Work on character mapping. Implemented replace rules.

[idzebra-moved-to-github.git] / index / extract.c

diff --git a/index/extract.c b/index/extract.c
index 3e29b83..3a9c4b1 100644 (file)
--- a/index/extract.c
+++ b/index/extract.c
@@ -4,7 +4,26 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: extract.c,v $
- * Revision 1.92  1999-03-09 16:27:49  adam
+ * Revision 1.98  1999-09-07 07:19:21  adam
+ * Work on character mapping. Implemented replace rules.
+ *
+ * Revision 1.97  1999/07/06 12:28:04  adam
+ * Updated record index structure. Format includes version ID. Compression
+ * algorithm ID is stored for each record block.
+ *
+ * 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
@@ -394,24 +413,32 @@ int key_open (struct recordGroup *rGroup, int mem)
     BFiles bfs = rGroup->bfs;
     int rw = rGroup->flagRw;
     data1_handle dh = rGroup->dh;
+    char *recordCompression;
+    int record_compression = REC_COMPRESS_NONE;
     if (!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;
     }
     assert (!records);
-    records = rec_open (bfs, rw);
+    recordCompression = res_get_def (common_resource,
+                                    "recordCompression", "none");
+    if (!strcmp (recordCompression, "none"))
+       record_compression = REC_COMPRESS_NONE;
+    if (!strcmp (recordCompression, "bzip2"))
+       record_compression = REC_COMPRESS_BZIP2;
+    records = rec_open (bfs, rw, record_compression);
     if (!records)
     {
        dict_close (matchDict);
@@ -625,6 +652,7 @@ static void wordInit (struct recExtractCtrl *p, RecWord *w)
     w->attrSet = VAL_BIB1;
     w->attrUse = 1016;
     w->reg_type = 'w';
+    w->extractCtrl = p;
 }
 
 static struct sortKey {
@@ -657,7 +685,7 @@ 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);
@@ -719,11 +747,11 @@ 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);
 
@@ -832,6 +860,13 @@ static void addCompleteField (RecWord *p)
 
 static void addRecordKey (RecWord *p)
 {
+    WRBUF wrbuf;
+    if ((wrbuf = zebra_replace(p->zebra_maps, p->reg_type, 0,
+                              p->string, p->length)))
+    {
+       p->string = wrbuf_buf(wrbuf);
+       p->length = wrbuf_len(wrbuf);
+    }
     if (zebra_maps_is_complete (p->zebra_maps, p->reg_type))
        addCompleteField (p);
     else
@@ -987,7 +1022,8 @@ struct file_read_info {
 
 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;
@@ -1004,7 +1040,7 @@ 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;
@@ -1013,13 +1049,13 @@ static off_t file_seek (void *handle, off_t offset)
 
 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;
     if (p->sdrbuf)
@@ -1043,7 +1079,7 @@ 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->sdrbuf && p->file_moffset)
@@ -1053,7 +1089,7 @@ static void file_begin (void *handle)
 
 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;
@@ -1217,7 +1253,7 @@ 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)
@@ -1235,7 +1271,7 @@ 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;
@@ -1266,8 +1302,8 @@ static int recordExtract (SYSNO *sysno, const char *fname,
        extractCtrl.fh = fi;
        extractCtrl.subType = subType;
        extractCtrl.init = wordInit;
-       extractCtrl.addWord = addRecordKey;
-       extractCtrl.addSchema = addSchema;
+       extractCtrl.tokenAdd = addRecordKey;
+       extractCtrl.schemaAdd = addSchema;
        extractCtrl.dh = rGroup->dh;
        for (i = 0; i<256; i++)
        {
@@ -1287,18 +1323,20 @@ 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;
         }
@@ -1480,7 +1518,8 @@ 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",
@@ -1526,6 +1565,7 @@ 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));
    
@@ -1565,7 +1605,8 @@ 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;
@@ -1602,7 +1643,7 @@ int fileExtract (SYSNO *sysno, const char *fname,
     {
         if (zebraExplain_newDatabase (zti, rGroup->databaseName,
                                      rGroup->explainDatabase))
-            abort ();
+           return 0;
     }
 
     if (rGroup->flagStoreData == -1)
@@ -1704,7 +1745,7 @@ 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)
@@ -1730,8 +1771,8 @@ static int explain_extract (void *handle, Record rec, data1_node *n)
     reckeys.prevSeqNo = 0;
     
     extractCtrl.init = wordInit;
-    extractCtrl.addWord = addRecordKey;
-    extractCtrl.addSchema = addSchema;
+    extractCtrl.tokenAdd = addRecordKey;
+    extractCtrl.schemaAdd = addSchema;
     extractCtrl.dh = rGroup->dh;
     for (i = 0; i<256; i++)
        extractCtrl.seqno[i] = 0;