-/*
- * Copyright (C) 1995-1996, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: index.h,v $
- * Revision 1.44 1996-06-06 12:08:40 quinn
- * Added showRecord function
- *
- * Revision 1.43 1996/06/04 10:18:12 adam
- * Search/scan uses character mapping module.
- *
- * Revision 1.42 1996/06/04 08:20:16 quinn
- * Smallish
- *
- * Revision 1.41 1996/06/04 07:54:55 quinn
- * Added output-map.
- *
- * Revision 1.40 1996/05/31 09:06:58 quinn
- * Work on character-set handling
- *
- * Revision 1.39 1996/05/14 14:04:33 adam
- * In zebraidx, the 'stat' command is improved. Statistics about ISAM/DICT
- * is collected.
- *
- * Revision 1.38 1996/04/12 07:02:23 adam
- * File update of single files.
- *
- * Revision 1.37 1996/03/26 16:01:13 adam
- * New setting lockPath: directory of various lock files.
- *
- * Revision 1.36 1996/03/21 14:50:09 adam
- * File update uses modify-time instead of change-time.
- *
- * Revision 1.35 1996/02/12 18:45:36 adam
- * New fileVerboseFlag in record group control.
- *
- * Revision 1.34 1995/12/11 11:43:29 adam
- * Locking based on fcntl instead of flock.
- * Setting commitEnable removed. Command line option -n can be used to
- * prevent commit if commit setting is defined in the configuration file.
- *
- * Revision 1.33 1995/12/08 16:22:53 adam
- * Work on update while servers are running. Three lock files introduced.
- * The servers reload their registers when necessary, but they don't
- * reestablish result sets yet.
- *
- * Revision 1.32 1995/12/07 17:38:46 adam
- * Work locking mechanisms for concurrent updates/commit.
- *
- * Revision 1.31 1995/12/06 12:41:22 adam
- * New command 'stat' for the index program.
- * Filenames can be read from stdin by specifying '-'.
- * Bug fix/enhancement of the transformation from terms to regular
- * expressons in the search engine.
- *
- * Revision 1.30 1995/12/05 11:25:02 adam
- * Include of zebraver.h.
- *
- * Revision 1.29 1995/11/28 09:09:40 adam
- * Zebra config renamed.
- * Use setting 'recordId' to identify record now.
- * Bug fix in recindex.c: rec_release_blocks was invokeded even
- * though the blocks were already released.
- * File traversal properly deletes records when needed.
- *
- * Revision 1.28 1995/11/27 13:58:53 adam
- * New option -t. storeStore data implemented in server.
- *
- * Revision 1.27 1995/11/25 10:24:06 adam
- * More record fields - they are enumerated now.
- * New options: flagStoreData flagStoreKey.
- *
- * Revision 1.26 1995/11/22 17:19:17 adam
- * Record management uses the bfile system.
- *
- * Revision 1.25 1995/11/21 15:29:12 adam
- * Config file 'base' read by default by both indexer and server.
- *
- * Revision 1.24 1995/11/21 15:01:15 adam
- * New general match criteria implemented.
- * New feature: document groups.
- *
- * Revision 1.23 1995/11/20 16:59:45 adam
- * New update method: the 'old' keys are saved for each records.
- *
- * Revision 1.22 1995/11/20 11:56:26 adam
- * Work on new traversal.
- *
- * Revision 1.21 1995/11/16 15:34:55 adam
- * Uses new record management system in both indexer and server.
- *
- * Revision 1.20 1995/11/15 14:46:18 adam
- * Started work on better record management system.
- *
- * Revision 1.19 1995/10/27 14:00:11 adam
- * Implemented detection of database availability.
- *
- * Revision 1.18 1995/10/17 18:02:08 adam
- * New feature: databases. Implemented as prefix to words in dictionary.
- *
- * Revision 1.17 1995/10/13 16:01:49 adam
- * Work on relations.
- *
- * Revision 1.16 1995/10/10 12:24:38 adam
- * Temporary sort files are compressed.
- *
- * Revision 1.15 1995/10/04 16:57:19 adam
- * Key input and merge sort in one pass.
- *
- * Revision 1.14 1995/09/29 14:01:40 adam
- * Bug fixes.
- *
- * Revision 1.13 1995/09/28 14:22:56 adam
- * Sort uses smaller temporary files.
- *
- * Revision 1.12 1995/09/28 12:10:32 adam
- * Bug fixes. Field prefix used in queries.
- *
- * Revision 1.11 1995/09/27 12:22:28 adam
- * More work on extract in record control.
- * Field name is not in isam keys but in prefix in dictionary words.
- *
- * Revision 1.10 1995/09/14 07:48:23 adam
- * Record control management.
- *
- * Revision 1.9 1995/09/11 13:09:33 adam
- * More work on relevance feedback.
- *
- * Revision 1.8 1995/09/08 14:52:27 adam
- * Minor changes. Dictionary is lower case now.
- *
- * Revision 1.7 1995/09/06 16:11:16 adam
- * Option: only one word key per file.
- *
- * Revision 1.6 1995/09/05 15:28:39 adam
- * More work on search engine.
- *
- * Revision 1.5 1995/09/04 12:33:42 adam
- * Various cleanup. YAZ util used instead.
- *
- * Revision 1.4 1995/09/04 09:10:35 adam
- * More work on index add/del/update.
- * Merge sort implemented.
- * Initial work on z39 server.
- *
- * Revision 1.3 1995/09/01 14:06:35 adam
- * Split of work into more files.
- *
- * Revision 1.2 1995/09/01 10:30:24 adam
- * More work on indexing. Not working yet.
- *
- * Revision 1.1 1995/08/31 14:50:24 adam
- * New simple file index tool.
- *
- */
+/* $Id: index.h,v 1.175 2006-09-08 14:40:52 adam Exp $
+ Copyright (C) 1995-2006
+ Index Data ApS
+
+This file is part of the Zebra server.
+
+Zebra is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; if not, write to the Free Software
+Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+*/
+
+#ifndef INDEX_H
+#define INDEX_H
#include <time.h>
-#include <zebraver.h>
-#include <alexutil.h>
-#include <dict.h>
-#include <isam.h>
+#include <stdlib.h>
+#include <idzebra/version.h>
+#include <idzebra/util.h>
+#include <idzebra/flock.h>
+#include <sortidx.h>
+
+#if HAVE_SYS_TIMES_H
+#include <sys/times.h>
+#endif
+#include <sys/stat.h>
+
+#include <idzebra/dict.h>
+#include <idzebra/isams.h>
+#include <idzebra/isamc.h>
+#include <idzebra/isamb.h>
+#include <d1_absyn.h>
+#include <idzebra/recgrs.h>
+#include "recindex.h"
+#include <idzebra/api.h>
+#include "zinfo.h"
+#include <passwddb.h>
+#include <rset.h>
+#include <zebramap.h>
+
+YAZ_BEGIN_CDECL
+
+#define ISAM_DEFAULT "b"
+
+#define SU_SCHEME 1
#define IT_MAX_WORD 256
-#define IT_KEY_HAVE_SEQNO 1
-#define IT_KEY_HAVE_FIELD 0
+#define IT_KEY_LEVEL_MAX 5
struct it_key {
- int sysno;
- int seqno;
+ int len;
+ zint mem[IT_KEY_LEVEL_MAX];
};
enum dirsKind { dirs_dir, dirs_file };
time_t mtime;
};
-struct recordGroup {
- char *groupName;
- char *databaseName;
- char *path;
- char *recordId;
- char *recordType;
- int flagStoreData;
- int flagStoreKeys;
- int flagShowRecords;
- int fileVerboseFlag;
-};
-
+void getFnameTmp (Res res, char *fname, int no);
-struct dirs_info *dirs_open (Dict dict, const char *rep);
-struct dirs_info *dirs_fopen (Dict dict, const char *path);
+struct dirs_info *dirs_open (Dict dict, const char *rep, int rw);
+struct dirs_info *dirs_fopen (Dict dict, const char *path, int rw);
struct dirs_entry *dirs_read (struct dirs_info *p);
struct dirs_entry *dirs_last (struct dirs_info *p);
void dirs_mkdir (struct dirs_info *p, const char *src, time_t mtime);
void dirs_rmdir (struct dirs_info *p, const char *src);
-void dirs_add (struct dirs_info *p, const char *src, int sysno, time_t mtime);
+void dirs_add (struct dirs_info *p, const char *src, SYSNO sysno, time_t mtime);
void dirs_del (struct dirs_info *p, const char *src);
void dirs_free (struct dirs_info **pp);
-struct dir_entry *dir_open (const char *rep);
+struct dir_entry *dir_open (const char *rep, const char *base,
+ int follow_links);
void dir_sort (struct dir_entry *e);
void dir_free (struct dir_entry **e_p);
-void repositoryUpdate (struct recordGroup *rGroup);
-void repositoryAdd (struct recordGroup *rGroup);
-void repositoryDelete (struct recordGroup *rGroup);
-void repositoryShow (struct recordGroup *rGroup);
+void repositoryUpdate (ZebraHandle zh, const char *path);
+void repositoryAdd (ZebraHandle zh, const char *path);
+void repositoryDelete (ZebraHandle zh, const char *path);
+void repositoryShow (ZebraHandle zh, const char *path);
-void key_open (int mem);
-int key_close (void);
-void key_write (int cmd, struct it_key *k, const char *str);
+int key_open (ZebraHandle zh, int mem);
+int key_close (ZebraHandle zh);
int key_compare (const void *p1, const void *p2);
+void key_init(struct it_key *k);
+char *key_print_it (const void *p, char *buf);
+zint key_get_seq (const void *p);
+zint key_get_segment (const void *p);
+int key_compare_it (const void *p1, const void *p2);
int key_qsort_compare (const void *p1, const void *p2);
void key_logdump (int mask, const void *p);
-void inv_prstat (const char *dict_fname, const char *isam_fname);
-void key_input (const char *dict_fname, const char *isam_fname,
- int nkeys, int cache);
+void key_logdump_txt (int logmask, const void *p, const char *txt);
+void inv_prstat (ZebraHandle zh);
+void inv_compact (BFiles bfs);
+void key_input (ZebraHandle zh, int nkeys, int cache, Res res);
+ISAMS_M *key_isams_m (Res res, ISAMS_M *me);
+ISAMC_M *key_isamc_m (Res res, ISAMC_M *me);
int merge_sort (char **buf, int from, int to);
+int key_SU_code (int ch, char *out);
-#define TEMP_FNAME "keys%d.tmp"
-#define FNAME_WORD_DICT "worddict"
-#define FNAME_WORD_ISAM "wordisam"
+#define FNAME_DICT "dict"
+#define FNAME_ISAM "isam"
+#define FNAME_ISAMC "isamc"
+#define FNAME_ISAMS "isams"
+#define FNAME_ISAMH "isamh"
+#define FNAME_ISAMD "isamd"
#define FNAME_CONFIG "zebra.cfg"
#define GMATCH_DICT "gmatch"
void strtab_del (struct strtab *t,
void (*func)(const char *name, void *info, void *data),
void *data);
-int index_char_cvt (int c);
-int index_word_prefix (char *string, int attset_ordinal,
- int local_attribute, const char *databaseName);
-
-int fileExtract (SYSNO *sysno, const char *fname,
- const struct recordGroup *rGroup, int deleteFlag);
-void rec_prstat (void);
+void zebraIndexLockMsg (ZebraHandle zh, const char *str);
+void zebraIndexUnlock (ZebraHandle zh);
+int zebraIndexLock (BFiles bfs, ZebraHandle zh, int commitNow, const char *rval);
+int zebraIndexWait (ZebraHandle zh, int commitPhase);
-void zebraLockPrefix (char *pathPrefix);
-
-void zebraIndexLockMsg (const char *str);
-void zebraIndexUnlock (void);
-void zebraIndexLock (int commitNow);
-int zebraIndexWait (int commitPhase);
+void zebra_lock_prefix (Res res, char *dst);
#define FNAME_MAIN_LOCK "zebraidx.LCK"
#define FNAME_COMMIT_LOCK "zebracmt.LCK"
#define FNAME_ORG_LOCK "zebraorg.LCK"
#define FNAME_TOUCH_TIME "zebraidx.time"
-int zebraLock (int fd, int wr);
-int zebraLockNB (int fd, int wr);
-int zebraUnlock (int fd);
+int key_SU_decode (int *ch, const unsigned char *out);
+int key_SU_encode (int ch, char *out);
+
+typedef struct zebra_set *ZebraSet;
+
+typedef struct zebra_rank_class {
+ struct rank_control *control;
+ int init_flag;
+ void *class_handle;
+ struct zebra_rank_class *next;
+} *ZebraRankClass;
+
+#include "reckeys.h"
+
+struct zebra_register {
+ char *name;
+
+ ISAMS isams;
+ ISAMC isamc;
+ ISAMB isamb;
+ Dict dict;
+ Dict matchDict;
+ SortIdx sortIdx;
+ int registerState; /* 0 (no commit pages), 1 (use commit pages) */
+ time_t registerChange;
+ BFiles bfs;
+ Records records;
+ ZebraExplainInfo zei;
+
+ char *server_path_prefix;
+ data1_handle dh;
+ ZebraMaps zebra_maps;
+ ZebraRankClass rank_classes;
+ RecTypes recTypes;
+ int seqno;
+ int last_val;
+ int stop_flag;
+
+ zebra_rec_keys_t keys;
+ zebra_rec_keys_t sortKeys;
+ char **key_buf;
+ size_t ptr_top;
+ size_t ptr_i;
+ size_t key_buf_used;
+ int key_file_no;
+};
+
+struct zebra_service {
+ int stop_flag;
+ Res global_res;
+ struct zebra_session *sessions;
+ struct zebra_register *regs;
+ Zebra_mutex_cond session_lock;
+ Passwd_db passwd_db;
+ Res dbaccess;
+ const char *path_root;
+ RecTypeClass record_classes;
+ NMEM nmem;
+};
+
+
+struct zebra_session {
+ struct zebra_session *next;
+ struct zebra_service *service;
+ struct zebra_register *reg;
+
+ char *xadmin_databaseName;
+
+ char **basenames;
+ int num_basenames;
+ zint approx_limit;
+ char *reg_name;
+ char *path_reg;
+
+ ZebraLockHandle lock_normal;
+ ZebraLockHandle lock_shadow;
+
+ int trans_no;
+ int trans_w_no;
+
+ int destroyed;
+ ZebraSet sets;
+ Res res;
+ Res session_res;
+ char *user_perm;
+ char *dbaccesslist;
+ int errCode;
+ zint hits;
+ char *errString;
+#if HAVE_SYS_TIMES_H
+ struct tms tms1;
+ struct tms tms2;
+#endif
+ int shadow_enable;
+
+ int m_staticrank;
+ int m_segment_indexing;
+
+ zint records_inserted;
+ zint records_updated;
+ zint records_deleted;
+ zint records_processed;
+ char *record_encoding;
+
+ yaz_iconv_t iconv_to_utf8;
+ yaz_iconv_t iconv_from_utf8;
+
+ int m_follow_links;
+ const char *m_group;
+ const char *m_record_id;
+ const char *m_record_type;
+ int m_store_data;
+ int m_store_keys;
+ int m_explain_database;
+ int m_flag_rw;
+ int m_file_verbose_limit;
+
+ void *store_data_buf;
+ size_t store_data_size;
+ NMEM nmem_error;
+
+ struct zebra_limit *m_limit;
+};
+
+
+struct term_set_entry {
+ char *term;
+ struct term_set_entry *next;
+};
+
+struct term_set_list {
+ struct term_set_entry *first;
+ struct term_set_entry *last;
+};
+
+
+void zebra_limit_destroy(struct zebra_limit *zl);
+struct zebra_limit *zebra_limit_create(int exclude_flag, zint *ids);
+void zebra_limit_for_rset(struct zebra_limit *zl,
+ int (**filter_func)(const void *buf, void *data),
+ void (**filter_destroy)(void *data),
+ void **filter_data);
+
+struct rset_key_control *zebra_key_control_create(ZebraHandle zh);
+
+ZEBRA_RES rpn_search_top(ZebraHandle zh, Z_RPNStructure *zs,
+ oid_value attributeSet,
+ NMEM stream, NMEM rset_nmem,
+ Z_SortKeySpecList *sort_sequence,
+ int num_bases, char **basenames,
+ RSET *result_set);
+
+ZEBRA_RES rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt,
+ oid_value attributeset,
+ int num_bases, char **basenames,
+ int *position, int *num_entries, ZebraScanEntry **list,
+ int *is_partial, RSET limit_set, int return_zero);
+
+RSET rset_trunc(ZebraHandle zh, ISAM_P *isam_p, int no,
+ const char *term, int length_term, const char *flags,
+ int preserve_position, int term_type, NMEM rset_nmem,
+ struct rset_key_control *kctrl, int scope,
+ struct ord_list *ol, int reg_type,
+ zint hits_limit, const char *term_ref_id);
+
+void resultSetAddTerm(ZebraHandle zh, ZebraSet s, int reg_type,
+ const char *db, const char *index_name,
+ const char *term);
+ZebraSet resultSetAdd (ZebraHandle zh, const char *name, int ov);
+ZebraSet resultSetGet (ZebraHandle zh, const char *name);
+ZEBRA_RES resultSetAddRPN (ZebraHandle zh, NMEM m, Z_RPNQuery *rpn,
+ int num_bases, char **basenames,
+ const char *setname);
+RSET resultSetRef (ZebraHandle zh, const char *resultSetId);
+void resultSetDestroy (ZebraHandle zh, int num_names, char **names,
+ int *statuses);
+ZEBRA_RES resultSetSort (ZebraHandle zh, NMEM nmem,
+ int num_input_setnames, const char **input_setnames,
+ const char *output_setname,
+ Z_SortKeySpecList *sort_sequence, int *sort_status);
+ZEBRA_RES resultSetSortSingle (ZebraHandle zh, NMEM nmem,
+ ZebraSet sset, RSET rset,
+ Z_SortKeySpecList *sort_sequence,
+ int *sort_status);
+ZEBRA_RES resultSetRank (ZebraHandle zh, ZebraSet zebraSet, RSET rset,
+ NMEM nmem);
+void resultSetInvalidate (ZebraHandle zh);
+
+int zebra_server_lock_init (ZebraService zh);
+int zebra_server_lock_destroy (ZebraService zh);
+int zebra_server_lock (ZebraService zh, int lockCommit);
+void zebra_server_unlock (ZebraService zh, int commitPhase);
+int zebra_server_lock_get_state (ZebraService zh, time_t *timep);
+
+int zebra_record_fetch (ZebraHandle zh, SYSNO sysno, int score,
+ zebra_snippets *hit_snippet, ODR stream,
+ oid_value input_format, Z_RecordComposition *comp,
+ oid_value *output_format, char **rec_bufp,
+ int *rec_lenp, char **basenamep,
+ char **addinfo);
+
+void extract_get_fname_tmp (ZebraHandle zh, char *fname, int no);
+
+void zebra_index_merge (ZebraHandle zh);
+
+ZEBRA_RES zebra_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 zebra_extract_record_stream(ZebraHandle zh,
+ struct ZebraRecStream *stream,
+ int delete_flag,
+ int test_mode,
+ const char *recordType,
+ SYSNO *sysno,
+ const char *match_criteria,
+ const char *fname,
+ int force_update,
+ int allow_update,
+ RecType recType,
+ void *recTypeClientData);
+
+#if 0
+int extract_rec_in_mem (ZebraHandle zh, const char *recordType,
+ const char *buf, size_t buf_size,
+ const char *databaseName, int delete_flag,
+ int test_mode, int *sysno,
+ int store_keys, int store_data,
+ const char *match_criteria);
+#endif
+void extract_flushWriteKeys (ZebraHandle zh, int final);
+
+YAZ_EXPORT void zebra_create_stream_mem(struct ZebraRecStream *stream,
+ const char *buf, size_t sz);
+YAZ_EXPORT void zebra_create_stream_fd(struct ZebraRecStream *stream,
+ int fd, off_t start_offset);
+void print_rec_keys(ZebraHandle zh, zebra_rec_keys_t reckeys);
+
+ZEBRA_RES zebra_snippets_rec_keys(ZebraHandle zh, zebra_rec_keys_t reckeys,
+ zebra_snippets *snippets);
+ZEBRA_RES zebra_snippets_hit_vector(ZebraHandle zh, const char *setname,
+ zint sysno, zebra_snippets *snippets);
+
+ZEBRA_RES zebra_extract_explain(void *handle, Record rec, data1_node *n);
+
+ZEBRA_RES zebra_extract_file(ZebraHandle zh, SYSNO *sysno, const char *fname,
+ int deleteFlag);
+
+ZEBRA_RES zebra_begin_read (ZebraHandle zh);
+ZEBRA_RES zebra_end_read (ZebraHandle zh);
+
+int zebra_file_stat (const char *file_name, struct stat *buf,
+ int follow_links);
+
+void zebra_livcode_transform(ZebraHandle zh, Z_RPNQuery *query);
+
+void *iscz1_start ();
+void iscz1_reset (void *vp);
+void iscz1_stop (void *p);
+void iscz1_decode (void *vp, char **dst, const char **src);
+void iscz1_encode (void *vp, char **dst, const char **src);
+
+Dict dict_open_res (BFiles bfs, const char *name, int cache, int rw,
+ int compact_flag, Res res);
+
+void zebra_setError(ZebraHandle zh, int code, const char *addinfo);
+void zebra_setError_zint(ZebraHandle zh, int code, zint i);
+
+void zebra_term_untrans_iconv(ZebraHandle zh, NMEM stream, int reg_type,
+ char **dst, const char *src);
+
+ZEBRA_RES zebra_get_hit_vector(ZebraHandle zh, const char *setname, zint sysno);
+
+void zebra_term_untrans(ZebraHandle zh, int reg_type,
+ char *dst, const char *src);
+
+ZEBRA_RES zebra_apt_get_ord(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ int index_type,
+ const char *xpath_use,
+ oid_value curAttributeSet,
+ int *ord);
+
+ZEBRA_RES zebra_attr_list_get_ord(ZebraHandle zh,
+ Z_AttributeList *attr_list,
+ zinfo_index_category_t cat,
+ int index_type,
+ oid_value curAttributeSet,
+ int *ord);
+
+ZEBRA_RES zebra_sort_get_ord(ZebraHandle zh,
+ Z_SortAttributes *sortAttributes,
+ int *ord,
+ int *numerical);
+
+ZEBRA_RES zebra_update_file_match(ZebraHandle zh, const char *path);
+ZEBRA_RES zebra_update_from_path(ZebraHandle zh, const char *path);
+ZEBRA_RES zebra_delete_from_path(ZebraHandle zh, const char *path);
+
+#define FIRST_IN_FIELD_STR "\001^"
+#define FIRST_IN_FIELD_LEN 2
+
+YAZ_END_CDECL
+
+#endif
+/*
+ * Local variables:
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ * vim: shiftwidth=4 tabstop=8 expandtab
+ */
-void init_charmap(void);
-char **map_chrs_input(char **from, int len);
-const char *map_chrs_output(const char **from);
projects / idzebra-moved-to-github.git / blobdiff