X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fextract.c;h=4309b60e2a99673b0ce50f5b6997affdda740926;hb=ef696645cc3b7e0f4027008d1dc589c0f0f90c1f;hp=de9e04ff9d105b5e0e9e151afa836b2ac876fd4b;hpb=75049be3951292bbbbd53d83ca38ccbd191e4b08;p=idzebra-moved-to-github.git diff --git a/index/extract.c b/index/extract.c index de9e04f..4309b60 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,10 +1,34 @@ /* - * Copyright (C) 1994-1998, Index Data + * Copyright (C) 1994-1999, Index Data * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: extract.c,v $ - * Revision 1.88 1998-10-16 08:14:29 adam + * Revision 1.96 1999-05-26 07:49:13 adam + * C++ compilation. + * + * Revision 1.95 1999/05/21 12:00:17 adam + * Better diagnostics for extraction process. + * + * Revision 1.94 1999/05/20 12:57:18 adam + * Implemented TCL filter. Updated recctrl system. + * + * Revision 1.93 1999/05/15 14:36:38 adam + * Updated dictionary. Implemented "compression" of dictionary. + * + * Revision 1.92 1999/03/09 16:27:49 adam + * More work on SDRKit integration. + * + * Revision 1.91 1999/02/12 13:29:22 adam + * Implemented position-flag for registers. + * + * Revision 1.90 1999/02/02 14:50:52 adam + * Updated WIN32 code specific sections. Changed header. + * + * Revision 1.89 1998/10/28 10:54:38 adam + * SDRKit integration. + * + * Revision 1.88 1998/10/16 08:14:29 adam * Updated record control system. * * Revision 1.87 1998/10/15 13:10:33 adam @@ -323,7 +347,7 @@ */ #include #include -#ifdef WINDOWS +#ifdef WIN32 #include #else #include @@ -337,6 +361,14 @@ #include "zinfo.h" +#ifndef ZEBRASDR +#define ZEBRASDR 0 +#endif + +#if ZEBRASDR +#include "zebrasdr.h" +#endif + static Dict matchDict; static Records records = NULL; @@ -378,14 +410,14 @@ int key_open (struct recordGroup *rGroup, int mem) mem = atoi(res_get_def (common_resource, "memMax", "4"))*1024*1024; if (mem < 50000) mem = 50000; - key_buf = xmalloc (mem); + key_buf = (char **) xmalloc (mem); ptr_top = mem/sizeof(char*); ptr_i = 0; key_buf_used = 0; key_file_no = 0; - if (!(matchDict = dict_open (bfs, GMATCH_DICT, 50, rw))) + if (!(matchDict = dict_open (bfs, GMATCH_DICT, 50, rw, 0))) { logf (LOG_FATAL, "dict_open fail of %s", GMATCH_DICT); return -1; @@ -412,6 +444,7 @@ int key_open (struct recordGroup *rGroup, int mem) struct encode_info { int sysno; int seqno; + int cmd; char buf[768]; }; @@ -419,6 +452,7 @@ void encode_key_init (struct encode_info *i) { i->sysno = 0; i->seqno = 0; + i->cmd = -1; } char *encode_key_int (int d, char *bp) @@ -460,8 +494,11 @@ void encode_key_write (char *k, struct encode_info *i, FILE *outf) 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; if (fwrite (i->buf, bp - i->buf, 1, outf) != 1) { logf (LOG_FATAL|LOG_ERRNO, "fwrite"); @@ -632,7 +669,7 @@ static void addIndexString (RecWord *p, const char *string, int length) { char *b; - b = xmalloc (reckeys.buf_max += 128000); + b = (char *) xmalloc (reckeys.buf_max += 128000); if (reckeys.buf_used > 0) memcpy (b, reckeys.buf, reckeys.buf_used); xfree (reckeys.buf); @@ -682,7 +719,8 @@ static void addIndexString (RecWord *p, const char *string, int length) dst += sizeof(*pseqno); } reckeys.buf_used = dst - reckeys.buf; - (*pseqno)++; + if (*pseqno) + (*pseqno)++; } static void addSortString (RecWord *p, const char *string, int length) @@ -693,11 +731,11 @@ static void addSortString (RecWord *p, const char *string, int length) if (sk->attrSet == p->attrSet && sk->attrUse == p->attrUse) return; - sk = xmalloc (sizeof(*sk)); + sk = (struct sortKey *) xmalloc (sizeof(*sk)); sk->next = sortKeys; sortKeys = sk; - sk->string = xmalloc (length); + sk->string = (char *) xmalloc (length); sk->length = length; memcpy (sk->string, string, length); @@ -950,20 +988,25 @@ static const char **searchRecordKey (struct recKeys *reckeys, } struct file_read_info { - off_t file_max; - off_t file_offset; - off_t file_moffset; + off_t file_max; /* maximum offset so far */ + off_t file_offset; /* current offset */ + 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) { - struct file_read_info *fi = xmalloc (sizeof(*fi)); + struct file_read_info *fi = (struct file_read_info *) + xmalloc (sizeof(*fi)); fi->fd = fd; fi->file_max = 0; fi->file_moffset = 0; + fi->sdrbuf = 0; + fi->sdrmax = 0; return fi; } @@ -974,23 +1017,34 @@ static void file_read_stop (struct file_read_info *fi) static off_t file_seek (void *handle, off_t offset) { - struct file_read_info *p = handle; + 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); } static off_t file_tell (void *handle) { - struct file_read_info *p = handle; + struct file_read_info *p = (struct file_read_info *) handle; return p->file_offset; } static int file_read (void *handle, char *buf, size_t count) { - struct file_read_info *p = handle; + struct file_read_info *p = (struct file_read_info *) handle; int fd = p->fd; int r; - r = read (fd, buf, count); + 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); if (r > 0) { p->file_offset += r; @@ -1002,17 +1056,17 @@ static int file_read (void *handle, char *buf, size_t count) static void file_begin (void *handle) { - struct file_read_info *p = handle; + struct file_read_info *p = (struct file_read_info *) handle; p->file_offset = p->file_moffset; - if (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 = handle; + struct file_read_info *p = (struct file_read_info *) handle; assert (p->file_more == 0); p->file_more = 1; @@ -1176,7 +1230,7 @@ struct recordLogInfo { static void recordLogPreamble (int level, const char *msg, void *info) { - struct recordLogInfo *p = info; + struct recordLogInfo *p = (struct recordLogInfo *) info; FILE *outf = log_file (); if (level & LOG_LOG) @@ -1194,7 +1248,7 @@ void addSchema (struct recExtractCtrl *p, Odr_oid *oid) static int recordExtract (SYSNO *sysno, const char *fname, struct recordGroup *rGroup, int deleteFlag, struct file_read_info *fi, - RecType recType, char *subType) + RecType recType, char *subType, void *clientData) { RecordAttr *recordAttr; int r; @@ -1229,7 +1283,12 @@ static int recordExtract (SYSNO *sysno, const char *fname, extractCtrl.addSchema = addSchema; extractCtrl.dh = rGroup->dh; for (i = 0; i<256; i++) - extractCtrl.seqno[i] = 0; + { + if (zebra_maps_is_positioned(rGroup->zebra_maps, i)) + extractCtrl.seqno[i] = 1; + else + extractCtrl.seqno[i] = 0; + } extractCtrl.zebra_maps = rGroup->zebra_maps; extractCtrl.flagShowRecords = !rGroup->flagRw; @@ -1241,18 +1300,20 @@ static int recordExtract (SYSNO *sysno, const char *fname, logInfo.rGroup = rGroup; log_event_start (recordLogPreamble, &logInfo); - r = (*recType->extract)(&extractCtrl); + r = (*recType->extract)(clientData, &extractCtrl); log_event_start (NULL, NULL); - if (r) - { + if (r == RECCTRL_EXTRACT_EOF) + return 0; + else if (r == RECCTRL_EXTRACT_ERROR) + { /* error occured during extraction ... */ if (rGroup->flagRw && records_processed < rGroup->fileVerboseLimit) { - logf (LOG_WARN, "fail %s %s %ld code = %d", rGroup->recordType, - fname, (long) recordOffset, r); + logf (LOG_WARN, "fail %s %s %ld", rGroup->recordType, + fname, (long) recordOffset); } return 0; } @@ -1434,7 +1495,8 @@ static int recordExtract (SYSNO *sysno, const char *fname, if (rGroup->flagStoreData == 1) { rec->size[recInfo_storeData] = recordAttr->recordSize; - rec->info[recInfo_storeData] = xmalloc (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 %ld in %s", @@ -1480,6 +1542,7 @@ int fileExtract (SYSNO *sysno, const char *fname, struct recordGroup rGroupM; struct recordGroup *rGroup = &rGroupM; struct file_read_info *fi; + void *clientData; memcpy (rGroup, rGroupP, sizeof(*rGroupP)); @@ -1519,7 +1582,8 @@ int fileExtract (SYSNO *sysno, const char *fname, if (!*rGroup->recordType) return 0; if (!(recType = - recType_byName (rGroup->recTypes, rGroup->recordType, subType))) + recType_byName (rGroup->recTypes, rGroup->recordType, subType, + &clientData))) { logf (LOG_WARN, "No such record type: %s", rGroup->recordType); return 0; @@ -1590,6 +1654,59 @@ int fileExtract (SYSNO *sysno, const char *fname, if (rGroup->flagStoreKeys == -1) rGroup->flagStoreKeys = 0; +#if ZEBRASDR + if (rGroup->useSDR) + { + ZebraSdrHandle h; + char xname[128], *xp; + + strncpy (xname, fname, 127); + if (!(xp = strchr (xname, '.'))) + return 0; + *xp = '\0'; + if (strcmp (xp+1, "sdr.bits")) + return 0; + + h = zebraSdr_open (xname); + if (!h) + { + logf (LOG_WARN, "sdr open %s", xname); + return 0; + } + for (;;) + { + unsigned char *buf; + char sdr_name[128]; + int r, segmentno; + + segmentno = zebraSdr_segment (h, 0); + sprintf (sdr_name, "%%%s.%d", xname, segmentno); + +#if 0 + if (segmentno > 20) + break; +#endif + r = zebraSdr_read (h, &buf); + + if (!r) + break; + + fi = file_read_start (0); + fi->sdrbuf = buf; + fi->sdrmax = r; + do + { + file_begin (fi); + r = recordExtract (sysno, sdr_name, rGroup, deleteFlag, fi, + recType, subType); + } while (r && !sysno && fi->file_more); + file_read_stop (fi); + free (buf); + } + zebraSdr_close (h); + return 1; + } +#endif if (sysno && deleteFlag) fd = -1; else @@ -1605,7 +1722,7 @@ int fileExtract (SYSNO *sysno, const char *fname, { file_begin (fi); r = recordExtract (sysno, fname, rGroup, deleteFlag, fi, - recType, subType); + recType, subType, clientData); } while (r && !sysno && fi->file_more); file_read_stop (fi); if (fd != -1)