-/* $Id: retrieve.c,v 1.72 2007-08-21 13:27:04 adam Exp $
+/* $Id: retrieve.c,v 1.78 2007-12-03 13:04:04 adam Exp $
Copyright (C) 1995-2007
Index Data ApS
#include <yaz/snprintf.h>
#include <direntz.h>
#include <yaz/oid_db.h>
+#include <zebra_strmap.h>
+
+#define MAX_SYSNOS_PER_RECORD 40
#define ZEBRA_XML_HEADER_STR "<record xmlns=\"http://www.indexdata.com/zebra/\""
else
strcpy(full_rep, (*rec)->info[recInfo_filename]);
- if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1){
- yaz_log (YLOG_WARN|YLOG_ERRNO, "Retrieve fail; missing file: %s",
+ if ((fd = open(full_rep, O_BINARY|O_RDONLY)) == -1){
+ yaz_log(YLOG_WARN|YLOG_ERRNO, "Retrieve fail; missing file: %s",
full_rep);
rec_free(rec);
return YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
}
+struct index_spec {
+ const char *index_name;
+ const char *index_type;
+ struct index_spec *next;
+};
+
+
+struct index_spec *parse_index_spec(const char *elem, NMEM nmem,
+ int *error)
+{
+ struct index_spec *first = 0;
+ struct index_spec **last = &first;
+ const char *cp = elem;
+
+ *error = 0;
+ if (cp[0] == ':' && cp[1] == ':')
+ {
+
+ cp++; /* skip first ':' */
+ for (;;)
+ {
+ const char *cp0;
+ struct index_spec *spec = nmem_malloc(nmem, sizeof(*spec));
+ spec->index_type = 0;
+ spec->next = 0;
+
+ if (!first)
+ first = spec;
+ *last = spec;
+ last = &spec->next;
+
+ cp++; /* skip ',' or second ':' */
+ cp0 = cp;
+ while (*cp != ':' && *cp != '\0' && *cp != ',')
+ cp++;
+ spec->index_name = nmem_strdupn(nmem, cp0, cp - cp0);
+ if (*cp == ':') /* type as well */
+ {
+ cp++;
+ cp0 = cp;
+
+ while (*cp != '\0' && *cp != ',')
+ cp++;
+ spec->index_type = nmem_strdupn(nmem, cp0, cp - cp0);
+ }
+ if (*cp != ',')
+ break;
+ }
+ }
+ if (*cp != '\0')
+ *error = 1;
+ return first;
+}
+
static int parse_zebra_elem(const char *elem,
const char **index, size_t *index_len,
const char **type, size_t *type_len)
const char *retrieval_type;
size_t retrieval_type_len;
char retrieval_index_cstr[256];
+ char retrieval_type_cstr[256];
int ord;
/* only accept XML and SUTRS requests */
memcpy(retrieval_index_cstr, retrieval_index, retrieval_index_len);
retrieval_index_cstr[retrieval_index_len] = '\0';
+ memcpy(retrieval_type_cstr, retrieval_type, retrieval_type_len);
+ retrieval_type_cstr[retrieval_type_len] = '\0';
+
ord = zebraExplain_lookup_attr_str(zh->reg->zei,
zinfo_index_category_sort,
- retrieval_type[0],
+ retrieval_type_cstr,
retrieval_index_cstr);
if (ord == -1)
return -1; /* is not a sort index */
{
char dst_buf[IT_MAX_WORD];
char str[IT_MAX_WORD];
- int index_type;
+ const char *index_type;
const char *db = 0;
const char *string_index = 0;
WRBUF wrbuf = wrbuf_alloc();
zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, &db, &string_index);
zebra_term_untrans(zh, index_type, dst_buf, str);
-
if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
{
wrbuf_printf(wrbuf, " <index name=\"%s\"",
string_index);
- wrbuf_printf(wrbuf, " type=\"%c\">", index_type);
+ wrbuf_printf(wrbuf, " type=\"%s\">", index_type);
wrbuf_xmlputs(wrbuf, dst_buf);
wrbuf_printf(wrbuf, "</index>\n");
wrbuf_printf(wrbuf, "</record>\n");
{
*output_format = yaz_oid_recsyn_sutrs;
- wrbuf_printf(wrbuf, "%s %c %s\n", string_index, index_type,
+ wrbuf_printf(wrbuf, "%s %s %s\n", string_index, index_type,
dst_buf);
}
*rec_lenp = wrbuf_len(wrbuf);
size_t retrieval_type_len;
zebra_rec_keys_t keys;
int ret_code = 0;
+ char retrieval_type_cstr[256];
/* set output variables before processing possible error states */
/* *rec_lenp = 0; */
&retrieval_type, &retrieval_type_len))
return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
- if (retrieval_type_len != 0 && retrieval_type_len != 1)
+ if (retrieval_type_len)
{
- return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
+ memcpy(retrieval_type_cstr, retrieval_type, retrieval_type_len);
+ retrieval_type_cstr[retrieval_type_len] = '\0';
}
-
+
if (retrieval_index_len)
{
char retrieval_index_cstr[256];
if (zebraExplain_lookup_attr_str(zh->reg->zei,
zinfo_index_category_index,
- (retrieval_type_len == 0 ? -1 :
- retrieval_type[0]),
+ (retrieval_type_len == 0 ? 0 :
+ retrieval_type_cstr),
retrieval_index_cstr) == -1)
return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
}
{
int i;
int ord = CAST_ZINT_TO_INT(key_in.mem[0]);
- int index_type;
+ const char *index_type;
const char *db = 0;
const char *string_index = 0;
size_t string_index_len;
{
/* process only if type is not defined, or is matching */
if (retrieval_type == 0
- || (retrieval_type_len == 1
- && retrieval_type[0] == index_type))
+ || !strcmp(retrieval_type_cstr, index_type))
{
zebra_term_untrans(zh, index_type, dst_buf, str);
if (strlen(dst_buf))
wrbuf_printf(wrbuf, " <index name=\"%s\"",
string_index);
- wrbuf_printf(wrbuf, " type=\"%c\"", index_type);
+ wrbuf_printf(wrbuf, " type=\"%s\"", index_type);
wrbuf_printf(wrbuf, " seq=\"" ZINT_FORMAT "\">",
key_in.mem[key_in.len -1]);
{
wrbuf_printf(wrbuf, "%s ", string_index);
- wrbuf_printf(wrbuf, "%c", index_type);
+ wrbuf_printf(wrbuf, "%s", index_type);
for (i = 1; i < key_in.len; i++)
wrbuf_printf(wrbuf, " " ZINT_FORMAT,
key_in.mem[i]);
- /* zebra_term_untrans(zh, index_type, dst_buf, str); */
wrbuf_printf(wrbuf, " %s", dst_buf);
wrbuf_printf(wrbuf, "\n");
{
if (doc_w->mark)
{
- int index_type;
+ const char *index_type;
const char *db = 0;
const char *string_index = 0;
if (mark_state == 0)
{
wrbuf_printf(wrbuf, " <snippet name=\"%s\"", string_index);
- wrbuf_printf(wrbuf, " type=\"%c\">", index_type);
+ wrbuf_printf(wrbuf, " type=\"%s\">", index_type);
}
if (doc_w->match)
wrbuf_puts(wrbuf, "<s>");
return return_code;
}
-int zebra_special_snippet_fetch(ZebraHandle zh, const char *setname,
- zint sysno, ODR odr,
- const char *elemsetname,
- const Odr_oid *input_format,
- const Odr_oid **output_format,
- char **rec_bufp, int *rec_lenp)
+static int snippet_fetch(ZebraHandle zh, const char *setname,
+ zint sysno, ODR odr,
+ const char *elemsetname,
+ const Odr_oid *input_format,
+ const Odr_oid **output_format,
+ char **rec_bufp, int *rec_lenp)
{
zebra_snippets *rec_snippets = zebra_snippets_create();
int return_code = zebra_get_rec_snippets(zh, sysno, rec_snippets);
return return_code;
}
+struct term_collect {
+ const char *term;
+ int oc;
+ zint set_occur;
+};
+
+zint freq_term(ZebraHandle zh, int ord, const char *term, RSET rset_set)
+{
+ struct rset_key_control *kc = zebra_key_control_create(zh);
+ char ord_buf[IT_MAX_WORD];
+ int ord_len = key_SU_encode(ord, ord_buf);
+ char *info;
+ zint hits = 0;
+ NMEM nmem = nmem_create();
+
+ strcpy(ord_buf + ord_len, term);
+
+ info = dict_lookup(zh->reg->dict, ord_buf);
+ if (info)
+ {
+ ISAM_P isam_p;
+ RSET rsets[2], rset;
+ memcpy(&isam_p, info+1, sizeof(ISAM_P));
+
+ rsets[0] = zebra_create_rset_isam(zh, nmem, kc, kc->scope, isam_p, 0);
+ rsets[1] = rset_dup(rset_set);
+
+ rset = rset_create_and(nmem, kc, kc->scope, 2, rsets);
+
+ zebra_count_set(zh, rset, &hits, zh->approx_limit);
+
+ rset_delete(rsets[0]);
+ rset_delete(rset);
+ }
+ (*kc->dec)(kc);
+ nmem_destroy(nmem);
+ return hits;
+}
+
+void term_collect_freq(ZebraHandle zh,
+ struct term_collect *col, int no_terms_collect,
+ int ord, RSET rset)
+{
+ int i;
+ for (i = 0; i < no_terms_collect; i++)
+ {
+ if (col[i].term)
+ col[i].set_occur = freq_term(zh, ord, col[i].term, rset);
+ }
+}
+
+struct term_collect *term_collect_create(zebra_strmap_t sm,
+ int no_terms_collect,
+ NMEM nmem)
+{
+ const char *term;
+ void *data_buf;
+ size_t data_len;
+ zebra_strmap_it it;
+ struct term_collect *col = nmem_malloc(nmem,
+ sizeof *col *no_terms_collect);
+ int i;
+ for (i = 0; i < no_terms_collect; i++)
+ {
+ col[i].term = 0;
+ col[i].oc = 0;
+ col[i].set_occur = 0;
+ }
+ /* iterate over terms and collect the most frequent ones */
+ it = zebra_strmap_it_create(sm);
+ while ((term = zebra_strmap_it_next(it, &data_buf, &data_len)))
+ {
+ int oc = *(int*) data_buf;
+ int j = 0;
+ /* insertion may be slow but terms terms will be "infrequent" and
+ thus number of iterations should be small below */
+ while (j < no_terms_collect && oc > col[j].oc)
+ j++;
+ if (j)
+ {
+ --j;
+ memmove(col, col+1, sizeof(*col) * j);
+ col[j].term = term;
+ col[j].oc = oc;
+ }
+ }
+ zebra_strmap_it_destroy(it);
+ return col;
+}
+
+static ZEBRA_RES facet_fetch(ZebraHandle zh, const char *setname,
+ ODR odr,
+ const char *elemsetname,
+ const Odr_oid *input_format,
+ const Odr_oid **output_format,
+ char **rec_bufp, int *rec_lenp)
+{
+ zint *pos_array;
+ int i;
+ int num_recs = 10; /* number of records to analyze */
+ int no_collect_terms = 20; /* number of term candidates */
+ ZebraMetaRecord *poset;
+ ZEBRA_RES ret = ZEBRA_OK;
+ int *ord_array;
+ WRBUF wr = wrbuf_alloc();
+
+ int no_ord = 0;
+ struct index_spec *spec, *spec_list;
+ int error;
+
+
+ spec_list = parse_index_spec(elemsetname, odr_getmem(odr), &error);
+
+ if (!spec_list || error)
+ return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
+
+ for (spec = spec_list; spec; spec = spec->next)
+ {
+ if (!spec->index_type)
+ return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
+ no_ord++;
+ }
+
+ ord_array = odr_malloc(odr, sizeof(*ord_array) * no_ord);
+
+ for (spec = spec_list, i = 0; spec; spec = spec->next, i++)
+ {
+ int ord = zebraExplain_lookup_attr_str(zh->reg->zei,
+ zinfo_index_category_index,
+ spec->index_type,
+ spec->index_name);
+ if (ord == -1)
+ {
+ return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
+ }
+ ord_array[i] = ord;
+ }
+
+ pos_array = (zint *) xmalloc(num_recs * sizeof(*pos_array));
+ for (i = 0; i < num_recs; i++)
+ pos_array[i] = i+1;
+ poset = zebra_meta_records_create(zh, setname, num_recs, pos_array);
+ if (!poset)
+ {
+ zebra_setError(zh, YAZ_BIB1_SPECIFIED_RESULT_SET_DOES_NOT_EXIST,
+ setname);
+ xfree(pos_array);
+ ret = ZEBRA_FAIL;
+ }
+ else
+ {
+ zebra_strmap_t *map_array
+ = odr_malloc(odr, sizeof *map_array * no_ord);
+ for (i = 0; i < no_ord; i++)
+ map_array[i] = zebra_strmap_create();
+
+ for (i = 0; i < num_recs; i++)
+ {
+ int j;
+ zint sysnos[MAX_SYSNOS_PER_RECORD];
+ int no_sysnos = MAX_SYSNOS_PER_RECORD;
+ if (!poset[i].sysno)
+ continue;
+ ret = zebra_result_recid_to_sysno(zh, setname,
+ poset[i].sysno,
+ sysnos, &no_sysnos);
+ assert(no_sysnos > 0);
+ for (j = 0; j < no_sysnos; j++)
+ {
+ size_t slen;
+ const char *str;
+ struct it_key key_in;
+ Record rec = rec_get(zh->reg->records, sysnos[j]);
+ zebra_rec_keys_t keys = zebra_rec_keys_open();
+ zebra_rec_keys_set_buf(keys, rec->info[recInfo_delKeys],
+ rec->size[recInfo_delKeys], 0);
+
+ if (zebra_rec_keys_rewind(keys))
+ {
+ while (zebra_rec_keys_read(keys, &str, &slen, &key_in))
+ {
+ int i;
+ struct index_spec *spec;
+ for (spec = spec_list, i = 0; i < no_ord;
+ i++, spec = spec->next)
+ {
+ int ord = CAST_ZINT_TO_INT(key_in.mem[0]);
+ if (ord == ord_array[i])
+ {
+ int *freq;
+ zebra_strmap_t sm = map_array[i];
+
+ freq = zebra_strmap_lookup(sm, str, 0, 0);
+ if (freq)
+ (*freq)++;
+ else
+ {
+ int v = 1;
+ zebra_strmap_add(sm, str, &v, sizeof v);
+ }
+ }
+ }
+ }
+ }
+ zebra_rec_keys_close(keys);
+ rec_free(&rec);
+ }
+ }
+ wrbuf_puts(wr, "<facets>\n");
+ for (spec = spec_list, i = 0; i < no_ord; i++, spec = spec->next)
+ {
+ int j;
+ NMEM nmem = nmem_create();
+ struct term_collect *col = term_collect_create(map_array[i],
+ no_collect_terms,
+ nmem);
+ term_collect_freq(zh, col, no_collect_terms, ord_array[i],
+ resultSetRef(zh, setname));
+
+ wrbuf_printf(wr, " <facet type=\"%s\" index=\"%s\">\n",
+ spec->index_type, spec->index_name);
+ for (j = 0; j < no_collect_terms; j++)
+ {
+ if (col[j].term)
+ {
+ char dst_buf[IT_MAX_WORD];
+ zebra_term_untrans(zh, spec->index_type, dst_buf, col[j].term);
+ wrbuf_printf(wr, " <term coccur=\"%d\"", col[j].oc);
+ if (col[j].set_occur)
+ wrbuf_printf(wr, " occur=\"" ZINT_FORMAT "\"",
+ col[j].set_occur);
+ wrbuf_printf(wr, ">");
+ wrbuf_xmlputs(wr, dst_buf);
+ wrbuf_printf(wr, "</term>\n");
+ }
+ }
+ wrbuf_puts(wr, " </facet>\n");
+ nmem_destroy(nmem);
+ }
+ wrbuf_puts(wr, "</facets>\n");
+ for (i = 0; i < no_ord; i++)
+ zebra_strmap_destroy(map_array[i]);
+ }
+
+
+ *rec_bufp = odr_strdup(odr, wrbuf_cstr(wr));
+ wrbuf_destroy(wr);
+ *rec_lenp = strlen(*rec_bufp);
+ *output_format = yaz_oid_recsyn_xml;
+
+ xfree(pos_array);
+ zebra_meta_records_destroy(zh, poset, num_recs);
+ return ret;
+}
+
int zebra_special_fetch(ZebraHandle zh, const char *setname,
zint sysno, int score, ODR odr,
const char *elemsetname,
/* set output variables before processing possible error states */
/* *rec_lenp = 0; */
+ if (elemsetname && 0 == strncmp(elemsetname, "facet", 5))
+ {
+ return facet_fetch(zh, setname, odr,
+ elemsetname + 5,
+ input_format, output_format,
+ rec_bufp, rec_lenp);
+ }
if (elemsetname && 0 == strcmp(elemsetname, "snippet"))
{
- return zebra_special_snippet_fetch(zh, setname, sysno, odr,
- elemsetname + 7,
- input_format, output_format,
- rec_bufp, rec_lenp);
+ return snippet_fetch(zh, setname, sysno, odr,
+ elemsetname + 7,
+ input_format, output_format,
+ rec_bufp, rec_lenp);
}
/* processing zebra::meta::sysno elemset without fetching binary data */
int zebra_record_fetch(ZebraHandle zh, const char *setname,
zint sysno, int score,
- zebra_snippets *hit_snippet, ODR odr,
+ ODR odr,
const Odr_oid *input_format, Z_RecordComposition *comp,
const Odr_oid **output_format,
char **rec_bufp, int *rec_lenp, char **basenamep,
RecordAttr *recordAttr;
void *clientData;
int return_code = 0;
+ zint sysnos[MAX_SYSNOS_PER_RECORD];
+ int no_sysnos = MAX_SYSNOS_PER_RECORD;
+ ZEBRA_RES res;
+ res = zebra_result_recid_to_sysno(zh, setname, sysno, sysnos, &no_sysnos);
+ if (res != ZEBRA_OK)
+ return ZEBRA_FAIL;
+
+ sysno = sysnos[0];
*basenamep = 0;
*addinfo = 0;
elemsetname = yaz_get_esn(comp);
file_type = rec->info[recInfo_fileType];
fname = rec->info[recInfo_filename];
basename = rec->info[recInfo_databaseName];
- *basenamep = (char *) odr_malloc (odr, strlen(basename)+1);
- strcpy (*basenamep, basename);
+ *basenamep = (char *) odr_malloc(odr, strlen(basename)+1);
+ strcpy(*basenamep, basename);
yaz_log(YLOG_DEBUG, "retrieve localno=" ZINT_FORMAT " score=%d",
sysno, score);
if (rec)
{
- zebra_rec_keys_t reckeys = zebra_rec_keys_open();
RecType rt;
struct recRetrieveCtrl retrieveCtrl;
retrieveCtrl.res = zh->res;
retrieveCtrl.rec_buf = 0;
retrieveCtrl.rec_len = -1;
- retrieveCtrl.hit_snippet = hit_snippet;
- retrieveCtrl.doc_snippet = zebra_snippets_create();
-
- zebra_rec_keys_set_buf(reckeys,
- rec->info[recInfo_delKeys],
- rec->size[recInfo_delKeys],
- 0);
- zebra_rec_keys_to_snippets(zh, reckeys, retrieveCtrl.doc_snippet);
- zebra_rec_keys_close(reckeys);
if (!(rt = recType_byName(zh->reg->recTypes, zh->res,
file_type, &clientData)))
*addinfo = retrieveCtrl.addinfo;
}
- zebra_snippets_destroy(retrieveCtrl.doc_snippet);
-
stream.destroy(&stream);
rec_free(&rec);
}
projects / idzebra-moved-to-github.git / blobdiff