/*
- * Copyright (C) 1994, Index Data I/S
+ * Copyright (C) 1994-1995, Index Data I/S
* All rights reserved.
* Sebastian Hammer, Adam Dickmeiss
*
* $Log: main.c,v $
- * Revision 1.1 1995-08-31 14:50:24 adam
+ * Revision 1.25 1995-12-01 16:24:39 adam
+ * Commit files use separate meta file area.
+ *
+ * Revision 1.24 1995/11/30 17:01:38 adam
+ * New setting commitCache: points to commit directories/files.
+ * New command commit: commits at the end of a zebraidx run.
+ *
+ * Revision 1.23 1995/11/30 08:34:31 adam
+ * Started work on commit facility.
+ * Changed a few malloc/free to xmalloc/xfree.
+ *
+ * Revision 1.22 1995/11/28 09:09:42 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 1.21 1995/11/27 14:27:39 adam
+ * Renamed 'update' command to 'dir'.
+ *
+ * Revision 1.20 1995/11/27 13:58:53 adam
+ * New option -t. storeStore data implemented in server.
+ *
+ * Revision 1.19 1995/11/25 10:24:06 adam
+ * More record fields - they are enumerated now.
+ * New options: flagStoreData flagStoreKey.
+ *
+ * Revision 1.18 1995/11/22 17:19:17 adam
+ * Record management uses the bfile system.
+ *
+ * Revision 1.17 1995/11/21 15:01:16 adam
+ * New general match criteria implemented.
+ * New feature: document groups.
+ *
+ * Revision 1.16 1995/11/20 11:56:27 adam
+ * Work on new traversal.
+ *
+ * Revision 1.15 1995/11/01 16:25:51 quinn
+ * *** empty log message ***
+ *
+ * Revision 1.14 1995/10/17 18:02:09 adam
+ * New feature: databases. Implemented as prefix to words in dictionary.
+ *
+ * Revision 1.13 1995/10/10 12:24:39 adam
+ * Temporary sort files are compressed.
+ *
+ * Revision 1.12 1995/10/04 16:57:20 adam
+ * Key input and merge sort in one pass.
+ *
+ * Revision 1.11 1995/09/29 14:01:45 adam
+ * Bug fixes.
+ *
+ * Revision 1.10 1995/09/28 14:22:57 adam
+ * Sort uses smaller temporary files.
+ *
+ * Revision 1.9 1995/09/14 07:48:24 adam
+ * Record control management.
+ *
+ * Revision 1.8 1995/09/06 16:11:18 adam
+ * Option: only one word key per file.
+ *
+ * Revision 1.7 1995/09/05 15:28:39 adam
+ * More work on search engine.
+ *
+ * Revision 1.6 1995/09/04 12:33:43 adam
+ * Various cleanup. YAZ util used instead.
+ *
+ * Revision 1.5 1995/09/04 09:10:39 adam
+ * More work on index add/del/update.
+ * Merge sort implemented.
+ * Initial work on z39 server.
+ *
+ * Revision 1.4 1995/09/01 14:06:36 adam
+ * Split of work into more files.
+ *
+ * Revision 1.3 1995/09/01 10:57:07 adam
+ * Minor changes.
+ *
+ * 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.
*
*/
#include <stdio.h>
#include <assert.h>
#include <unistd.h>
-#include <dirent.h>
-#include <sys/stat.h>
-#include <sys/types.h>
-#include <fcntl.h>
-#include <ctype.h>
-#include <util.h>
+#include <alexutil.h>
+#include <data1.h>
#include "index.h"
char *prog;
-
-static int key_fd = -1;
-#define KEY_BUF_SIZE 100000
-static char *key_buf;
-int key_offset;
-SYSNO sysno_next;
-Dict file_idx;
-static char *base_path = NULL;
-
-void key_open (const char *fname)
-{
- void *file_key;
- if (key_fd != -1)
- return;
- if ((key_fd = open (fname, O_RDWR|O_CREAT, 0666)) == -1)
- {
- log (LOG_FATAL|LOG_ERRNO, "Creat %s", fname);
- exit (1);
- }
- if (!(key_buf = malloc (KEY_BUF_SIZE)))
- {
- log (LOG_FATAL|LOG_ERRNO, "malloc");
- exit (1);
- }
- key_offset = 0;
- if (!(file_idx = dict_open ("fileidx", 10, 1)))
- {
- log (LOG_FATAL, "dict_open fail of %s", "fileidx");
- exit (1);
- }
- file_key = dict_lookup (file_idx, ".");
- if (file_key)
- memcpy (&sysno_next, (char*)file_key+1, sizeof(sysno_next));
- else
- sysno_next = 1;
-}
-
-void key_close (void)
-{
- if (key_fd == -1)
- return;
- close (key_fd);
- dict_insert (file_idx, ".", sizeof(sysno_next), &sysno_next);
- dict_close (file_idx);
- key_fd = -1;
-}
-
-void key_flush (void)
-{
- size_t i = 0;
- int w;
-
- while (i < key_offset)
- {
- w = write (key_fd, key_buf + i, key_offset - i);
- if (w == -1)
- {
- log (LOG_FATAL|LOG_ERRNO, "Write key fail");
- exit (1);
- }
- i += w;
- }
- key_offset = 0;
-}
-
-void key_write (int cmd, struct it_key *k, const char *str)
-{
- char x = cmd;
- size_t slen = strlen(str);
-
- if (key_offset + sizeof(*k) + slen >= KEY_BUF_SIZE - 2)
- key_flush ();
- memcpy (key_buf + key_offset, &x, 1);
- key_offset++;
- memcpy (key_buf + key_offset, k, sizeof(*k));
- key_offset += sizeof(*k);
- memcpy (key_buf + key_offset, str, slen+1);
- key_offset += slen+1;
-}
-
-void text_extract (SYSNO sysno, int cmd, const char *fname)
-{
- FILE *inf;
- struct it_key k;
- int seqno = 1;
- int c;
- char w[256];
-
- log (LOG_DEBUG, "Text extract of %d", sysno);
- k.sysno = sysno;
- inf = fopen (fname, "r");
- if (!inf)
- {
- log (LOG_WARN|LOG_ERRNO, "open %s", fname);
- return;
- }
- while ((c=getc (inf)) != EOF)
- {
- int i = 0;
- while (i < 254 && c != EOF && isalnum(c))
- {
- w[i++] = c;
- c = getc (inf);
- }
- if (i)
- {
- w[i] = 0;
-
- k.seqno = seqno++;
- k.field = 0;
- key_write (cmd, &k, w);
- }
- if (c == EOF)
- break;
- }
- fclose (inf);
-}
-
-void file_extract (int cmd, struct stat *fs, const char *fname)
-{
- int i;
- char ext[128];
- SYSNO sysno;
- char ext_res[128];
- const char *file_type;
- void *file_info;
-
- log (LOG_DEBUG, "%c %s", cmd, fname);
- for (i = strlen(fname); --i >= 0; )
- if (fname[i] == '/')
- {
- strcpy (ext, "");
- break;
- }
- else if (fname[i] == '.')
- {
- strcpy (ext, fname+i+1);
- break;
- }
- sprintf (ext_res, "fileExtension.%s", ext);
- if (!(file_type = res_get (common_resource, ext_res)))
- return;
-
- file_info = dict_lookup (file_idx, fname);
- if (!file_info)
- {
- sysno = sysno_next++;
- dict_insert (file_idx, fname, sizeof(sysno), &sysno);
- }
- else
- memcpy (&sysno, (char*) file_info+1, sizeof(sysno));
- if (!strcmp (file_type, "text"))
- text_extract (sysno, cmd, fname);
-}
-
-static void repository_extract_r (int cmd, char *rep)
-{
- DIR *dir;
- size_t rep_len;
- struct dirent *dent;
- struct stat fs;
-
- rep_len = strlen(rep);
- dir = opendir(rep);
- if (!dir)
- {
- if (errno == ENOENT)
- {
- log (LOG_WARN|LOG_ERRNO, "opendir %s", rep);
- return;
- }
- log (LOG_FATAL|LOG_ERRNO, "opendir %s", rep);
- exit (1);
- }
- if (rep[rep_len-1] != '/')
- rep[rep_len] = '/';
- else
- --rep_len;
- while ((dent = readdir (dir)))
- {
- strcpy (rep +rep_len+1, dent->d_name);
- stat (rep, &fs);
- switch (fs.st_mode & S_IFMT)
- {
- case S_IFREG:
- file_extract (cmd, &fs, rep);
- break;
- case S_IFDIR:
- if (strcmp (dent->d_name, ".") && strcmp(dent->d_name, ".."))
- repository_extract_r (cmd, rep);
- break;
- }
- }
- closedir (dir);
-}
-
-void repository_update_r (int cmd, const char *rep, const char *with_rep)
-{
-
-}
-
-void repository_traverse (int cmd, const char *rep)
-{
- char rep_tmp1[2048];
- char rep_tmp2[2048];
-
- strcpy (rep_tmp1, rep);
- if (base_path)
- {
- strcpy (rep_tmp2, base_path);
- repository_update_r (cmd, rep_tmp1, rep_tmp2);
- }
- else
- repository_extract_r (cmd, rep_tmp1);
-}
-
+size_t mem_max = 4*1024*1024;
+extern char *data1_tabpath;
int main (int argc, char **argv)
{
+ int commit_at_end = 0;
int ret;
int cmd = 0;
char *arg;
- char *base_name;
+ char *configName = NULL;
+ int nsections;
+ int key_open_flag = 0;
+
+ struct recordGroup rGroupDef;
+
+ rGroupDef.groupName = NULL;
+ rGroupDef.databaseName = NULL;
+ rGroupDef.path = NULL;
+ rGroupDef.recordId = NULL;
+ rGroupDef.recordType = NULL;
+ rGroupDef.flagStoreData = -1;
+ rGroupDef.flagStoreKeys = -1;
prog = *argv;
- while ((ret = options ("b:v:", argv, argc, &arg)) != -2)
+ if (argc < 2)
+ {
+ fprintf (stderr, "zebraidx [options] command <dir> ...\n"
+ "Commands:\n"
+ " update <dir> Update index with files below <dir>.\n"
+ " If <dir> is empty filenames are read from stdin.\n"
+ " delete <dir> Delete index with files below <dir>.\n"
+ "Options:\n"
+ " -t <type> Index files as <type> (grs or text).\n"
+ " -c <config> Read configuration file <config>.\n"
+ " -g <group> Index files according to group settings.\n"
+ " -d <database> Records belong to Z39.50 database <database>.\n"
+ " -m <mbytes> Use <mbytes> before flushing keys to disk.\n"
+ " -v <level> Set logging to <level>.\n");
+ exit (1);
+ }
+ while ((ret = options ("t:c:g:d:m:v:", argv, argc, &arg)) != -2)
{
if (ret == 0)
{
- if (!base_name)
+ if(cmd == 0) /* command */
{
- base_name = arg;
-
- common_resource = res_open (base_name);
if (!common_resource)
{
- log (LOG_FATAL, "Cannot open resource `%s'", base_name);
- exit (1);
- }
- }
- else if(cmd == 0) /* command */
- {
- if (!strcmp (arg, "add"))
- {
- cmd = 'a';
+ const char *rval;
+ common_resource = res_open (configName ?
+ configName : FNAME_CONFIG);
+ if (!common_resource)
+ {
+ logf (LOG_FATAL, "Cannot open resource `%s'",
+ configName);
+ exit (1);
+ }
+ data1_tabpath = res_get (common_resource, "profilePath");
+ rval = res_get (common_resource, "commitEnable");
+ if (rval && atoi(rval))
+ bf_cache ();
}
- else if (!strcmp (arg, "del"))
- {
+ if (!strcmp (arg, "update"))
+ cmd = 'u';
+ else if (!strcmp (arg, "del") || !strcmp(arg, "delete"))
cmd = 'd';
- }
+ else if (!strcmp (arg, "commit"))
+ commit_at_end = 1;
else
{
- log (LOG_FATAL, "Unknown command: %s", arg);
+ logf (LOG_FATAL, "Unknown command: %s", arg);
exit (1);
}
}
else
{
- key_open ("keys.tmp");
- repository_traverse (cmd, arg);
+ struct recordGroup rGroup;
+
+ memcpy (&rGroup, &rGroupDef, sizeof(rGroup));
+ if (!key_open_flag)
+ {
+ key_open (mem_max);
+ key_open_flag = 1;
+ }
+ rGroup.path = arg;
+ if (cmd == 'u')
+ repositoryUpdate (&rGroup);
+ else if (cmd == 'd')
+ repositoryDelete (&rGroup);
cmd = 0;
}
}
{
log_init (log_mask_str(arg), prog, NULL);
}
- else if (ret == 'b')
+ else if (ret == 'm')
+ {
+ mem_max = 1024*1024*atoi(arg);
+ }
+ else if (ret == 'd')
+ {
+ rGroupDef.databaseName = arg;
+ }
+ else if (ret == 'g')
{
- base_path = arg;
+ rGroupDef.groupName = arg;
}
+ else if (ret == 'c')
+ configName = arg;
+ else if (ret == 't')
+ rGroupDef.recordType = arg;
else
{
- log (LOG_FATAL, "Unknown option '-%s'", arg);
+ logf (LOG_FATAL, "Unknown option '-%s'", arg);
exit (1);
}
}
- key_flush ();
- key_close ();
+ if (key_open_flag)
+ {
+ nsections = key_close ();
+ if (nsections)
+ {
+ logf (LOG_LOG, "Merging with index");
+ key_input (FNAME_WORD_DICT, FNAME_WORD_ISAM, nsections, 60);
+ }
+ }
+ if (commit_at_end)
+ {
+ logf (LOG_LOG, "commiting");
+ bf_commit ();
+ }
exit (0);
}
+
projects / idzebra-moved-to-github.git / blobdiff