/[cvs]/nfo/perl/libs/Data/Transfer/Sync/Core.pm
ViewVC logotype

Diff of /nfo/perl/libs/Data/Transfer/Sync/Core.pm

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.1 by joko, Sun Jan 19 01:23:04 2003 UTC revision 1.4 by joko, Sun Feb 9 05:01:10 2003 UTC
# Line 6  Line 6 
6  ##  ##
7  ##    ----------------------------------------------------------------------------------------  ##    ----------------------------------------------------------------------------------------
8  ##    $Log$  ##    $Log$
9    ##    Revision 1.4  2003/02/09 05:01:10  joko
10    ##    + major structure changes
11    ##    - refactored code to sister modules
12    ##
13    ##    Revision 1.3  2003/01/20 17:01:14  joko
14    ##    + cosmetics and debugging
15    ##    + probably refactored code to new plugin-modules 'Metadata.pm' and/or 'StorageInterface.pm' (guess it was the last one...)
16    ##
17    ##    Revision 1.2  2003/01/19 02:05:42  joko
18    ##    - removed pod-documentation: now in Data/Transfer/Sync.pod
19    ##
20  ##    Revision 1.1  2003/01/19 01:23:04  joko  ##    Revision 1.1  2003/01/19 01:23:04  joko
21  ##    + new from Data/Transfer/Sync.pm  ##    + new from Data/Transfer/Sync.pm
22  ##  ##
# Line 69  use mixin::with qw( Data::Transfer::Sync Line 80  use mixin::with qw( Data::Transfer::Sync
80  use Data::Dumper;  use Data::Dumper;
81    
82  use misc::HashExt;  use misc::HashExt;
 use libp qw( md5_base64 );  
 use libdb qw( quotesql hash2Sql );  
 use Data::Transform::Deep qw( hash2object refexpr2perlref );  
83  use Data::Compare::Struct qw( getDifference isEmpty );  use Data::Compare::Struct qw( getDifference isEmpty );
84  use Data::Storage::Container;  use Data::Storage::Container;
85  use DesignPattern::Object;  use DesignPattern::Object;
# Line 111  sub _init { Line 119  sub _init {
119    
120  sub _initV1 {  sub _initV1 {
121    my $self = shift;    my $self = shift;
122      $logger->debug( __PACKAGE__ . "->_initV1" );
123      die("this should not be reached!");
124    # tag storages with id-authority and checksum-provider information    # tag storages with id-authority and checksum-provider information
125    # TODO: better store tag inside metadata to hold bits together!    # TODO: better store tag inside metadata to hold bits together!
126    map { $self->{container}->{storage}->{$_}->{isIdentAuthority} = 1 } @{$self->{id_authorities}};    map { $self->{container}->{storage}->{$_}->{isIdentAuthority} = 1 } @{$self->{id_authorities}};
# Line 119  sub _initV1 { Line 129  sub _initV1 {
129  }  }
130    
131    
 sub _preCheckOptions {  
   
   my $self = shift;  
   my $opts = shift;  
   
 #print Dumper($opts);  
 #exit;  
   
   # the type of the to-be-synced item  
   if (!$opts->{source}->{nodeType}) {  
     $logger->error( __PACKAGE__ . "->_preCheckOptions failed: Please specify \"source-type\".");  
     return;  
   }  
   # the name of the (container-) node the items are listed in  
   if (!$opts->{source}->{nodeName}) {  
     $logger->error( __PACKAGE__ . "->_preCheckOptions failed: Please specify \"source-node\".");  
     return;  
   }  
   
   # a "map"-declaration which module to use for mapping- and/or lookup-purposes  
   if (!$opts->{map}) {  
     $logger->warning( __PACKAGE__ . "->_preCheckOptions: No mapping supplied - please check key 'map|mappings' in global configuration or specify additional argument '--mapping-module'.");  
     return;  
   }  
   if (!$opts->{map}->{moduleName}) {  
     $logger->warning( __PACKAGE__ . "->_preCheckOptions: Currently only perl-modules can provide mappings: Please specify one with '--mapping-module'.");  
     return;  
   }  
   
   return 1;  
   
 }  
   
   
   
   
   
   
   
   
   
 sub _buildFieldmappingV1 {  
   my $self = shift;  
   
   # build mapping  
   # incoming: and Array of node map entries (Array or Hash) - e.g.  
   #   [ 'source:item_name' => 'target:class_val' ]  
   #   { source => 'event->startDateTime', target => 'begindate' }  
   foreach (@{$self->{args}->{mapping}}) {  
     if (ref $_ eq 'ARRAY') {  
       my @entry1 = split(':', $_->[0]);  
       my @entry2 = split(':', $_->[1]);  
       my $descent = [];  
       my $node = [];  
       $descent->[0] = $entry1[0];  
       $descent->[1] = $entry2[0];  
       $node->[0] = $entry1[1];  
       $node->[1] = $entry2[1];  
       push @{$self->{meta}->{$descent->[0]}->{childnodes}}, $node->[0];  
       push @{$self->{meta}->{$descent->[1]}->{childnodes}}, $node->[1];  
     } elsif (ref $_ eq 'HASH') {  
       foreach my $entry_key (keys %$_) {  
         my $entry_val = $_->{$entry_key};  
         push @{$self->{meta}->{$entry_key}->{childnodes}}, $entry_val;  
       }  
     }  
   
   }  
   
 }  
   
 sub _buildMetadataV1 {  
   my $self = shift;  
   
   # decompose identifiers for each partner  
   # TODO: refactor!!! take this list from already established/given metadata  
   foreach ('source', 'target') {  
       
     # get/set metadata for further processing  
   
     # Partner and Node (e.g.: "L:Country" or "R:countries.csv")  
     if (my $item = $self->{args}->{$_}) {  
       my @item = split(':', $item);  
       $self->{meta}->{$_}->{dbkey} = $item[0];  
       $self->{meta}->{$_}->{node} = $item[1];  
     }  
       
     # Filter  
     if (my $item_filter = $self->{args}->{$_ . '_filter'}) {  
       $self->{meta}->{$_}->{filter} = $item_filter;  
     }  
   
     # IdentProvider  
     if (my $item_ident = $self->{args}->{$_ . '_ident'}) {  
       my @item_ident = split(':', $item_ident);  
       $self->{meta}->{$_}->{IdentProvider} = { method => $item_ident[0], arg => $item_ident[1] };  
     }  
   
 #print Dumper($self->{meta});  
   
     # TODO: ChecksumProvider  
   
     # exclude properties/subnodes  
     if (my $item_exclude = $self->{args}->{$_ . '_exclude'}) {  
       $self->{meta}->{$_}->{subnodes_exclude} = $item_exclude;  
     }  
       
     # TypeProvider  
     if (my $item_type = $self->{args}->{$_ . '_type'}) {  
       my @item_type = split(':', $item_type);  
       $self->{meta}->{$_}->{TypeProvider} = { method => $item_type[0], arg => $item_type[1] };  
     }  
       
     # Callbacks - writers (will be triggered _before_ writing to target)  
     if (my $item_writers = $self->{args}->{$_ . '_callbacks_write'}) {  
       my $descent = $_;     # this is important since the following code inside the map wants to use its own context variables  
       map { $self->{meta}->{$descent}->{Callback}->{write}->{$_}++; } @$item_writers;  
     }  
       
     # Callbacks - readers (will be triggered _after_ reading from source)  
     if (my $item_readers = $self->{args}->{$_ . '_callbacks_read'}) {  
       my $descent = $_;  
       map { $self->{meta}->{$descent}->{Callback}->{read}->{$_}++; } @$item_readers;  
     }  
       
     # resolve storage objects  
     #$self->{$_} = $self->{container}->{storage}->{$self->{meta}->{$_}->{dbkey}};  
     # relink references to metainfo  
     $self->{meta}->{$_}->{storage} = $self->{container}->{storage}->{$self->{meta}->{$_}->{dbkey}};  
     #print "iiiiisprov: ", Dumper($self->{meta}->{$_}->{storage}), "\n";  
   }  
   
 }  
   
132  # TODO: abstract the hardwired use of "source" and "target" in here somehow - hmmmm.......  /(="§/%???  # TODO: abstract the hardwired use of "source" and "target" in here somehow - hmmmm.......  /(="§/%???
133  sub _syncNodes {  sub _run {
134    
135    my $self = shift;    my $self = shift;
136        
137      #print "verbose: ", $self->{verbose}, "\n";
138      $self->{verbose} = 1;
139      
140      $logger->debug( __PACKAGE__ . "->_run" );
141    
142      # for statistics
143    my $tc = OneLineDumpHash->new( {} );    my $tc = OneLineDumpHash->new( {} );
144    my $results;    my $results;
145        
# Line 273  sub _syncNodes { Line 155  sub _syncNodes {
155    }    }
156        
157    # get reference to node list from convenient method provided by CORE-HANDLE    # get reference to node list from convenient method provided by CORE-HANDLE
   #$results ||= $self->{source}->getListUnfiltered($self->{meta}->{source}->{node});  
   #$results ||= $self->{meta}->{source}->{storage}->getListUnfiltered($self->{meta}->{source}->{node});  
158    $results ||= $self->_getNodeList('source');    $results ||= $self->_getNodeList('source');
159    
160    # checkpoint: do we actually have a list to iterate through?    # checkpoint: do we actually have a list to iterate through?
161    if (!$results || !@{$results}) {    if (!$results || !@{$results}) {
162      $logger->notice( __PACKAGE__ . "->syncNodes: No nodes to synchronize." );      $logger->notice( __PACKAGE__ . "->_run: No nodes to synchronize." );
163      return;      return;
164    }    }
165    
166    
167    
168      # check if we actually *have* a synchronization method
169      if (!$self->{options}->{metadata}->{syncMethod}) {
170        $logger->critical( __PACKAGE__ . "->_run: No synchronization method (checksum|timestamp) specified" );
171        return;
172      }
173        
174        
175    # dereference    # dereference
176    my @results = @{$results};    my @results = @{$results};
# Line 318  sub _syncNodes { Line 207  sub _syncNodes {
207          my $value; # = $source_node->{$node};          my $value; # = $source_node->{$node};
208    
209          # ------------  half-redundant: make $self->callCallback($object, $value, $opts)          # ------------  half-redundant: make $self->callCallback($object, $value, $opts)
210          my $perl_callback = $self->{meta}->{$descent}->{node} . '::' . $node . '_read';          #my $perl_callback = $self->{meta}->{$descent}->{node} . '::' . $node . '_read';
211            my $perl_callback = $self->{meta}->{$descent}->{nodeType} . '::' . $node . '_read';
212          my $evalstring = 'return ' . $perl_callback . '( { object => $object, property => $node, value => $value, storage => $self->{meta}->{$descent}->{storage} } );';          my $evalstring = 'return ' . $perl_callback . '( { object => $object, property => $node, value => $value, storage => $self->{meta}->{$descent}->{storage} } );';
213          #print $evalstring, "\n"; exit;          #print $evalstring, "\n"; exit;
214          my $cb_result = eval($evalstring);          my $cb_result = eval($evalstring);
# Line 351  sub _syncNodes { Line 241  sub _syncNodes {
241      #if (!$identOK && lc $self->{args}->{direction} ne 'import') {      #if (!$identOK && lc $self->{args}->{direction} ne 'import') {
242      if (!$identOK) {      if (!$identOK) {
243        #print Dumper($self->{meta}->{source});        #print Dumper($self->{meta}->{source});
244        $logger->critical( __PACKAGE__ . "->syncNodes: No ident found in source node \"$self->{meta}->{source}->{node}\", try to \"prepare\" this node first?" );        $logger->critical( __PACKAGE__ . "->_run: No ident found in source node ( nodeName='$self->{meta}->{source}->{nodeName}', nodeType='$self->{meta}->{source}->{nodeType}') try to \"prepare\" this node first?" );
245        return;        return;
246      }      }
247    
248  #print "statload", "\n";      #print "l" if $self->{verbose};
 #print "ident: ", $self->{node}->{source}->{ident}, "\n";  
 #print Dumper($self->{node});  
       
249      my $statOK = $self->_statloadNode('target', $self->{node}->{source}->{ident});      my $statOK = $self->_statloadNode('target', $self->{node}->{source}->{ident});
250    
 #print Dumper($self->{node});  
       
251      # mark node as new either if there's no ident or if stat/load failed      # mark node as new either if there's no ident or if stat/load failed
252      if (!$statOK) {      if (!$statOK) {
253        $self->{node}->{status}->{new} = 1;        $self->{node}->{status}->{new} = 1;
# Line 371  sub _syncNodes { Line 256  sub _syncNodes {
256    
257  #print "checksum", "\n";  #print "checksum", "\n";
258            
259        print ":" if $self->{verbose};
260    
261        #print Dumper($self);
262        
263      # determine status of entry by synchronization method      # determine status of entry by synchronization method
264      if ( (lc $self->{args}->{method} eq 'checksum') ) {      if ( lc $self->{options}->{metadata}->{syncMethod} eq 'checksum' ) {
265      #if ( $statOK && (lc $self->{args}->{method} eq 'checksum') ) {      #if ( $statOK && (lc $self->{args}->{method} eq 'checksum') ) {
266      #if ( !$self->{node}->{status}->{new} && (lc $self->{args}->{method} eq 'checksum') ) {      #if ( !$self->{node}->{status}->{new} && (lc $self->{args}->{method} eq 'checksum') ) {
267                
268        # TODO:      # calculate checksum of source node
       # is this really worth a "critical"???  
       # no - it should just be a debug appendix i believe  
   
 #print "readcs", "\n";  
         
       # calculate checksum of source node  
269        #$self->_calcChecksum('source');        #$self->_calcChecksum('source');
270        if (!$self->_readChecksum('source')) {        if (!$self->_readChecksum('source')) {
271          $logger->critical( __PACKAGE__ . "->_readChecksum: Could not find \"source\" entry with ident=\"$self->{node}->{source}->{ident}\"" );          $logger->warning( __PACKAGE__ . "->_run: Could not find \"source\" entry with ident=\"$self->{node}->{source}->{ident}\"" );
272          $tc->{skip}++;          $tc->{skip}++;
273          print "s" if $self->{verbose};          print "s" if $self->{verbose};
274          next;          next;
# Line 407  sub _syncNodes { Line 290  sub _syncNodes {
290        # determine if entry is "new" or "dirty"        # determine if entry is "new" or "dirty"
291        # after all, this seems to be the point where the hammer falls.....        # after all, this seems to be the point where the hammer falls.....
292        print "c" if $self->{verbose};        print "c" if $self->{verbose};
293    #print Dumper($self->{node});
294    #exit;
295        $self->{node}->{status}->{new} = !$self->{node}->{target}->{checksum};        $self->{node}->{status}->{new} = !$self->{node}->{target}->{checksum};
296        if (!$self->{node}->{status}->{new}) {        if (!$self->{node}->{status}->{new}) {
297          $self->{node}->{status}->{dirty} =          $self->{node}->{status}->{dirty} =
# Line 416  sub _syncNodes { Line 301  sub _syncNodes {
301            $self->{args}->{force};            $self->{args}->{force};
302        }        }
303    
304        } else {
305          $logger->warning( __PACKAGE__ . "->_run: Synchronization method '$self->{options}->{metadata}->{syncMethod}' is not implemented" );
306          $tc->{skip}++;
307          print "s" if $self->{verbose};
308          next;
309          
310      }      }
311    
312      # first reaction on entry-status: continue with next entry if the current is already "in sync"      # first reaction on entry-status: continue with next entry if the current is already "in sync"
# Line 425  sub _syncNodes { Line 316  sub _syncNodes {
316      }      }
317    
318      # build map to actually transfer the data from source to target      # build map to actually transfer the data from source to target
319      $self->_buildMap();      $self->buildAttributeMap();
320    
321    
322  #print Dumper($self->{node}); exit;  #print Dumper($self->{node}); exit;
# Line 436  sub _syncNodes { Line 327  sub _syncNodes {
327      if ($self->{meta}->{target}->{storage}->{isWriteProtected}) {      if ($self->{meta}->{target}->{storage}->{isWriteProtected}) {
328        $tc->{attempt_transfer}++;        $tc->{attempt_transfer}++;
329        print "\n" if $self->{verbose};        print "\n" if $self->{verbose};
330        $logger->notice( __PACKAGE__ . "->syncNodes: Target is write-protected. Will not insert or modify node. " .        $logger->notice( __PACKAGE__ . "->_run: Target is write-protected. Will not insert or modify node. " .
331            "(Ident: $self->{node}->{source}->{ident} " . "Dump:\n" . Dumper($self->{node}->{source}->{payload}) . ")" );            "(Ident: $self->{node}->{source}->{ident} " . "Dump:\n" . Dumper($self->{node}->{source}->{payload}) . ")" );
332        print "\n" if $self->{verbose};        print "\n" if $self->{verbose};
333        $tc->{skip}++;        $tc->{skip}++;
334        next;        next;
335      }      }
336    
337    #print Dumper($self);
338    #exit;
339    
340      # transfer contents of map to target      # transfer contents of map to target
341      if ($self->{node}->{status}->{new}) {      if ($self->{node}->{status}->{new}) {
342        $tc->{attempt_new}++;        $tc->{attempt_new}++;
# Line 471  sub _syncNodes { Line 365  sub _syncNodes {
365        print "e" if $self->{verbose};        print "e" if $self->{verbose};
366      }      }
367            
368        # trace
369          #print Dumper($self);
370          #exit;
371        
372      # change ident in source (take from target), if transfer was ok and target is an IdentAuthority      # change ident in source (take from target), if transfer was ok and target is an IdentAuthority
373      # this is (for now) called a "retransmit" indicated by a "r"-character when verbosing      # this is (for now) called a "retransmit" indicated by a "r"-character when verbosing
374      if ($self->{node}->{status}->{ok} && $self->{meta}->{target}->{storage}->{isIdentAuthority}) {      if ($self->{node}->{status}->{ok} && $self->{options}->{target}->{storage}->{idAuthority}) {
375        print "r" if $self->{verbose};        print "r" if $self->{verbose};
376        #print Dumper($self->{meta});        #print Dumper($self->{meta});
377        #print Dumper($self->{node});        #print Dumper($self->{node});
# Line 481  sub _syncNodes { Line 379  sub _syncNodes {
379        $self->_doModifySource_IdentChecksum($self->{node}->{target}->{ident});        $self->_doModifySource_IdentChecksum($self->{node}->{target}->{ident});
380      }      }
381    
     print ":" if $self->{verbose};  
   
382    }    }
383    
384    print "\n" if $self->{verbose};    print "\n" if $self->{verbose};
# Line 498  sub _syncNodes { Line 394  sub _syncNodes {
394            
395      # todo!!!      # todo!!!
396      #sysevent( { usermsg => $msg, level => $level }, $taskEvent );      #sysevent( { usermsg => $msg, level => $level }, $taskEvent );
397      $logger->info( __PACKAGE__ . "->syncNodes: $msg" );      $logger->info( __PACKAGE__ . "->_run: $msg" );
398    
399    return $tc;    return $tc;
400    
# Line 549  sub _dumpCompact { Line 445  sub _dumpCompact {
445  }  }
446    
447    
 sub _calcChecksum {  
   
   my $self = shift;  
   my $descent = shift;  
   my $specifier = shift;  
   
   # calculate checksum for current object  
     my $ident = $self->{node}->{$descent}->{ident};  
     
   # build dump of this node  
     my $payload = $self->{node}->{$descent}->{payload};  
     #my $dump = $ident . "\n" . $item->quickdump();  
     #my $dump = $ident . "\n" . Dumper($item);  
     my $dump = $ident . "\n" . $self->_dumpCompact($payload);  
     
   # TODO: $logger->dump( ... );  
     #$logger->debug( __PACKAGE__ . ": " . $dump );  
     #$logger->dump( __PACKAGE__ . ": " . $dump );  
     
   # calculate checksum from dump  
     # note: the 32-bit integer hash from DBI seems  
     # to generate duplicates with small payloads already in ranges of hundreds of items/rows!!!  
     # try to avoid to use it or try to use it only for payloads greater than, hmmm, let's say 30 chars?  
     # (we had about 15 chars average per item (row))  
   
     # md5-based fingerprint, base64 encoded (from Digest::MD5)  
       $self->{node}->{$descent}->{checksum} = md5_base64($dump) . '==';  
     # 32-bit integer "hash" value (maybe faster?) (from DBI)  
       #$self->{node}->{$descent}->{checksum} = DBI::hash($dump, 1);  
   
   # signal good  
   return 1;  
   
 }  
   
   
 sub _readChecksum {  
   my $self = shift;  
   
   my $descent = shift;  
   
   #print "getcheck:", "\n"; print Dumper($self->{node}->{$descent});  
     
   if (!$self->{node}->{$descent}) {  
     # signal checksum bad  
     return;  
   }  
   
   # get checksum for current entry  
   # TODO: don't have the checksum column/property hardcoded as "cs" here, make this configurable somehow  
   
   if ($self->{meta}->{$descent}->{storage}->{isChecksumAuthority}) {  
     #$self->{node}->{$descent}->{checksum} = $entry->{cs};  
     #$self->{node}->{$descent}->{checksum} = $self->_calcChecksum($descent); # $entry->{cs};  
     #print "descent: $descent", "\n";  
     $self->_calcChecksum($descent);  
     #print "checksum: ", $self->{node}->{$descent}->{checksum}, "\n";  
   } else {  
   
     #$self->{node}->{$descent}->{checksum} = $entry->{cs};  
     $self->{node}->{$descent}->{checksum} = $self->{node}->{$descent}->{payload}->{cs};  
   }  
   
   # signal checksum good  
   return 1;  
   
 }  
   
   
 sub _buildMap {  
   
  my $self = shift;  
   
   # field-structure for building sql  
   # mapping of sql-fieldnames to object-attributes  
     $self->{node}->{map} = {};  
   
     # manually set ...  
       # ... object-id  
       $self->{node}->{map}->{$self->{meta}->{target}->{IdentProvider}->{arg}} = $self->{node}->{source}->{ident};  
       # ... checksum  
       $self->{node}->{map}->{cs} = $self->{node}->{source}->{checksum};  
   
 #print "sqlmap: ", Dumper($self->{node}->{map}), "\n";  
   
     # for transferring flat structures via simple (1:1) mapping  
     # TODO: diff per property / property value  
   
     if ($self->{args}->{mapping}) {  
       # apply mapping from $self->{args}->{mapping} to $self->{node}->{map}  
       #foreach my $key (@{$self->{meta}->{source}->{childnodes}}) {  
       my @childnodes = @{$self->{meta}->{source}->{childnodes}};  
       for (my $mapidx = 0; $mapidx <= $#childnodes; $mapidx++) {  
         #my $map_right = $self->{args}->{mapping}->{$key};  
           
         $self->{node}->{source}->{propcache} = {};  
         $self->{node}->{target}->{propcache} = {};  
           
         # get property name  
         $self->{node}->{source}->{propcache}->{property} = $self->{meta}->{source}->{childnodes}->[$mapidx];  
         $self->{node}->{target}->{propcache}->{property} = $self->{meta}->{target}->{childnodes}->[$mapidx];  
         #print "map: $map_right", "\n";  
   
         # get property value  
         my $value;  
           
         # detect for callback - old style - (maybe the better???)  
         if (ref($self->{node}->{target}->{map}) eq 'CODE') {  
           #$value = &$map_right($objClone);  
         } else {  
           # plain (scalar?) value  
           #$value = $objClone->{$map_right};  
           $self->{node}->{source}->{propcache}->{value} = $self->{node}->{source}->{payload}->{$self->{node}->{source}->{propcache}->{property}};  
         }  
         #$self->{node}->{map}->{$key} = $value;  
           
         # detect expression  
         # for transferring deeply nested structures described by expressions  
         #print "val: $self->{node}->{source}->{propcache}->{value}", "\n";  
         if ($self->{node}->{source}->{propcache}->{property} =~ s/^expr://) {  
             
           # create an anonymous sub to act as callback target dispatcher  
             my $cb_dispatcher = sub {  
               #print "===============  CALLBACK DISPATCHER", "\n";  
               #print "ident: ", $self->{node}->{source}->{ident}, "\n";  
               #return $self->{node}->{source}->{ident};  
                 
             };  
             
   
 #print Dumper($self->{node});  
             
           # build callback map for helper function  
           #my $cbmap = { $self->{meta}->{source}->{IdentProvider}->{arg} => $cb_dispatcher };  
           my $cbmap = {};  
           my $value = refexpr2perlref($self->{node}->{source}->{payload}, $self->{node}->{source}->{propcache}->{property}, $cbmap);  
           $self->{node}->{source}->{propcache}->{value} = $value;  
         }  
   
         # encode values dependent on type of underlying storage here - expand cases...  
         my $storage_type = $self->{meta}->{target}->{storage}->{locator}->{type};  
         if ($storage_type eq 'DBI') {  
           # ...for sql  
           $self->{node}->{source}->{propcache}->{value} = quotesql($self->{node}->{source}->{propcache}->{value});  
         }  
          elsif ($storage_type eq 'Tangram') {  
           # iso? utf8 already possible?  
           
         } elsif ($storage_type eq 'LDAP') {  
           # TODO: encode utf8 here?  
         }  
   
         # store value to transfer map  
         $self->{node}->{map}->{$self->{node}->{target}->{propcache}->{property}} = $self->{node}->{source}->{propcache}->{value};  
   
       }  
     }  
   
       
   # TODO: $logger->dump( ... );  
   #$logger->debug( "sqlmap:" . "\n" . Dumper($self->{node}->{map}) );  
 #print "sqlmap: ", Dumper($self->{node}->{map}), "\n";  
 #print "entrystatus: ", Dumper($self->{node}), "\n";  
   
 }  
   
 sub _resolveNodeIdent {  
   my $self = shift;  
   my $descent = shift;  
     
   #print Dumper($self->{node}->{$descent});  
     
   # get to the payload  
     #my $item = $specifier->{item};  
     my $payload = $self->{node}->{$descent}->{payload};  
   
   # resolve method to get to the id of the given item  
   # we use global metadata and the given descent for this task  
     #my $ident = $self->{$descent}->id($item);  
     #my $ident = $self->{meta}->{$descent}->{storage}->id($item);  
   
     my $ident;  
     my $provider_method = $self->{meta}->{$descent}->{IdentProvider}->{method};  
     my $provider_arg = $self->{meta}->{$descent}->{IdentProvider}->{arg};  
   
     # resolve to ident  
     if ($provider_method eq 'property') {  
       $ident = $payload->{$provider_arg};  
   
     } elsif ($provider_method eq 'storage_method') {  
       #$ident = $self->{meta}->{$descent}->{storage}->id($item);  
       $ident = $self->{meta}->{$descent}->{storage}->$provider_arg($payload);  
     }  
       
     $self->{node}->{$descent}->{ident} = $ident;  
       
   return 1 if $ident;  
   
 }  
   
   
 sub _modifyNode {  
   my $self = shift;  
   my $descent = shift;  
   my $action = shift;  
   my $map = shift;  
   my $crit = shift;  
   
   # map for new style callbacks  
   my $map_callbacks = {};  
   
   # checks go first!  
     
     # TODO: this should be reviewed first - before extending  ;-)  
     # TODO: this should be extended:  
     # count this cases inside the caller to this sub and provide a better overall message  
     # if this counts still zero in the end:  
     #     "No nodes have been touched for modify: Do you have column-headers in your csv file?"  
     if (not defined $self->{node}) {  
       #$logger->critical( __PACKAGE__ . "->_modifyNode failed: \"$descent\" node is empty." );  
       #return;  
     }  
   
   # transfer callback nodes from value map to callback map - handle them afterwards! - (new style callbacks)  
   if (my $callbacks = $self->{meta}->{$descent}->{Callback}) {  
     foreach my $callback (keys %{$callbacks->{write}}) {  
       $map_callbacks->{write}->{$callback} = $map->{$callback};  
       delete $map->{$callback};  
     }  
   }  
     
     
   #print Dumper($self->{meta});  
   
   # DBI speaks SQL  
   if ($self->{meta}->{$descent}->{storage}->{locator}->{type} eq 'DBI') {  
   
 #print Dumper($self->{node});  
     my $sql_main;  
     # translate map to sql  
     #print $action, "\n"; exit;  
     #print $self->{meta}->{$descent}->{node}, "\n"; exit;  
     #print "action:";  
     #print $action, "\n";  
 #$action = "anc";  
 #print "yai", "\n";  
   
 #print Dumper($map);  
 #delete $map->{cs};  
   
     if (lc($action) eq 'insert') {  
       $sql_main = hash2Sql($self->{meta}->{$descent}->{node}, $map, 'SQL_INSERT');  
     } elsif (lc $action eq 'update') {  
       $crit ||= "$self->{meta}->{$descent}->{IdentProvider}->{arg}='$self->{node}->{$descent}->{ident}'";  
       $sql_main = hash2Sql($self->{meta}->{$descent}->{node}, $map, 'SQL_UPDATE', $crit);  
     }  
   
 #$sql_main = "UPDATE currencies_csv SET oid='abcdef' WHERE text='Australian Dollar' AND key='AUD';";  
 #$sql_main = "UPDATE currencies_csv SET oid='huhu2' WHERE ekey='AUD'";  
   
 #print "sql: ", $sql_main, "\n";  
 #exit;  
   
     # transfer data  
     my $sqlHandle = $self->{meta}->{$descent}->{storage}->sendCommand($sql_main);  
   
 #exit;  
   
     # handle errors  
     if ($sqlHandle->err) {  
       #if ($self->{args}->{debug}) { print "sql-error with statement: $sql_main", "\n"; }  
       $self->{node}->{status}->{error} = {  
         statement => $sql_main,  
         state => $sqlHandle->state,  
         err => $sqlHandle->err,  
         errstr => $sqlHandle->errstr,  
       };  
     } else {  
       $self->{node}->{status}->{ok} = 1;  
     }  
   
   # Tangram does it the oo-way (naturally)  
   } elsif ($self->{meta}->{$descent}->{storage}->{locator}->{type} eq 'Tangram') {  
     my $sql_main;  
     my $object;  
   
     # determine classname  
     my $classname = $self->{meta}->{$descent}->{node};  
       
     # properties to exclude  
     my @exclude = @{$self->{meta}->{$descent}->{subnodes_exclude}};  
   
   
     if (my $identProvider = $self->{meta}->{$descent}->{IdentProvider}) {  
       push @exclude, $identProvider->{arg};  
     }  
   
     # new feature:  
     #     - check TypeProvider metadata property from other side  
     #     - use argument (arg) inside as a classname for object creation on this side  
     #my $otherSide = $self->_otherSide($descent);  
     if (my $typeProvider = $self->{meta}->{$descent}->{TypeProvider}) {  
       #print Dumper($map);  
       $classname = $map->{$typeProvider->{arg}};  
       # remove nodes from map also (push nodes to "subnodes_exclude" list)  
       push @exclude, $typeProvider->{arg};  
     }  
       
     # exclude banned properties (remove from map)  
     #map { delete $self->{node}->{map}->{$_} } @{$self->{args}->{exclude}};  
     map { delete $map->{$_} } @exclude;  
   
     # list of properties  
     my @props = keys %{$map};  
       
     # transfer data  
     if (lc $action eq 'insert') {  
   
       # build array to initialize object  
       #my @initarray = ();  
       #map { push @initarray, $_, undef; } @props;  
   
       # make the object persistent in four steps:  
       #   - raw create (perl / class tangram scope)  
       #   - engine insert (tangram scope)   ... this establishes inheritance - don't try to fill in inherited properties before!  
       #   - raw fill-in from hash (perl scope)  
       #   - engine update (tangram scope)  ... this updates all properties just filled in  
         
       # create new object ...  
       #my $object = $classname->new( @initarray );  
       $object = $classname->new();  
         
       # ... pass to orm ...  
       $self->{meta}->{$descent}->{storage}->insert($object);  
   
       # ... and initialize with empty (undef'd) properties.  
       #print Dumper(@props);  
       map { $object->{$_} = undef; } @props;  
   
       # mix in values ...  
       hash2object($object, $map);  
   
       # ... and re-update@orm.  
 #print Dumper($object);  
       $self->{meta}->{$descent}->{storage}->update($object);  
   
       # asymmetry: get ident after insert  
       # TODO:  
       #   - just do this if it is an IdentAuthority  
       #   - use IdentProvider metadata here  
 #print Dumper($self->{meta}->{$descent});  
       my $oid = $self->{meta}->{$descent}->{storage}->id($object);  
 #print "oid: $oid", "\n";  
       $self->{node}->{$descent}->{ident} = $oid;  
   
   
     } elsif (lc $action eq 'update') {  
         
       # get fresh object from orm first  
       $object = $self->{meta}->{$descent}->{storage}->load($self->{node}->{$descent}->{ident});  
   
 #print Dumper($self->{node});  
         
       # mix in values  
       #print Dumper($object);  
       hash2object($object, $map);  
       #print Dumper($object);  
       #exit;  
       $self->{meta}->{$descent}->{storage}->update($object);  
     }  
   
     my $error = 0;  
   
     # handle new style callbacks - this is a HACK - do this without an eval!  
     #print Dumper($map);  
     #print "cb: ", Dumper($self->{meta}->{$descent}->{Callback});  
     #print Dumper($map_callbacks);  
     foreach my $node (keys %{$map_callbacks->{write}}) {  
       #print Dumper($node);  
       my $perl_callback = $self->{meta}->{$descent}->{node} . '::' . $node . '_write';  
       my $evalstring = $perl_callback . '( { object => $object, value => $map_callbacks->{write}->{$node}, storage => $self->{meta}->{$descent}->{storage} } );';  
       #print $evalstring, "\n"; exit;  
       eval($evalstring);  
       if ($@) {  
         $error = 1;  
         print $@, "\n";  
       }  
         
       #print "after eval", "\n";  
         
       if (!$error) {  
         # re-update@orm  
         $self->{meta}->{$descent}->{storage}->update($object);  
       }  
     }  
     
     # handle errors  
     if ($error) {  
       #print "error", "\n";  
 =pod  
       my $sqlHandle;  
       #if ($self->{args}->{debug}) { print "sql-error with statement: $sql_main", "\n"; }  
       $self->{node}->{status}->{error} = {  
         statement => $sql_main,  
         state => $sqlHandle->state,  
         err => $sqlHandle->err,  
         errstr => $sqlHandle->errstr,  
       };  
 =cut  
       # rollback....  
       #print "rollback", "\n";  
       $self->{meta}->{$descent}->{storage}->erase($object);  
       #print "after rollback", "\n";  
     } else {  
       $self->{node}->{status}->{ok} = 1;  
     }  
   
   
   }  
   
 }  
   
 # TODO:  
 # this should be split up into...  
 #   - a "_statNode" (should just touch the node to check for existance)  
 #   - a "_loadNode" (should load node completely)  
 #   - maybe additionally a "loadNodeProperty" (may specify properties to load)  
 #   - introduce $self->{nodecache} for this purpose  
 # TODO:  
 #   should we:  
 #     - not pass ident in here but resolve it via "$descent"?  
 #     - refactor this and stuff it with additional debug/error message  
 #       - this = the way the implicit load mechanism works  
 sub _statloadNode {  
   
   my $self = shift;  
   my $descent = shift;  
   my $ident = shift;  
   my $force = shift;  
   
   # fetch entry to retrieve checksum from  
   # was:  
   if (!$self->{node}->{$descent} || $force) {  
   # is:  
   #if (!$self->{node}->{$descent}->{item} || $force) {  
       
     if (!$ident) {  
       #print "\n", "Attempt to fetch entry implicitely by ident failed: no ident given! This may result in an insert if no write-protection is in the way.", "\n";  
       return;  
     }  
       
     # patch for DBD::CSV  
     if ($ident && $ident eq 'Null') {  
       return;  
     }  
   
 #print "yai!", "\n";  
   
     my $query = {  
       node => $self->{meta}->{$descent}->{node},  
       subnodes => [qw( cs )],  
       criterias => [  
         { key => $self->{meta}->{$descent}->{IdentProvider}->{arg},  
            op => 'eq',  
            val => $ident },  
       ]  
     };  
   
 #print Dumper($query);  
   
     my $result = $self->{meta}->{$descent}->{storage}->sendQuery($query);  
   
     my $entry = $result->getNextEntry();  
   
 #print Dumper($entry);  
 #print "pers: " . $self->{meta}->{$descent}->{storage}->is_persistent($entry), "\n";  
 #my $state = $self->{meta}->{$descent}->{storage}->_fetch_object_state($entry, { name => 'TransactionHop' } );  
 #print Dumper($state);  
   
     my $status = $result->getStatus();  
   
 #print Dumper($status);  
       
     # TODO: enhance error handling (store inside tc)  
     #if (!$row) {  
     #  print "\n", "row error", "\n";  
     #  next;  
     #}  
   
     # these checks run before actually loading payload- and meta-data to node-container  
       
       # 1st level - hard error  
       if ($status && $status->{err}) {  
         $logger->debug( __PACKAGE__ . "->_statloadNode (ident=\"$ident\") failed - hard error (that's ok): $status->{err}" );  
         return;  
       }  
     
       # 2nd level - logical (empty/notfound) error  
       if (($status && $status->{empty}) || !$entry) {  
         $logger->debug( __PACKAGE__ . "->_statloadNode (ident=\"$ident\") failed - logical error (that's ok)" );  
         #print "no entry (logical)", "\n";  
         return;  
       }  
   
 #print Dumper($entry);  
   
     # was:  
     # $self->{node}->{$descent}->{ident} = $ident;    
     # is:  
     # TODO: re-resolve ident from entry via metadata "IdentProvider" here - like elsewhere  
     $self->{node}->{$descent}->{ident} = $ident;  
     $self->{node}->{$descent}->{payload} = $entry;  
   
   }  
     
   return 1;  
     
 }  
448    
449  sub _doTransferToTarget {  sub _doTransferToTarget {
450    my $self = shift;    my $self = shift;
451    my $action = shift;    my $action = shift;
452      #print Dumper($self->{node});
453      #exit;
454    $self->_modifyNode('target', $action, $self->{node}->{map});    $self->_modifyNode('target', $action, $self->{node}->{map});
455  }  }
456    
457    
458  sub _doModifySource_IdentChecksum {  sub _doModifySource_IdentChecksum {
459    my $self = shift;    my $self = shift;
460    my $ident_new = shift;    my $ident_new = shift;
# Line 1094  sub _doModifySource_IdentChecksum { Line 475  sub _doModifySource_IdentChecksum {
475  }  }
476    
477    
 # this is a shortcut method  
 # ... let's try to avoid _any_ redundant code in here (ok... - at the cost of method lookups...)  
 sub _getNodeList {  
   my $self = shift;  
   my $descent = shift;  
   my $filter = shift;  
   return $self->{meta}->{$descent}->{storage}->getListFiltered($self->{meta}->{$descent}->{node}, $filter);  
 }  
   
478    
479  sub _prepareNode_MetaProperties {  sub _prepareNode_MetaProperties {
480    my $self = shift;    my $self = shift;
# Line 1204  sub _erase_all { Line 576  sub _erase_all {
576    $self->{meta}->{$descent}->{storage}->eraseAll($node);    $self->{meta}->{$descent}->{storage}->eraseAll($node);
577  }  }
578    
   
 =pod  
   
   
 =head1 DESCRIPTION  
   
 Data::Transfer::Sync is a module providing a generic synchronization process  
 across arbitrary/multiple storages based on a ident/checksum mechanism.  
 It sits on top of Data::Storage.  
   
   
 =head1 REQUIREMENTS  
   
   For full functionality:  
     Data::Storage  
     Data::Transform  
     Data::Compare  
     ... and all their dependencies  
   
   
 =head1 AUTHORS / COPYRIGHT  
   
 The Data::Storage module is Copyright (c) 2002 Andreas Motl.  
 All rights reserved.  
   
 You may distribute it under the terms of either the GNU General Public  
 License or the Artistic License, as specified in the Perl README file.  
   
   
 =head1 SUPPORT / WARRANTY  
   
 Data::Storage is free software. IT COMES WITHOUT WARRANTY OF ANY KIND.  
   
   
   
 =head1 BUGS  
   
   When in "import" mode for windows file - DBD::AutoCSV may hang.  
   Hint: Maybe the source node contains an ident-, but no checksum-column?  
   
   
 =head1 USER LEVEL ERRORS  
   
 =head4 Mapping  
   
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   info: BizWorks::Process::Setup->syncResource( source_node Currency mode PULL erase 0 import 0 )critical: BizWorks::Process::Setup->startSync: Can't access mapping for node "Currency" - please check BizWorks::ResourceMapping.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   You have to create a sub for each node used in synchronization inside named Perl module. The name of this sub _must_ match  
   the name of the node you want to sync. This sub holds mapping metadata to give the engine hints about how  
   to access the otherwise generic nodes.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   
   
 =head4 DBD::AutoCSV's rulebase  
     
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   info: BizWorks::Process::Setup->syncResource( source_node Currency mode PULL erase 0 import 0 )  
   info: Data::Transfer::Sync->syncNodes: source=L/Currency <- target=R/currencies.csv  
     
   Execution ERROR: Error while scanning: Missing first row or scanrule not applied. at C:/home/amo/develop/netfrag.org/nfo/perl/libs/DBD/CSV.p  
   m line 165, <GEN9> line 1.  
    called from C:/home/amo/develop/netfrag.org/nfo/perl/libs/Data/Storage/Handler/DBI.pm at 123.  
     
   DBI-Error: DBD::AutoCSV::st fetchrow_hashref failed: Attempt to fetch row from a Non-SELECT statement  
   notice: Data::Transfer::Sync->syncNodes: No nodes to synchronize.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   DBD::AutoCSV contains a rulebase which is spooled down while attempting to guess the style of the csv file regarding  
   parameters like newline (eol), column-seperation-character (sep_char), quoting character (quote_char).  
   If this spool runs out of entries and no style could be resolved, DBD::CSV dies causing this "Execution ERROR" which  
   results in a "DBI-Error" afterwards.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   
   
 =head4 Check structure of source node  
     
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   info: Data::Transfer::Sync->syncNodes: source=L/Currency <- target=R/currencies.csv  
   critical: Data::Transfer::Sync->syncNodes: Can not synchronize: No ident found in source node, maybe try to "import" this node first.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   If lowlevel detection succeeds, but no other required informations are found, this message is issued.  
   "Other informations" might be:  
     - column-header-row completely missing  
     - ident column is empty  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   
   
 =head4 Modify structure of source node  
     
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   info: Data::Transfer::Sync->syncNodes: source=L/Currency <- target=R/currencies.csv  
   info: Data::Transfer::Sync->_prepareNode_MetaProperties( descent source )  
   warning: Data::Transfer::Sync->_prepareNode_MetaProperties: node is lacking meta properties - will try to alter...  
   SQL ERROR: Command 'ALTER' not recognized or not supported!  
     
   SQL ERROR: Command 'ALTER' not recognized or not supported!  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   The Engine found a node which structure does not match the required. It tries to alter this automatically - only when doing "import" -  
   but the DBD driver (in this case DBD::CSV) gets in the way croaking not to be able to do this.  
   This could also appear if your database connection has insufficient rights to modify the database structure.  
   DBD::CSV croaks because it doesn't implement the ALTER command, so please edit your columns manually.  
   Hint: Add columns with the names of your "ident" and "checksum" property specifications.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   
   
 =head4 Load source node by ident  
   
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   info: Data::Transfer::Sync->_prepareNode_DummyIdent( descent source )  
   pcritical: Data::Transfer::Sync->_modifyNode failed: "source" node is empty.  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   The source node could not be loaded. Maybe the ident is missing. Please check manually.  
   Hint: Like above, the ident and/or checksum columns may be missing....  
   - - - - - - - - - - - - - - - - - - - - - - - - - -  
   
   
 =head1 TODO  
   
   - sub _resolveIdentProvider  
   - wrap _doModifySource and _doTransferTarget around a core function which can change virtually any type of node  
   - split this module up into Sync.pm, Sync/Core.pm, Sync/Compare.pm and Sync/Compare/Checksum.pm  
      - introduce _compareNodes as a core method and wrap it around methods in Sync/Compare/Checksum.pm  
   - introduce Sync/Compare/MyComparisonImplementation.pm  
   - some generic deferring method - e.g. "$self->defer(action)" - to be able to accumulate a bunch of actions for later processing  
      - this implies everything done is _really_ split up into generic actions - how else would we defer them???  
      - example uses:  
         - fetch whole checksum list from node  
         - remember source ident retransmits  
      - remember: this is convenient - and maybe / of course faster - but we'll loose "per-node-atomic" operations  
   - feature: mechanism to implicit inject checksum property to nodes (alter table / modify schema)  
   - expand statistics / keep track of:  
     - touched/untouched nodes  
   - full sync  
     - just do a push and a pull for now but use stats for touched nodes in between to speed up things  
   - introduce some new metadata flags for a synchronization partner which is (e.g.) of "source" or "target":  
     - isNewNodePropagator  
     - isWriteProtected  
   
   
 =cut  
   
579  1;  1;

Legend:
Removed from v.1.1  
changed lines
  Added in v.1.4

MailToCvsAdmin">MailToCvsAdmin
ViewVC Help
Powered by ViewVC 1.1.26 RSS 2.0 feed