-/* $Id: extract.c,v 1.256 2007-04-25 08:22:01 adam Exp $
+/* $Id: extract.c,v 1.267 2007-10-31 16:56:14 adam Exp $
Copyright (C) 1995-2007
Index Data ApS
*/
+/** \file
+ \brief indexes records and extract tokens for indexing and sorting
+*/
+
#include <stdio.h>
#include <assert.h>
#include <ctype.h>
#endif
#include <fcntl.h>
+
#include "index.h"
#include "orddict.h"
#include <direntz.h>
#include <charmap.h>
+#include <yaz/snprintf.h>
static int log_level_extract = 0;
static int log_level_details = 0;
}
}
-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 extract_token_add2(RecWord *p);
static void check_log_limit(ZebraHandle zh)
{
}
}
-static void logRecord (ZebraHandle zh)
+static void logRecord(ZebraHandle zh)
{
check_log_limit(zh);
++zh->records_processed;
}
}
-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);
{
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_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_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,
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 ;
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;
}
*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;
}
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 == '\'')
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;
}
}
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;
}
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)
{
RecWord word;
extract_init(ctrl, &word);
word.index_name = "_ALLRECORDS";
- word.index_type = 'w';
+ word.index_type = "w";
word.seqno = 1;
- extract_add_index_string (&word, zinfo_index_category_alwaysmatches,
+ extract_add_index_string(&word, zinfo_index_category_alwaysmatches,
"", 0);
}
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);
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);
{
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;
}
{
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;
}
}
+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,
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;
}
stream->endf(stream, &null_offset);;
extractCtrl.init = extract_init;
- extractCtrl.tokenAdd = extract_token_add;
+ if (zh->reg->index_types)
+ {
+ extractCtrl.tokenAdd = extract_token_add2;
+ }
+ else
+ {
+ extractCtrl.tokenAdd = extract_token_add;
+ }
extractCtrl.schemaAdd = extract_schema_add;
extractCtrl.dh = zh->reg->dh;
extractCtrl.handle = 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;
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);
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;
}
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))
{
/* new record */
if (action == action_delete)
{
- yaz_log (YLOG_LOG, "delete %s %s " ZINT_FORMAT, recordType,
+ 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_WARN, "cannot delete record above (seems new)");
return ZEBRA_FAIL;
}
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);
+ recordAttr = rec_init_attr(zh->reg->zei, rec);
if (extractCtrl.staticrank < 0)
{
yaz_log(YLOG_WARN, "Negative staticrank for record. Set to 0");
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);
- 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,
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);
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,
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],
/* 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)
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);
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;
}
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);
}
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]);
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,
{
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);
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);
}
{
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);
}
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_rec_keys_t reckeys,
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];
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);
}
}
{
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);
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));
&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)
{
{
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;
}
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;
}
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)
{
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;
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;
}
if (!i)
return;
- extract_add_string (p, buf, i);
+ extract_add_string(p, zm, buf, i);
}
+static void extract_token_add2_index(ZebraHandle zh, zebra_index_type_t type,
+ RecWord *p)
+{
+ 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);
+
+ 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_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_details)
{
yaz_log(log_level_details, "extract_token_add "
- "type=%c index=%s seqno=" ZINT_FORMAT " s=%.*s",
+ "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_complete (zh->reg->zebra_maps, p->index_type))
- extract_add_complete_field (p);
+ if (zebra_maps_is_complete(zm))
+ extract_add_complete_field(p, zm);
else
- extract_add_incomplete_field(p);
+ extract_add_incomplete_field(p, zm);
}
static void extract_set_store_data_cb(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,