizeof(char*)) - extract_flushWriteKeys (zh,0); - assert(zh->reg->ptr_i >= 0); - ++(zh->reg->ptr_i); - assert(zh->reg->ptr_i > 0); - (zh->reg->key_buf)[zh->reg->ptr_top - zh->reg->ptr_i] = - (char*)zh->reg->key_buf + zh->reg->key_buf_used; -#if SU_SCHEME -#else - ch = zebraExplain_lookupSU (zei, attrSet, attrUse); - if (ch < 0) - ch = zebraExplain_addSU (zei, attrSet, attrUse); -#endif - assert (ch > 0); - zh->reg->key_buf_used += - key_SU_encode (ch,((char*)zh->reg->key_buf) + - zh->reg->key_buf_used); + else + key_out.len = 3; - while (*src) - ((char*)zh->reg->key_buf) [(zh->reg->key_buf_used)++] = *src++; - src++; - ((char*)(zh->reg->key_buf))[(zh->reg->key_buf_used)++] = '\0'; - ((char*)(zh->reg->key_buf))[(zh->reg->key_buf_used)++] = cmd; + if (key_in.mem[1]) /* filter specified record ID */ + *keyp++ = key_in.mem[1]; + else + *keyp++ = sysno; + *keyp++ = key_in.mem[2]; /* section_id */ + *keyp++ = key_in.mem[3]; /* sequence .. */ - if (lead & 60) - seqno += ((lead>>2) & 15)-1; - else - { - memcpy (&seqno, src, sizeof(seqno)); - src += sizeof(seqno); - } - key.seqno = seqno; - key.sysno = sysno; - memcpy ((char*)zh->reg->key_buf + zh->reg->key_buf_used, &key, sizeof(key)); - (zh->reg->key_buf_used) += sizeof(key); + memcpy ((char*)zh->reg->key_buf + zh->reg->key_buf_used, + &key_out, sizeof(key_out)); + (zh->reg->key_buf_used) += sizeof(key_out); off = src - reckeys->buf; } -#endif assert (off == reckeys->buf_used); -#if IT_KEY_NEW iscz1_stop(decode_handle); -#endif } void extract_flushWriteKeys (ZebraHandle zh, int final) @@ -1454,19 +1376,19 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) #endif if (!zh->reg->key_buf || ptr_i <= 0) { - logf (LOG_DEBUG, " nothing to flush section=%d buf=%p i=%d", + yaz_log (YLOG_DEBUG, " nothing to flush section=%d buf=%p i=%d", zh->reg->key_file_no, zh->reg->key_buf, ptr_i); - logf (LOG_DEBUG, " buf=%p ", + yaz_log (YLOG_DEBUG, " buf=%p ", zh->reg->key_buf); - logf (LOG_DEBUG, " ptr=%d ",zh->reg->ptr_i); - logf (LOG_DEBUG, " reg=%p ",zh->reg); + yaz_log (YLOG_DEBUG, " ptr=%d ",zh->reg->ptr_i); + yaz_log (YLOG_DEBUG, " reg=%p ",zh->reg); return; } (zh->reg->key_file_no)++; - logf (LOG_LOG, "sorting section %d", (zh->reg->key_file_no)); - logf (LOG_DEBUG, " sort_buff at %p n=%d", + yaz_log (YLOG_LOG, "sorting section %d", (zh->reg->key_file_no)); + yaz_log (YLOG_DEBUG, " sort_buff at %p n=%d", zh->reg->key_buf + zh->reg->ptr_top - ptr_i,ptr_i); #if !SORT_EXTRA qsort (zh->reg->key_buf + zh->reg->ptr_top - ptr_i, ptr_i, @@ -1480,7 +1402,7 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) temp_policy=toupper(res_get_def(zh->res,"tempfiles","auto")[0]); if (temp_policy != 'Y' && temp_policy != 'N' && temp_policy != 'A') { - logf (LOG_WARN, "Illegal tempfiles setting '%c'. using 'Auto' ", + yaz_log (YLOG_WARN, "Illegal tempfiles setting '%c'. using 'Auto' ", temp_policy); temp_policy='A'; } @@ -1502,10 +1424,10 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) if (!(outf = fopen (out_fname, "wb"))) { - logf (LOG_FATAL|LOG_ERRNO, "fopen %s", out_fname); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fopen %s", out_fname); exit (1); } - logf (LOG_LOG, "writing section %d", zh->reg->key_file_no); + yaz_log (YLOG_LOG, "writing section %d", zh->reg->key_file_no); prevcp = cp = (zh->reg->key_buf)[zh->reg->ptr_top - ptr_i]; encode_key_init (&encode_info); @@ -1531,10 +1453,10 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) if (!(outf = fopen (out_fname, "wb"))) { - logf (LOG_FATAL|LOG_ERRNO, "fopen %s", out_fname); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fopen %s", out_fname); exit (1); } - logf (LOG_LOG, "writing section %d", key_file_no); + yaz_log (YLOG_LOG, "writing section %d", key_file_no); i = ptr_i; prevcp = key_buf[ptr_top-i]; while (1) @@ -1542,7 +1464,7 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) { key_y_len = strlen(prevcp)+1; #if 0 - logf (LOG_LOG, "key_y_len: %2d %02x %02x %s", + yaz_log (YLOG_LOG, "key_y_len: %2d %02x %02x %s", key_y_len, prevcp[0], prevcp[1], 2+prevcp); #endif qsort (key_buf + ptr_top-ptr_i, ptr_i - i, @@ -1564,36 +1486,25 @@ void extract_flushWriteKeys (ZebraHandle zh, int final) #endif if (fclose (outf)) { - logf (LOG_FATAL|LOG_ERRNO, "fclose %s", out_fname); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fclose %s", out_fname); exit (1); } - logf (LOG_LOG, "finished section %d", zh->reg->key_file_no); + yaz_log (YLOG_LOG, "finished section %d", zh->reg->key_file_no); zh->reg->ptr_i = 0; zh->reg->key_buf_used = 0; } -void extract_add_index_string (RecWord *p, const char *str, int length) +void extract_add_it_key (ZebraHandle zh, + struct recKeys *keys, + int reg_type, + const char *str, int slen, struct it_key *key) { char *dst; - ZebraHandle zh = p->extractCtrl->handle; - struct recKeys *keys = &zh->reg->keys; -#if IT_KEY_NEW - struct it_key key; - const char *src = (char*) &key; -#else - unsigned char attrSet; - unsigned short attrUse; - int lead = 0; - int diff = 0; - int *pseqno = &p->seqno; - ZebraExplainInfo zei = zh->reg->zei; -#endif + const char *src = (char*) key; if (keys->buf_used+1024 > keys->buf_max) { - char *b; - - b = (char *) xmalloc (keys->buf_max += 128000); + char *b = (char *) xmalloc (keys->buf_max += 128000); if (keys->buf_used > 0) memcpy (b, keys->buf, keys->buf_used); xfree (keys->buf); @@ -1601,92 +1512,169 @@ void extract_add_index_string (RecWord *p, const char *str, int length) } dst = keys->buf + keys->buf_used; -#if IT_KEY_NEW - key.len = 3; - key.mem[0] = p->attrSet; - key.mem[1] = p->attrUse; - key.mem[2] = p->seqno; + iscz1_encode(keys->codec_handle, &dst, &src); -#if 0 - /* just for debugging .. */ - yaz_log(LOG_LOG, "set=%d use=%d seqno=%d", p->attrSet, p->attrUse, - p->seqno); +#if REG_TYPE_PREFIX + *dst++ = reg_type; #endif + memcpy (dst, str, slen); + dst += slen; + *dst++ = '\0'; + keys->buf_used = dst - keys->buf; +} - iscz1_encode(keys->codec_handle, &dst, &src); +ZEBRA_RES zebra_snippets_rec_keys(ZebraHandle zh, struct recKeys *reckeys, + zebra_snippets *snippets) +{ + void *decode_handle = iscz1_start(); + int off = 0; + int seqno = 0; + NMEM nmem = nmem_create(); - *dst++ = p->reg_type; - memcpy (dst, str, length); - dst += length; - *dst++ = '\0'; -#else - /* leader byte is encoded as follows: - bit 0 : 1 if attrset is unchanged; 0 if attrset is changed - bit 1 : 1 if attruse is unchanged; 0 if attruse is changed - */ - attrSet = p->attrSet; - if (keys->buf_used > 0 && keys->prevAttrSet == attrSet) - lead |= 1; - else - keys->prevAttrSet = attrSet; - attrUse = p->attrUse; - if (keys->buf_used > 0 && keys->prevAttrUse == attrUse) - lead |= 2; - else - keys->prevAttrUse = attrUse; + yaz_log(YLOG_LOG, "zebra_rec_keys_snippets buf=%p sz=%d", reckeys->buf, + reckeys->buf_used); + assert(reckeys->buf); + while (off < reckeys->buf_used) + { + const char *src = reckeys->buf + off; + struct it_key key; + char *dst = (char*) &key; + char dst_buf[IT_MAX_WORD]; + char *dst_term = dst_buf; - diff = 1 + *pseqno - keys->prevSeqNo; - if (diff >= 1 && diff <= 15) - lead |= (diff << 2); - else - diff = 0; + iscz1_decode(decode_handle, &dst, &src); + assert(key.len <= 4 && key.len > 2); - keys->prevSeqNo = *pseqno; - - *dst++ = lead; + seqno = (int) key.mem[key.len-1]; -#if SU_SCHEME - if ((lead & 3) < 3) + zebra_term_untrans_iconv(zh, nmem, src[0], &dst_term, src+1); + zebra_snippets_append(snippets, seqno, src[0], key.mem[0], dst_term); + while (*src++) + ; + off = src - reckeys->buf; + nmem_reset(nmem); + } + nmem_destroy(nmem); + iscz1_stop(decode_handle); + return ZEBRA_OK; +} + +void print_rec_keys(ZebraHandle zh, struct recKeys *reckeys) +{ + void *decode_handle = iscz1_start(); + int off = 0; + int seqno = 0; + NMEM nmem = nmem_create(); + + yaz_log(YLOG_LOG, "print_rec_keys buf=%p sz=%d", reckeys->buf, + reckeys->buf_used); + assert(reckeys->buf); + while (off < reckeys->buf_used) { - int ch = zebraExplain_lookupSU (zei, attrSet, attrUse); - if (ch < 0) - { - ch = zebraExplain_addSU (zei, attrSet, attrUse); - yaz_log (LOG_DEBUG, "addSU set=%d use=%d SU=%d", - attrSet, attrUse, ch); - } - assert (ch > 0); - memcpy (dst, &ch, sizeof(ch)); - dst += sizeof(ch); + const char *src = reckeys->buf + off; + struct it_key key; + char *dst = (char*) &key; + int attrSet, attrUse; + char dst_buf[IT_MAX_WORD]; + char *dst_term = dst_buf; + + iscz1_decode(decode_handle, &dst, &src); + assert(key.len <= 4 && key.len > 2); + + attrSet = (int) key.mem[0] >> 16; + attrUse = (int) key.mem[0] & 65535; + seqno = (int) key.mem[key.len-1]; + + zebra_term_untrans_iconv(zh, nmem, src[0], &dst_term, src+1); + + yaz_log(YLOG_LOG, "ord=" ZINT_FORMAT " seqno=%d term=%s", + key.mem[0], seqno, dst_term); + while (*src++) + ; + off = src - reckeys->buf; + nmem_reset(nmem); } -#else - if (!(lead & 1)) + nmem_destroy(nmem); + iscz1_stop(decode_handle); +} + +void extract_add_index_string (RecWord *p, const char *str, int length) +{ + struct it_key key; + + ZebraHandle zh = p->extractCtrl->handle; + ZebraExplainInfo zei = zh->reg->zei; + int ch; + + if (p->index_name) { - memcpy (dst, &attrSet, sizeof(attrSet)); - dst += sizeof(attrSet); + ch = zebraExplain_lookup_attr_str(zei, p->index_type, p->index_name); + if (ch < 0) + ch = zebraExplain_add_attr_str(zei, p->index_type, p->index_name); } - if (!(lead & 2)) + else { - memcpy (dst, &attrUse, sizeof(attrUse)); - dst += sizeof(attrUse); +#if NATTR + return; +#else + ch = zebraExplain_lookup_attr_su(zei, p->index_type, + p->attrSet, p->attrUse); + if (ch < 0) + ch = zebraExplain_add_attr_su(zei, p->index_type, + p->attrSet, p->attrUse); +#endif } + key.len = 4; + key.mem[0] = ch; + key.mem[1] = p->record_id; + key.mem[2] = p->section_id; + key.mem[3] = p->seqno; + +#if 0 + /* just for debugging .. */ + yaz_log(YLOG_LOG, "add: set=%d use=%d " + "record_id=%lld section_id=%lld seqno=%lld", + p->attrSet, p->attrUse, p->record_id, p->section_id, p->seqno); #endif - *dst++ = p->reg_type; - memcpy (dst, str, length); - dst += length; - *dst++ = '\0'; - if (!diff) + extract_add_it_key(p->extractCtrl->handle, + &zh->reg->keys, + p->index_type, str, + length, &key); +} + +#if NATTR +static void extract_add_sort_string (RecWord *p, const char *str, int length) +{ + struct it_key key; + + ZebraHandle zh = p->extractCtrl->handle; + ZebraExplainInfo zei = zh->reg->zei; + int ch; + + if (p->index_name) { - memcpy (dst, pseqno, sizeof(*pseqno)); - dst += sizeof(*pseqno); + ch = zebraExplain_lookup_attr_str(zei, p->index_type, p->index_name); + if (ch < 0) + ch = zebraExplain_add_attr_str(zei, p->index_type, p->index_name); } -#endif - keys->buf_used = dst - keys->buf; -} + else + { + return; + } + key.len = 4; + key.mem[0] = ch; + key.mem[1] = p->record_id; + key.mem[2] = p->section_id; + key.mem[3] = p->seqno; -static void extract_add_sort_string (RecWord *p, const char *str, - int length) + extract_add_it_key(p->extractCtrl->handle, + &zh->reg->sortKeys, + p->index_type, str, + length, &key); +} +#else +static void extract_add_sort_string (RecWord *p, const char *str, int length) { ZebraHandle zh = p->extractCtrl->handle; struct sortKeys *sk = &zh->reg->sortKeys; @@ -1721,11 +1709,12 @@ static void extract_add_sort_string (RecWord *p, const char *str, memcpy (sk->buf + off, str, length); sk->buf_used = off + length; } +#endif void extract_add_string (RecWord *p, const char *string, int length) { assert (length > 0); - if (zebra_maps_is_sort (p->zebra_maps, p->reg_type)) + if (zebra_maps_is_sort (p->zebra_maps, p->index_type)) extract_add_sort_string (p, string, length); else extract_add_index_string (p, string, length); @@ -1733,12 +1722,14 @@ void extract_add_string (RecWord *p, const char *string, int length) static void extract_add_incomplete_field (RecWord *p) { - const char *b = p->string; - int remain = p->length; + const char *b = p->term_buf; + int remain = p->term_len; const char **map = 0; + + yaz_log(YLOG_DEBUG, "Incomplete field, w='%.*s'", p->term_len, p->term_buf); if (remain > 0) - map = zebra_maps_input(p->zebra_maps, p->reg_type, &b, remain); + map = zebra_maps_input(p->zebra_maps, p->index_type, &b, remain, 0); while (map) { @@ -1748,9 +1739,10 @@ static void extract_add_incomplete_field (RecWord *p) /* Skip spaces */ while (map && *map && **map == *CHR_SPACE) { - remain = p->length - (b - p->string); + remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input(p->zebra_maps, p->reg_type, &b, remain); + map = zebra_maps_input(p->zebra_maps, p->index_type, &b, + remain, 0); else map = 0; } @@ -1763,9 +1755,9 @@ static void extract_add_incomplete_field (RecWord *p) while (i < IT_MAX_WORD && *cp) buf[i++] = *(cp++); - remain = p->length - (b - p->string); + remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input(p->zebra_maps, p->reg_type, &b, remain); + map = zebra_maps_input(p->zebra_maps, p->index_type, &b, remain, 0); else map = 0; } @@ -1778,22 +1770,28 @@ static void extract_add_incomplete_field (RecWord *p) static void extract_add_complete_field (RecWord *p) { - const char *b = p->string; + const char *b = p->term_buf; char buf[IT_MAX_WORD+1]; const char **map = 0; - int i = 0, remain = p->length; + int i = 0, remain = p->term_len; + + yaz_log(YLOG_DEBUG, "Complete field, w='%.*s'", + p->term_len, p->term_buf); if (remain > 0) - map = zebra_maps_input (p->zebra_maps, p->reg_type, &b, remain); + map = zebra_maps_input (p->zebra_maps, p->index_type, &b, remain, 1); while (remain > 0 && i < IT_MAX_WORD) { while (map && *map && **map == *CHR_SPACE) { - remain = p->length - (b - p->string); + remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input(p->zebra_maps, p->reg_type, &b, remain); + { + int first = i ? 0 : 1; /* first position */ + map = zebra_maps_input(p->zebra_maps, p->index_type, &b, remain, first); + } else map = 0; } @@ -1814,13 +1812,16 @@ static void extract_add_complete_field (RecWord *p) { if (i >= IT_MAX_WORD) break; + yaz_log(YLOG_DEBUG, "Adding string to index '%d'", **map); while (i < IT_MAX_WORD && *cp) buf[i++] = *(cp++); } - remain = p->length - (b - p->string); + remain = p->term_len - (b - p->term_buf); if (remain > 0) - map = zebra_maps_input (p->zebra_maps, p->reg_type, &b, - remain); + { + map = zebra_maps_input (p->zebra_maps, p->index_type, &b, + remain, 0); + } else map = 0; } @@ -1834,29 +1835,89 @@ void extract_token_add (RecWord *p) { WRBUF wrbuf; #if 0 - yaz_log (LOG_LOG, "token_add " + yaz_log (YLOG_LOG, "token_add " "reg_type=%c attrSet=%d attrUse=%d seqno=%d s=%.*s", p->reg_type, p->attrSet, p->attrUse, p->seqno, p->length, p->string); #endif - if ((wrbuf = zebra_replace(p->zebra_maps, p->reg_type, 0, - p->string, p->length))) + if ((wrbuf = zebra_replace(p->zebra_maps, p->index_type, 0, + p->term_buf, p->term_len))) { - p->string = wrbuf_buf(wrbuf); - p->length = wrbuf_len(wrbuf); + p->term_buf = wrbuf_buf(wrbuf); + p->term_len = wrbuf_len(wrbuf); } - if (zebra_maps_is_complete (p->zebra_maps, p->reg_type)) + if (zebra_maps_is_complete (p->zebra_maps, p->index_type)) extract_add_complete_field (p); else extract_add_incomplete_field(p); } +static void extract_set_store_data_cb(struct recExtractCtrl *p, + void *buf, size_t sz) +{ + ZebraHandle zh = (ZebraHandle) p->handle; + + xfree(zh->store_data_buf); + zh->store_data_buf = 0; + zh->store_data_size = 0; + if (buf && sz) + { + zh->store_data_buf = xmalloc(sz); + zh->store_data_size = sz; + memcpy(zh->store_data_buf, buf, sz); + } +} + +static void extract_set_store_data_prepare(struct recExtractCtrl *p) +{ + ZebraHandle zh = (ZebraHandle) p->handle; + xfree(zh->store_data_buf); + zh->store_data_buf = 0; + zh->store_data_size = 0; + p->setStoreData = extract_set_store_data_cb; +} + void extract_schema_add (struct recExtractCtrl *p, Odr_oid *oid) { - ZebraHandle zh = (ZebraHandle) (p->handle); + ZebraHandle zh = (ZebraHandle) p->handle; zebraExplain_addSchema (zh->reg->zei, oid); } +#if NATTR +void extract_flushSortKeys (ZebraHandle zh, SYSNO sysno, + int cmd, struct recKeys *reckeys) +{ + SortIdx sortIdx = zh->reg->sortIdx; + void *decode_handle = iscz1_start(); + int off = 0; + int ch = 0; + + while (off < reckeys->buf_used) + { + const char *src = reckeys->buf + off; + struct it_key key; + char *dst = (char*) &key; + + iscz1_decode(decode_handle, &dst, &src); + assert(key.len == 4); + + ch = (int) key.mem[0]; /* ordinal for field/use/attribute */ + + sortIdx_type(sortIdx, ch); + if (cmd == 1) + sortIdx_add(sortIdx, src, strlen(src)); + else + sortIdx_add(sortIdx, "", 1); + + src += strlen(src); + src++; + + off = src - reckeys->buf; + } + assert (off == reckeys->buf_used); + iscz1_stop(decode_handle); +} +#else void extract_flushSortKeys (ZebraHandle zh, SYSNO sysno, int cmd, struct sortKeys *sk) { @@ -1881,6 +1942,7 @@ void extract_flushSortKeys (ZebraHandle zh, SYSNO sysno, off += slen; } } +#endif void encode_key_init (struct encode_info *i) { @@ -1891,39 +1953,9 @@ void encode_key_init (struct encode_info *i) i->prevseq=0; i->prevcmd=-1; i->keylen=0; -#if IT_KEY_NEW i->encode_handle = iscz1_start(); -#endif } -#if IT_KEY_NEW -#else -char *encode_key_int (int d, char *bp) -{ - if (d <= 63) - *bp++ = d; - else if (d <= 16383) - { - *bp++ = 64 + (d>>8); - *bp++ = d & 255; - } - else if (d <= 4194303) - { - *bp++ = 128 + (d>>16); - *bp++ = (d>>8) & 255; - *bp++ = d & 255; - } - else - { - *bp++ = 192 + (d>>24); - *bp++ = (d>>16) & 255; - *bp++ = (d>>8) & 255; - *bp++ = d & 255; - } - return bp; -} -#endif - #define OLDENCODE 1 #ifdef OLDENCODE @@ -1942,35 +1974,20 @@ void encode_key_write (char *k, struct encode_info *i, FILE *outf) ; /* and copy & align key so we can mangle */ memcpy (&key, k+1, sizeof(struct it_key)); /* *k is insert/delete */ -#if IT_KEY_NEW + bp0 = bp++; iscz1_encode(i->encode_handle, &bp, &src); *bp0 = (*k * 128) + bp - bp0 - 1; /* length and insert/delete combined */ -#else - bp = encode_key_int ( (key.sysno - i->sysno) * 2 + *k, bp); - if (i->sysno != key.sysno) - { - i->sysno = key.sysno; - i->seqno = 0; - } - else if (!i->seqno && !key.seqno && i->cmd == *k) - return; - bp = encode_key_int (key.seqno - i->seqno, bp); - i->seqno = key.seqno; - i->cmd = *k; -#endif if (fwrite (i->buf, bp - i->buf, 1, outf) != 1) { - logf (LOG_FATAL|LOG_ERRNO, "fwrite"); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fwrite"); exit (1); } } void encode_key_flush (struct encode_info *i, FILE *outf) { /* dummy routine */ -#if IT_KEY_NEW iscz1_stop(i->encode_handle); -#endif } #else @@ -2032,7 +2049,7 @@ void encode_key_write (char *k, struct encode_info *i, FILE *outf) i->cmd = i->prevcmd; if (fwrite (i->buf, bp - i->buf, 1, outf) != 1) { - logf (LOG_FATAL|LOG_ERRNO, "fwrite"); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fwrite"); exit (1); } i->keylen=0; /* ok, it's written, forget it */ @@ -2066,7 +2083,7 @@ void encode_key_flush (struct encode_info *i, FILE *outf) i->cmd = i->prevcmd; if (fwrite (i->buf, bp - i->buf, 1, outf) != 1) { - logf (LOG_FATAL|LOG_ERRNO, "fwrite"); + yaz_log (YLOG_FATAL|YLOG_ERRNO, "fwrite"); exit (1); } i->keylen=0; /* ok, it's written, forget it */