[BACK]Return to searchcat.pl CVS log [TXT] [DIR] Up to [LON-CAPA] / loncom / metadata_database

Diff for /loncom/metadata_database/searchcat.pl between versions 1.29 and 1.30

version 1.29, 2003/02/03 13:42:16 version 1.30, 2003/02/03 17:01:55
Line 374  sub escape ($) Line 374  sub escape ($)
   
 B<build_on_the_fly_dynamic_metadata> - evaluate and store dynamic metadata.  B<build_on_the_fly_dynamic_metadata> - evaluate and store dynamic metadata.
   
 Dynamic metadata is stored in a nohist_resevaldata GDBM database.  Returns the dynamic metadata for an author, which will later be added to the
 The only thing that this subroutine really makes happen is adjusting  MySQL database (not yet implemented).
 a 'count' value inside the F<nohist_new_resevaldata.db> as well  
 as updating F<nohist_new_resevaldata.db> with information from  The vast majority of entries in F<nohist_resevaldata.db>, which contains
 F<nohist_resevaldata.db>.  the dynamic metadata for an author's resources, are "count", which make
   the file really large and evaluation really slow.
   
   While computing the current value of all dynamic metadata
   for later insertion into the MySQL metadata cache (not yet implemented),
   this routine also simply adds up all "count" type fields and replaces them by
   one new field with the to-date count.
   
   Only after successful completion of working with one author, copy new file to
   original file. Copy to tmp-"new"-db-file was necessary since db-file size 
   would not shrink after "delete" of key.
   
 It may need optmization, but since it gets called once a week. . .  
 =over 4  =over 4
   
 Parameters:  Parameters:
Line 399  Returns: Line 408  Returns:
   
 =cut  =cut
   
 sub build_on_the_fly_dynamic_metadata ($)  sub build_on_the_fly_dynamic_metadata {
   {  
     # some elements in here maybe non-obvious  
   
     # Need to compute the user's directory.      # Need to compute the user's directory.
     my $url = &declutter(shift(@_));      my $url=&declutter(shift);
     $url =~ s/\.meta$//;      $url=~s/\.meta$//;
     my %returnhash = ();      my %returnhash=();
     my ($adomain,$aauthor) = ($url =~ m!^(\w+)/(\w+)/!);      my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//);
     my $user_directory = &construct_path_to_user_directory($adomain,$aauthor);      my $user_directory=&construct_path_to_user_directory($adomain,$aauthor);
   
     # Attempt a GDBM database instantiation inside users directory and proceed.      # Attempt a GDBM database instantiation inside users directory and proceed.
     if ((tie(%evaldata,'GDBM_File',      if ((tie(%evaldata,'GDBM_File',
Line 416  sub build_on_the_fly_dynamic_metadata ($ Line 423  sub build_on_the_fly_dynamic_metadata ($
      '/nohist_resevaldata.db',&GDBM_READER(),0640)) &&       '/nohist_resevaldata.db',&GDBM_READER(),0640)) &&
         (tie(%newevaldata,'GDBM_File',          (tie(%newevaldata,'GDBM_File',
      $user_directory.       $user_directory.
      '/nohist_new_resevaldata.db',&GDBM_WRCREAT(),0640)))       '/nohist_new_resevaldata.db',&GDBM_WRCREAT(),0640))) {
       {  
  # For different variables, track the running sum and counts.   # For different variables, track the running sum and counts.
  my %sum = ();   my %sum=();
  my %cnt = ();   my %cnt=();
   
  # Define computed items as a sum (add) or an average (avg) or a raw   # Define computed items as a sum (add) or an average (avg) or a raw
  # count (cnt) or 'app'?   # count (cnt) or append (app)?
  my %listitems=('count'        => 'add',   my %listitems=('count'        => 'add',
        'course'       => 'add',         'course'       => 'add',
        'avetries'     => 'avg',         'avetries'     => 'avg',
Line 439  sub build_on_the_fly_dynamic_metadata ($ Line 445  sub build_on_the_fly_dynamic_metadata ($
        );         );
   
  # Untaint the url and use as part of a regular expression.   # Untaint the url and use as part of a regular expression.
  my $regexp = $url;   my $regexp=$url;
  $regexp =~ s/(\W)/\\$1/g;   $regexp=~s/(\W)/\\$1/g;
  $regexp = '___'.$regexp.'___([a-z]+)$'; #' emacs   $regexp='___'.$regexp.'___([a-z]+)$'; #' emacs
   
  # Check existing nohist database for this url.   # Check existing database for this author.
         # this is modfying the 'count' entries          # this is modifying the 'count' entries
         # and copying all othe entries over          # and copying all other entries over
  foreach (keys %evaldata)  
   {   foreach (keys %evaldata) {
     my $key = &unescape($_);      my $key=&unescape($_);
     if ($key =~ /$regexp/) # If url-based entry exists.      if ($key=~/$regexp/) { # If url-based entry exists.
       {   my $ctype=$1; # Set to specific category type.
  my $ctype = $1; # Set to specific category type.  
   
  # Do an increment for this category type.   # Do an increment for this category type.
  if (defined($cnt{$ctype}))   if (defined($cnt{$ctype})) {
   {  
     $cnt{$ctype}++;       $cnt{$ctype}++; 
   }   } else {
  else      $cnt{$ctype}=1; 
   {   }
     $cnt{$ctype} = 1;                   unless ($listitems{$ctype} eq 'app') { # append comments
   }  
                 unless ($listitems{$ctype} eq 'app') # WHAT DOES 'app' MEAN?  
   {  
     # Increment the sum based on the evaluated data in the db.      # Increment the sum based on the evaluated data in the db.
     if (defined($sum{$ctype}))      if (defined($sum{$ctype})) {
       {   $sum{$ctype}+=$evaldata{$_};
  $sum{$ctype} += $evaldata{$_};      } else {
       }   $sum{$ctype}=$evaldata{$_};
     else      }
       {    } else { # 'app' mode, means to use '<hr />' as a separator
  $sum{$ctype} = $evaldata{$_};      if (defined($sum{$ctype})) {
       }   if ($evaldata{$_}) {
    }      $sum{$ctype}.='<hr />'.$evaldata{$_};
  else # 'app' mode, means to use '<hr />' as a separator   }
   {      } else {
     if (defined($sum{$ctype}))   $sum{$ctype}=''.$evaldata{$_};
       {      }
  if ($evaldata{$_})          }
   {   if ($ctype ne 'count') {
     $sum{$ctype} .= '<hr />'.$evaldata{$_};  
   }  
       }  
     else  
       {  
  $sum{$ctype} = ''.$evaldata{$_};  
       }  
   }  
  if ($ctype ne 'count')  
   {  
                     # this is copying all data except 'count' attributes                      # this is copying all data except 'count' attributes
     $newevaldata{$_} = $evaldata{$_};      $newevaldata{$_}=$evaldata{$_};
   }          }
       }      }
   }   }
   
         # the only other time this loop is useful is for the 'count' hash          # these values will be returned (currently still unused)
         # element   foreach (keys %cnt) {
  foreach (keys %cnt)      if ($listitems{$_} eq 'avg') {
   {   $returnhash{$_}=int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0;
     if ($listitems{$_} eq 'avg')      } elsif ($listitems{$_} eq 'cnt') {
       {   $returnhash{$_}=$cnt{$_};
  $returnhash{$_} = int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0;      } else {
       }   $returnhash{$_}=$sum{$_};
     elsif ($listitems{$_} eq 'cnt')      }
       {   }
  $returnhash{$_} = $cnt{$_};  
       }          # generate new count key in resevaldata, insert sum
     else   if ($returnhash{'count'}) {
       {      my $newkey=$$.'_'.time.'_searchcat___'.&escape($url).'___count';
  $returnhash{$_} = $sum{$_};      $newevaldata{$newkey}=$returnhash{'count'};
       }   }
   }  
   
         # seems to be doing something useful  
  if ($returnhash{'count'})  
   {  
     my $newkey = $$.'_'.time.'_searchcat___'.&escape($url).'___count';  
     $newevaldata{$newkey} = $returnhash{'count'};  
   }  
   
  untie(%evaldata); # Close/release the original nohist database.   untie(%evaldata); # Close/release the original nohist database.
  untie(%newevaldata); # Close/release the new nohist database.   untie(%newevaldata); # Close/release the new nohist database.
       }      }
     return(%returnhash);      return %returnhash;
   }  }
   
 =pod  =pod
   
Removed from v.1.29  
changed lines
  Added in v.1.30

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>