Diff for /loncom/metadata_database/searchcat.pl between versions 1.29 and 1.54

version 1.29, 2003/02/03 13:42:16 version 1.54, 2004/01/05 15:54:22
Line 26 Line 26
 #  #
 # http://www.lon-capa.org/  # http://www.lon-capa.org/
 #  #
 # YEAR=2001  
 # 04/14/2001, 04/16/2001 Scott Harrison  
 #  
 # YEAR=2002  
 # 05/11/2002 Scott Harrison  
 #  
 # YEAR=2003  
 # Scott Harrison  
 #  
 ###  ###
   
 =pod  =pod
Line 52  filesystem installation location: F</etc Line 43  filesystem installation location: F</etc
 Here is the cron job entry.  Here is the cron job entry.
   
 C<# Repopulate and refresh the metadata database used for the search catalog.>  C<# Repopulate and refresh the metadata database used for the search catalog.>
   
 C<10 1 * * 7    www    /home/httpd/perl/searchcat.pl>  C<10 1 * * 7    www    /home/httpd/perl/searchcat.pl>
   
 This script only allows itself to be run as the user C<www>.  This script only allows itself to be run as the user C<www>.
Line 65  The metadata is entered into a SQL datab Line 55  The metadata is entered into a SQL datab
 This script also does general database maintenance such as reformatting  This script also does general database maintenance such as reformatting
 the C<loncapa:metadata> table if it is deprecated.  the C<loncapa:metadata> table if it is deprecated.
   
 This script also builds dynamic temporal metadata and stores this inside  This script evaluates dynamic metadata from the authors'
 a F<nohist_resevaldata.db> database file.  F<nohist_resevaldata.db> database file in order to store it in MySQL.
   
 This script is playing an increasingly important role for a loncapa  This script is playing an increasingly important role for a loncapa
 library server.  The proper operation of this script is critical for a smooth  library server.  The proper operation of this script is critical for a smooth
Line 74  and correct user experience. Line 64  and correct user experience.
   
 =cut  =cut
   
 # ========================================================== Setting things up.  use strict;
   
 # ------------------------------------------------------  Use external modules.  
   
 use lib '/home/httpd/lib/perl/';  use lib '/home/httpd/lib/perl/';
 use LONCAPA::Configuration;  use LONCAPA::Configuration;
Line 87  use DBI; Line 75  use DBI;
 use GDBM_File;  use GDBM_File;
 use POSIX qw(strftime mktime);  use POSIX qw(strftime mktime);
   
 # ----------------- Code to enable 'find' subroutine listing of the .meta files  require "find.pl";
 use File::Find;  
   
 # List of .meta files (used on a per-user basis).  
 my @metalist;  my @metalist;
   
 # ---------------  Read loncapa_apache.conf and loncapa.conf and get variables.  my $simplestatus='';
 my $perlvarref = LONCAPA::Configuration::read_conf('loncapa.conf');  my %countext=();
 my %perlvar = %{$perlvarref};  
 undef($perlvarref); # Remove since sensitive and not needed.  
 delete($perlvar{'lonReceipt'}); # Remove since sensitive and not needed.  
   
 # ------------------------------------- Only run if machine is a library server  
 if ($perlvar{'lonRole'} ne 'library')  
   {  
     exit(0);  
   }  
   
 # ------------------------------ Make sure this process is running as user=www.  
 my $wwwid = getpwnam('www');  
 if ($wwwid != $<)  
   {  
     $emailto = "$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";  
     $subj = "LON: $perlvar{'lonHostID'} User ID mismatch";  
     system("echo 'User ID mismatch. searchcat.pl must be run as user www.' | ".  
    "mailto $emailto -s '$subj' > /dev/null");  
     exit(1);  
   }  
   
 # ------------------------------------------------------ Initialize log output.  
 open(LOG,'>'.$perlvar{'lonDaemons'}.'/logs/searchcat.log');  
 print(LOG '==== Searchcat Run '.localtime().' ===='."\n\n");  
   
 my $dbh; # Database object reference handle.  
   
 # ----------------------------- Verify connection to loncapa:metadata database.  
 unless (  
  $dbh = DBI->connect('DBI:mysql:loncapa','www',  
     $perlvar{'lonSqlAccess'},  
     { RaiseError => 0,PrintError => 0})  
  )  
   {   
     print(LOG '**** ERROR **** Cannot connect to database!'."\n");  
     exit(0);  
   }  
   
 # ------------------------------ Create loncapa:metadata table if non-existent.  
 my $make_metadata_table = 'CREATE TABLE IF NOT EXISTS metadata ('.  
     'title TEXT, author TEXT, subject TEXT, url TEXT, keywords TEXT, '.  
     'version TEXT, notes TEXT, abstract TEXT, mime TEXT, language TEXT, '.  
     'creationdate DATETIME, lastrevisiondate DATETIME, owner TEXT, '.  
     'copyright TEXT, utilitysemaphore BOOL, FULLTEXT idx_title (title), '.  
     'FULLTEXT idx_author (author), FULLTEXT idx_subject (subject), '.  
     'FULLTEXT idx_url (url), FULLTEXT idx_keywords (keywords), '.  
     'FULLTEXT idx_version (version), FULLTEXT idx_notes (notes), '.  
     'FULLTEXT idx_abstract (abstract), FULLTEXT idx_mime (mime), '.  
     'FULLTEXT idx_language (language), FULLTEXT idx_owner (owner), '.  
     'FULLTEXT idx_copyright (copyright)) TYPE=MYISAM';  
   
 $dbh->do($make_metadata_table); # Generate the table.  
   
 # ----------------------------- Verify format of the loncapa:metadata database.  
 #                               (delete and recreate database if necessary).  
   
 # Make a positive control for verifying table structure.  
 my $make_metadata_table_CONTROL = $make_metadata_table;  
 $make_metadata_table_CONTROL =~  
     s/^(CREATE TABLE IF NOT EXISTS) metadata/$1 CONTROL_metadata/;  
   
 $dbh->do('DROP TABLE IF EXISTS CONTROL_metadata');  
 $dbh->do($make_metadata_table_CONTROL);  
   
 my $table_description; # selectall reference to the table description.  
   
 my $CONTROL_table_string; # What the table description should look like.  
 my $table_string; # What the table description does look like.  
   
 # Calculate the CONTROL table description (what it should be).  
 $table_description = $dbh->selectall_arrayref('describe CONTROL_metadata');  
 foreach my $table_row (@{$table_description})  
   {  
     $CONTROL_table_string .= join(',',@{$table_row})."\n";  
   }  
   
 # Calculate the current table description (what it currently looks like).  
 $table_description = $dbh->selectall_arrayref('describe metadata');  
 foreach my $table_row (@{$table_description})  
   {  
     $table_string .= join(',',@{$table_row})."\n";  
   }  
   
 if ($table_string ne $CONTROL_table_string)  
   {  
     # Log this incident.  
     print(LOG '**** WARNING **** Table structure mismatch, need to regenerate'.  
   '.'."\n");  
     # Delete the table.  
     $dbh->do('DROP TABLE IF EXISTS metadata');  
     # Generate the table.  
     $dbh->do($make_metadata_table);  
   }  
   
 $dbh->do('DROP TABLE IF EXISTS CONTROL_metadata'); # Okay.  Done with control.  
   
 # ----------------------------------------------- Set utilitysemaphore to zero.  
 $dbh->do('UPDATE metadata SET utilitysemaphore = 0');  
   
 # ========================================================= Main functionality.  
   
 # - Determine home authors on this server based on resources dir and user tree.  
   
 # RESOURCES: the resources directory (subdirs correspond to author usernames).  
 opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}") or  
     (print(LOG '=== /res/--lonDefDomain-- directory is not accessible'."\n")  
      and exit(0));  
   
 # query_home_server_status will look for user home directories on this machine.  
 my @homeusers =  
     grep {&query_home_server_status($perlvar{'lonDocRoot'}.'/res/'.  
     $perlvar{'lonDefDomain'}.'/'.$_)  
   } grep {!/^\.\.?$/} readdir(RESOURCES);  
 closedir(RESOURCES);  
   
 unless (@homeusers)  
   {  
     print(LOG '=== No home users found on this server.'."\n");  
   }  
   
 # Consider each author individually.  
 foreach my $user (@homeusers)  
   {  
     # Make a log entry.  
     print(LOG "\n".'=== User: '.$user."\n\n");  
   
     # Get filesystem path to this user's directory.  
     my $user_directory =  
  &construct_path_to_user_directory($perlvar{'lonDefDomain'},$user);  
   
     # Remove left-over db-files from a potentially crashed searchcat run.  
     unlink($user_directory.'/nohist_new_resevaldata.db');  
   
     # Cleanup the metalist array.  
     undef(@metalist);  
     @metalist = ();  
   
     # This will add entries to the @metalist array.  
     &File::Find::find(\&wanted,  
       $perlvar{'lonDocRoot'}.'/res/'.  
       $perlvar{'lonDefDomain'}.'/'.$user);  
   
     # -- process file to get metadata and put into search catalog SQL database  
     # Also, build and store dynamic metadata.  
     # Also, delete record entries before refreshing.  
     foreach my $m (@metalist)  
       {  
  # Log this action.  
  print(LOG "- ".$m."\n");  
   
  # Get metadata from the file.  
  my $ref = get_metadata_from_file($m);  
   
  # Make a datarecord identifier for this resource.  
  my $m2 = '/res/'.declutter($m);  
  $m2 =~ s/\.meta$//;  
   
  # Build and store dynamic metadata inside nohist_resevaldata.db.  
  build_on_the_fly_dynamic_metadata($m2);  
   
  # Delete record if it already exists.  
  my $q2 = 'select * from metadata where url like binary '."'".$m2."'";  
  my $sth = $dbh->prepare($q2);  
  $sth->execute();  
  my $r1 = $sth->fetchall_arrayref;  
  if (@$r1)  
   {  
     $sth =   
  $dbh->prepare('delete from metadata where url like binary '.  
       "'".$m2."'");  
     $sth->execute();  
   }  
   
  # Add new/replacement record into the loncapa:metadata table.  
  $sth = $dbh->prepare('insert into metadata values ('.  
      '"'.delete($ref->{'title'}).'"'.','.  
      '"'.delete($ref->{'author'}).'"'.','.  
      '"'.delete($ref->{'subject'}).'"'.','.  
      '"'.$m2.'"'.','.  
      '"'.delete($ref->{'keywords'}).'"'.','.  
      '"'.'current'.'"'.','.  
      '"'.delete($ref->{'notes'}).'"'.','.  
      '"'.delete($ref->{'abstract'}).'"'.','.  
      '"'.delete($ref->{'mime'}).'"'.','.  
      '"'.delete($ref->{'language'}).'"'.','.  
      '"'.sql_formatted_time(  
        delete($ref->{'creationdate'})).'"'.','.  
      '"'.sql_formatted_time(  
    delete($ref->{'lastrevisiondate'})).'"'.','.  
      '"'.delete($ref->{'owner'}).'"'.','.  
      '"'.delete($ref->{'copyright'}).'"'.','.  
      '1'.')');  
  $sth->execute();  
       }  
   
 # ----------------------- Clean up database, remove stale SQL database records.  
     $dbh->do('DELETE FROM metadata WHERE utilitysemaphore = 0');  
   
 # -------------------------------------------------- Copy over the new db-files  
     system('mv '.$user_directory.'/nohist_new_resevaldata.db '.  
          $user_directory.'/nohist_resevaldata.db');  
   }  
   
 # --------------------------------------------------- Close database connection  
 $dbh->disconnect;  
 print LOG "\n==== Searchcat completed ".localtime()." ====\n";  
 close(LOG);  
 exit(0);  
   
 # ================================================================ Subroutines.  
   
 =pod  
   
 =head1 SUBROUTINES  
   
 =cut  
   
 =pod  
   
 B<unescape> - translate to unstrange escaped syntax to strange characters.  
   
 =over 4  # ----------------------------------------------------- write out simple status
   sub writesimple {
       open(SMP,'>/home/httpd/html/lon-status/mysql.txt');
       print SMP $simplestatus."\n";
       close(SMP);
   }
   
   sub writecount {
       open(RSMP,'>/home/httpd/html/lon-status/rescount.txt');
       foreach (keys %countext) {
    print RSMP $_.'='.$countext{$_}.'&';
       }
       print RSMP 'time='.time."\n";
       close(RSMP);
   }
   
   # -------------------------------------- counts files with different extensions
   sub count {
       my $file=shift;
       $file=~/\.(\w+)$/;
       my $ext=lc($1);
       if (defined($countext{$ext})) {
    $countext{$ext}++;
       } else {
    $countext{$ext}=1;
       }
   }
   # ----------------------------------------------------- Un-Escape Special Chars
   
 Parameters:  sub unescape {
       my $str=shift;
 =item I<$str> - string with unweird characters.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<string> - string with potentially weird characters.  
   
 =back  
   
 =cut  
   
 sub unescape ($)  
   {  
     my $str = shift(@_);  
     $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;      $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;
     return($str);      return $str;
   }  }
   
 =pod  
   
 B<escape> - translate strange characters to unstrange escaped syntax.  
   
 =over 4  
   
 Parameters:  # -------------------------------------------------------- Escape Special Chars
   
 =item I<$str> - string with potentially weird characters to unweird-ify.  sub escape {
       my $str=shift;
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<string> - unweird-ified string.  
   
 =back  
   
 =cut  
   
 sub escape ($)  
   {  
     my $str = shift(@_);  
     $str =~ s/(\W)/"%".unpack('H2',$1)/eg;      $str =~ s/(\W)/"%".unpack('H2',$1)/eg;
     return($str);      return $str;
   }  }
   
 =pod  
   
 B<build_on_the_fly_dynamic_metadata> - evaluate and store dynamic metadata.  
   
 Dynamic metadata is stored in a nohist_resevaldata GDBM database.  
 The only thing that this subroutine really makes happen is adjusting  
 a 'count' value inside the F<nohist_new_resevaldata.db> as well  
 as updating F<nohist_new_resevaldata.db> with information from  
 F<nohist_resevaldata.db>.  
   
 It may need optmization, but since it gets called once a week. . .  # ------------------------------------------- Code to evaluate dynamic metadata
 =over 4  
   
 Parameters:  sub dynamicmeta {
       my $url=&declutter(shift);
 =item I<$url> - the filesystem path (url may be a misnomer...)      $url=~s/\.meta$//;
       my %returnhash=(
 =back      'count' => 0,
       'course' => 0,
 =over 4      'course_list' => '',
       'avetries' => 'NULL',
 Returns:      'avetries_list' => '',
       'stdno' => 0,
 =item C<hash> - key-value table of dynamically evaluated metadata.      'stdno_list' => '',
       'usage' => 0,
 =back      'usage_list' => '',
       'goto' => 0,
 =cut      'goto_list' => '',
       'comefrom' => 0,
 sub build_on_the_fly_dynamic_metadata ($)      'comefrom_list' => '',
   {      'difficulty' => 'NULL',
     # some elements in here maybe non-obvious      'difficulty_list' => '',
                       'clear' => 'NULL',
     # Need to compute the user's directory.                      'technical' => 'NULL',
     my $url = &declutter(shift(@_));      'correct' => 'NULL',
     $url =~ s/\.meta$//;      'helpful' => 'NULL',
     my %returnhash = ();      'depth' => 'NULL',
     my ($adomain,$aauthor) = ($url =~ m!^(\w+)/(\w+)/!);      'comments' => ''
     my $user_directory = &construct_path_to_user_directory($adomain,$aauthor);      );
       my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//);
     # Attempt a GDBM database instantiation inside users directory and proceed.      my $prodir=&propath($adomain,$aauthor);
     if ((tie(%evaldata,'GDBM_File',  
             $user_directory.  # Get metadata except counts
      '/nohist_resevaldata.db',&GDBM_READER(),0640)) &&      if (tie(my %evaldata,'GDBM_File',
         (tie(%newevaldata,'GDBM_File',              $prodir.'/nohist_resevaldata.db',&GDBM_READER(),0640)) {
      $user_directory.   my %sum=();
      '/nohist_new_resevaldata.db',&GDBM_WRCREAT(),0640)))   my %cnt=();
       {   my %concat=();
  # For different variables, track the running sum and counts.   my %listitems=(
  my %sum = ();  
  my %cnt = ();  
   
  # Define computed items as a sum (add) or an average (avg) or a raw  
  # count (cnt) or 'app'?  
  my %listitems=('count'        => 'add',  
        'course'       => 'add',         'course'       => 'add',
          'goto'         => 'add',
          'comefrom'     => 'add',
        'avetries'     => 'avg',         'avetries'     => 'avg',
        'stdno'        => 'add',         'stdno'        => 'add',
        'difficulty'   => 'avg',         'difficulty'   => 'avg',
Line 438  sub build_on_the_fly_dynamic_metadata ($ Line 178  sub build_on_the_fly_dynamic_metadata ($
        'usage'        => 'cnt'         'usage'        => 'cnt'
        );         );
   
  # 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]+)$';
  $regexp = '___'.$regexp.'___([a-z]+)$'; #' emacs   while (my ($esckey,$value)=each %evaldata) {
       my $key=&unescape($esckey);
  # Check existing nohist database for this url.      if ($key=~/$regexp/) {
         # this is modfying the 'count' entries   my ($item,$purl,$cat)=split(/___/,$key);
         # and copying all othe entries over   if (defined($cnt{$cat})) { $cnt{$cat}++; } else { $cnt{$cat}=1; }
  foreach (keys %evaldata)   unless ($listitems{$cat} eq 'app') {
   {      if (defined($sum{$cat})) {
     my $key = &unescape($_);   $sum{$cat}+=&unescape($evaldata{$esckey});
     if ($key =~ /$regexp/) # If url-based entry exists.   $concat{$cat}.=','.$item;
       {      } else {
  my $ctype = $1; # Set to specific category type.   $sum{$cat}=&unescape($evaldata{$esckey});
    $concat{$cat}=$item;
  # Do an increment for this category type.      }
  if (defined($cnt{$ctype}))   } else {
   {      if (defined($sum{$cat})) {
     $cnt{$ctype}++;    if ($evaldata{$esckey}=~/\w/) {
   }      $sum{$cat}.='<hr />'.&unescape($evaldata{$esckey});
  else   }
   {      } else {
     $cnt{$ctype} = 1;    $sum{$cat}=''.&unescape($evaldata{$esckey});
   }      }
                 unless ($listitems{$ctype} eq 'app') # WHAT DOES 'app' MEAN?   }
   {      }
     # Increment the sum based on the evaluated data in the db.   }
     if (defined($sum{$ctype}))   untie(%evaldata);
       {  # transfer gathered data to returnhash, calculate averages where applicable
  $sum{$ctype} += $evaldata{$_};   while (my $cat=each(%cnt)) {
       }      if ($cnt{$cat} eq 'nan') { next; }
     else      if ($sum{$cat} eq 'nan') { next; }
       {      if ($listitems{$cat} eq 'avg') {
  $sum{$ctype} = $evaldata{$_};   if ($cnt{$cat}) {
       }      $returnhash{$cat}=int(($sum{$cat}/$cnt{$cat})*100.0+0.5)/100.0;
    }   } else {
  else # 'app' mode, means to use '<hr />' as a separator      $returnhash{$cat}='NULL';
   {   }
     if (defined($sum{$ctype}))      } elsif ($listitems{$cat} eq 'cnt') {
       {   $returnhash{$cat}=$cnt{$cat};
  if ($evaldata{$_})      } else {
   {   $returnhash{$cat}=$sum{$cat};
     $sum{$ctype} .= '<hr />'.$evaldata{$_};      }
   }      $returnhash{$cat.'_list'}=$concat{$cat};
       }   }
     else      }
       {  # get count
  $sum{$ctype} = ''.$evaldata{$_};      if (tie(my %evaldata,'GDBM_File',
       }              $prodir.'/nohist_accesscount.db',&GDBM_READER(),0640)) {
   }   my $escurl=&escape($url);
  if ($ctype ne 'count')   if (! exists($evaldata{$escurl})) {
   {      $returnhash{'count'}=0;
                     # this is copying all data except 'count' attributes   } else {
     $newevaldata{$_} = $evaldata{$_};      $returnhash{'count'}=$evaldata{$escurl};
   }   }
       }   untie %evaldata;
   }      }
       return %returnhash;
         # the only other time this loop is useful is for the 'count' hash  }
         # element    
  foreach (keys %cnt)  # ---------------  Read loncapa_apache.conf and loncapa.conf and get variables
   {  my $perlvarref=LONCAPA::Configuration::read_conf('loncapa.conf');
     if ($listitems{$_} eq 'avg')  my %perlvar=%{$perlvarref};
       {  undef $perlvarref;
  $returnhash{$_} = int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0;  delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed
       }  
     elsif ($listitems{$_} eq 'cnt')  
       {  
  $returnhash{$_} = $cnt{$_};  
       }  
     else  
       {  
  $returnhash{$_} = $sum{$_};  
       }  
   }  
   
         # 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(%newevaldata); # Close/release the new nohist database.  
       }  
     return(%returnhash);  
   }  
   
 =pod  
   
 B<wanted> - used by B<File::Find::find> subroutine.  
   
 This evaluates whether a file is wanted, and pushes it onto the  
 I<@metalist> array.  This subroutine was, for the most part, auto-generated  
 by the B<find2perl> command.  
   
 =over 4  
   
 Parameters:  
   
 =item I<$file> - a path to the file.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<boolean> - true or false based on logical statement.  
   
 =back  
   
 =cut  
   
 sub wanted ($)  
   {  
     (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&  
     -f $_ &&  
     /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&  
     push(@metalist,$File::Find::dir.'/'.$_);  
   }  
   
 =pod  
   
 B<get_metadata_from_file> - read xml-tagged file and return parsed metadata.  
   
 I<Note that this is significantly altered from a subroutine present in lonnet.>  
   
 =over 4  
   
 Parameters:  
   
 =item I<$file> - a path.to the file.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<hash reference> - a hash array (keys and values).  
   
 =back  
   
 =cut  
   
 sub get_metadata_from_file ($)  
   {  
     my ($filename) = @_;  
     my %metatable; # Used to store return value of hash-tabled metadata.  
     $filename = &declutter($filename); # Remove non-identifying filesystem info  
     my $uri = ''; # The URI is not relevant in this scenario.  
     unless ($filename =~ m/\.meta$/) # Unless ending with .meta.  
       {  
  $filename .= '.meta'; # Append a .meta suffix.  
       }  
     # Get the file contents.  
     my $metadata_string =  
  &get_file_contents($perlvar{'lonDocRoot'}.'/res/'.$filename);  
   
     # Parse the file based on its XML tags.  
     my $parser = HTML::TokeParser->new(\$metadata_string);  
     my $token;  
     while ($token = $parser->get_token) # Loop through tokens.  
       {  
  if ($token->[0] eq 'S') # If it is a start token.  
   {  
     my $entry = $token->[1];  
     my $unikey = $entry; # A unique identifier for this xml tag key.  
     if (defined($token->[2]->{'part'}))  
       {   
  $unikey .= '_'.$token->[2]->{'part'};   
       }  
     if (defined($token->[2]->{'name'}))  
       {   
  $unikey .= '_'.$token->[2]->{'name'};   
       }  
     # Append $unikey to metatable's keys entry.  
     if ($metatable{$uri.'keys'})  
       {  
  $metatable{$uri.'keys'} .= ','.$unikey;  
       }  
     else  
       {  
  $metatable{$uri.'keys'} = $unikey;  
       }  
     # Insert contents into metatable entry for the unikey.  
     foreach my $t3 (@{$token->[3]})  
       {  
  $metatable{$uri.''.$unikey.'.'.$_} = $token->[2]->{$t3};  
       }  
     # If there was no text contained inside the tags, set = default.  
     unless  
       (  
         $metatable{$uri.''.$unikey} = $parser->get_text('/'.$entry)  
       )  
       {  
  $metatable{$uri.''.$unikey} =  
     $metatable{$uri.''.$unikey.'.default'};  
       }  
   }  
       }  
     # Return with a key-value table of XML tags and their tag contents.  
     return(\%metatable);  
   }  
   
 =pod  
   
 B<get_file_contents> - returns either the contents of the file or a -1.  
   
 =over 4  
   
 Parameters:  
   
 =item I<$file> - a complete filesystem path.to the file.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<string> - file contents or a -1.  
   
 =back  # ------------------------------------- Only run if machine is a library server
   exit unless $perlvar{'lonRole'} eq 'library';
 =cut  
   
 sub get_file_contents ($)  
   {  
     my $file = shift(@_);  
   
     # If file does not exist, then return a -1 value.  
     unless (-e $file)  
       {  
  return(-1);  
       }  
   
     # Read in file contents.  
     my $file_handle = IO::File->new($file);  
     my $file_contents = '';  
     while (<$file_handle>)  
       {  
  $file_contents .= $_;  
       }  
   
     # Return file contents.  
     return($file_contents);  
   }  
   
 =pod  
   
 B<declutter> - Declutters URLs (remove extraneous prefixed filesystem path).  
   
 =over 4  
   
 Parameters:  
   
 =item I<$filesystem_path> - a complete filesystem path.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<string> - remnants of the filesystem path (beginning portion removed).  
   
 =back  
   
 =cut  
   
 sub declutter  
   {  
     my $filesystem_path = shift(@_);  
   
     # Remove beginning portions of the filesystem path.  
     $filesystem_path =~ s/^$perlvar{'lonDocRoot'}//;  
     $filesystem_path =~ s!^/!!;  
     $filesystem_path =~ s!^res/!!;  
   
     # Return what is remaining for the filesystem path.  
     return($filesystem_path);  
   }  
   
 =pod  
   
 B<query_home_server_status> - Is this the home server of an author's directory?  
   
 =over 4  
   
 Parameters:  
   
 =item I<$author_filesystem_path> - directory path for a user.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<boolean> - 1 if true; 0 if false.  
   
 =back  
   
 =cut  
   
 sub query_home_server_status ($)  
   {  
     my $author_filesystem_path = shift(@_);  
   
     # Remove beginning portion of this filesystem path.  
     $author_filesystem_path =~ s!/home/httpd/html/res/([^/]*)/([^/]*).*!$1/$2!;  
   
     # Construct path to the author's ordinary user directory.  
     my ($user_domain,$username) = split(m!/!,$author_filesystem_path);  
     my $user_directory_path = construct_path_to_user_directory($user_domain,  
        $username);  
   
     # Return status of whether the user directory path is defined.  
     if (-e $user_directory_path)  
       {  
  return(1); # True.  
       }  
     else  
       {  
         return(0); # False.  
       }  
   }  
   
 =pod  
   
 B<construct_path_to_user_directory> ($$) - makes a filesystem path to user dir.  
   
 =over 4  
   
 Parameters:  
   
 =item I<$user_domain> - the loncapa domain of the user.  
   
 =item I<$username> - the unique username (user id) of the user.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<string> - representing the path on the filesystem.  
   
 =back  
   
 =cut  
   
 sub construct_path_to_user_directory ($$)  
   {  
     my ($user_domain,$username) = @_;  
   
     # Untaint.  
     $user_domain =~ s/\W//g;  
     $username =~ s/\W//g;  
   
     # Create three levels of sub-directoried filesystem path  
     # based on the first three characters of the username.  
     my $sub_filesystem_path = $username.'__';  
     $sub_filesystem_path =~ s!(.)(.)(.).*!$1/$2/$3/!;  
   
     # Use the sub-directoried levels and other variables to generate  
     # the complete filesystem path.  
     my $complete_filesystem_path =  
  join('/',($perlvar{'lonUsersDir'},  
   $user_domain,  
   $sub_filesystem_path,  
   $username));  
   
     # Return the complete filesystem path.  
     return($complete_filesystem_path);  
   }  
   
 =pod  
   
 B<sql_formatted_time> (@) - turns seconds since epoch into datetime sql format.  
   
 =over 4  
   
 Parameters:  
   
 =item I<$epochtime> - time in seconds since epoch (may need to be sanitized).  
   
 =back  # ----------------------------- Make sure this process is running from user=www
   
 =over 4  my $wwwid=getpwnam('www');
   if ($wwwid!=$<) {
       my $emailto="$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";
       my $subj="LON: $perlvar{'lonHostID'} User ID mismatch";
       system("echo 'User ID mismatch. searchcat.pl must be run as user www.' |\
    mailto $emailto -s '$subj' > /dev/null");
       exit 1;
   }
   
 Returns:  
   
 =item C<string> - datetime sql formatted string.  # ---------------------------------------------------------- We are in business
   
 =back  open(LOG,'>'.$perlvar{'lonDaemons'}.'/logs/searchcat.log');
   print LOG '==== Searchcat Run '.localtime()."====\n\n";
   $simplestatus='time='.time.'&';
   my $dbh;
   # ------------------------------------- Make sure that database can be accessed
   {
       unless (
       $dbh = DBI->connect("DBI:mysql:loncapa","www",$perlvar{'lonSqlAccess'},{ RaiseError =>0,PrintError=>0})
       ) { 
    print LOG "Cannot connect to database!\n";
    $simplestatus.='mysql=defunct';
    &writesimple();
    exit;
       }
   
   # Make temporary table
       $dbh->do("DROP TABLE IF EXISTS newmetadata");
       my $make_metadata_table = "CREATE TABLE IF NOT EXISTS newmetadata (".
           "title TEXT, author TEXT, subject TEXT, url TEXT, keywords TEXT, ".
           "version TEXT, notes TEXT, abstract TEXT, mime TEXT, language TEXT, ".
           "creationdate DATETIME, lastrevisiondate DATETIME, owner TEXT, ".
           "copyright TEXT, dependencies TEXT, ".
    "modifyinguser TEXT, authorspace TEXT, ".
    "lowestgradelevel INTEGER UNSIGNED, highestgradelevel INTEGER UNSIGNED, ".
    "standards TEXT, ".
           "count INTEGER UNSIGNED, ".
           "course INTEGER UNSIGNED, course_list TEXT, ".
           "goto INTEGER UNSIGNED, goto_list TEXT, ".
           "comefrom INTEGER UNSIGNED, comefrom_list TEXT, ".
           "sequsage INTEGER UNSIGNED, sequsage_list TEXT, ".
           "stdno INTEGER UNSIGNED, stdno_list TEXT, ".
    "avetries FLOAT, avetries_list TEXT, ".
           "difficulty FLOAT, difficulty_list TEXT, ".
    "clear FLOAT, technical FLOAT, correct FLOAT, helpful FLOAT, depth FLOAT, ".
    "comments TEXT, ".
   # For backward compatibility, only insert new fields below
   # ...
   # For backward compatibility, end new fields above
           "FULLTEXT idx_title (title), ".
           "FULLTEXT idx_author (author), FULLTEXT idx_subject (subject), ".
           "FULLTEXT idx_url (url), FULLTEXT idx_keywords (keywords), ".
           "FULLTEXT idx_notes (notes), ".
           "FULLTEXT idx_abstract (abstract), FULLTEXT idx_mime (mime), ".
           "FULLTEXT idx_owner (owner), ".
    "FULLTEXT idx_standards (standards))".
           "TYPE=MyISAM";
       # It would sure be nice to have some logging mechanism.
       unless ($dbh->do($make_metadata_table)) {
    print LOG "\nMySQL Error Create: ".$dbh->errstr."\n";
    die $dbh->errstr;
       }
   }
   
   # ------------------------------------------------------------- get .meta files
   opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}");
   my @homeusers = grep {
       &ishome("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$_")
       } grep {!/^\.\.?$/} readdir(RESOURCES);
   closedir RESOURCES;
   
 =cut  #
   # Create the statement handlers we need
   
 sub sql_formatted_time ($)  my $insert_sth = $dbh->prepare
   {      ("INSERT INTO newmetadata VALUES (".
     # Sanitize the time argument and convert to localtime array.       "?,".   # title
        "?,".   # author
        "?,".   # subject
        "?,".   # declutter url
        "?,".   # version
        "?,".   # current
        "?,".   # notes
        "?,".   # abstract
        "?,".   # mime
        "?,".   # language
        "?,".   # creationdate
        "?,".   # revisiondate
        "?,".   # owner
        "?,".   # copyright
        "?,".   # dependencies
        "?,".   # modifyinguser
        "?,".   # authorspace
        "?,".   # lowestgradelevel
        "?,".   # highestgradelevel
        "?,".   # standards
        "?,".   # count
        "?,".   # course
        "?,".   # course_list
        "?,".   # goto
        "?,".   # goto_list
        "?,".   # comefrom
        "?,".   # comefrom_list
        "?,".   # usage
        "?,".   # usage_list
        "?,".   # stdno
        "?,".   # stdno_list
        "?,".   # avetries
        "?,".   # avetries_list
        "?,".   # difficulty
        "?,".   # difficulty_list
        "?,".   # clear
        "?,".   # technical
        "?,".   # correct
        "?,".   # helpful
        "?,".   # depth
        "?".    # comments
        ")"
        );
   
   foreach my $user (@homeusers) {
       print LOG "\n=== User: ".$user."\n\n";
   
       my $prodir=&propath($perlvar{'lonDefDomain'},$user);
       # Use find.pl
       undef @metalist;
       @metalist=();
       &find("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$user");
       # -- process each file to get metadata and put into search catalog SQL
       # database.  Also, check to see if already there.
       # I could just delete (without searching first), but this works for now.
       foreach my $m (@metalist) {
           print LOG "- ".$m."\n";
           my $ref=&metadata($m);
           my $m2='/res/'.&declutter($m);
           $m2=~s/\.meta$//;
     if ($ref->{'obsolete'}) { print LOG "obsolete\n"; next; }
    if ($ref->{'copyright'} eq 'private') { print LOG "private\n"; next; }
    my %dyn=();
    if ($m2=~/\/default$/) {
       $m2=~s/\/default$/\//;
    } else {
       %dyn=&dynamicmeta($m2);
       &count($m2);
    }
           unless ($insert_sth->execute(
        $ref->{'title'},
                                $ref->{'author'},
                                $ref->{'subject'},
                                $m2,
                                $ref->{'keywords'},
                                'current',
                                $ref->{'notes'},
                                $ref->{'abstract'},
                                $ref->{'mime'},
                                $ref->{'language'},
                                sqltime($ref->{'creationdate'}),
                                sqltime($ref->{'lastrevisiondate'}),
                                $ref->{'owner'},
                                $ref->{'copyright'},
        $ref->{'dependencies'},
        $ref->{'modifyinguser'},
        $ref->{'authorspace'},
        $ref->{'lowestgradelevel'},
        $ref->{'highestgradelevel'},
        $ref->{'standards'},
        $dyn{'count'},
        $dyn{'course'},
        $dyn{'course_list'},
        $dyn{'goto'},
        $dyn{'goto_list'},
        $dyn{'comefrom'},
        $dyn{'comefrom_list'},
        $dyn{'usage'},
        $dyn{'usage_list'},
        $dyn{'stdno'},
        $dyn{'stdno_list'},
        $dyn{'avetries'},
        $dyn{'avetries_list'},
        $dyn{'difficulty'},
        $dyn{'difficulty_list'},     
        $dyn{'clear'},
        $dyn{'technical'},
        $dyn{'correct'},
        $dyn{'helpful'},
        $dyn{'depth'},
        $dyn{'comments'}     
        )) {
       print LOG "\nMySQL Error Insert: ".$dbh->errstr."\n";
       die $dbh->errstr;
    }
           $ref = undef;
       }
   }
   # --------------------------------------------------- Close database connection
   $dbh->do("DROP TABLE IF EXISTS metadata");
   unless ($dbh->do("RENAME TABLE newmetadata TO metadata")) {
       print LOG "\nMySQL Error Rename: ".$dbh->errstr."\n";
       die $dbh->errstr;
   }
   unless ($dbh->disconnect) {
       print LOG "\nMySQL Error Disconnect: ".$dbh->errstr."\n";
       die $dbh->errstr;
   }
   print LOG "\n==== Searchcat completed ".localtime()." ====\n";
   close(LOG);
   &writesimple();
   &writecount();
   exit 0;
   
   
   
   # =============================================================================
   
   # ---------------------------------------------------------------- Get metadata
   # significantly altered from subroutine present in lonnet
   sub metadata {
       my ($uri,$what)=@_;
       my %metacache=();
       $uri=&declutter($uri);
       my $filename=$uri;
       $uri=~s/\.meta$//;
       $uri='';
       unless ($metacache{$uri.'keys'}) {
           unless ($filename=~/\.meta$/) { $filename.='.meta'; }
    my $metastring=&getfile($perlvar{'lonDocRoot'}.'/res/'.$filename);
           my $parser=HTML::TokeParser->new(\$metastring);
           my $token;
           while ($token=$parser->get_token) {
               if ($token->[0] eq 'S') {
                   my $entry=$token->[1];
                   my $unikey=$entry;
                   if (defined($token->[2]->{'part'})) { 
                       $unikey.='_'.$token->[2]->{'part'}; 
                   }
                   if (defined($token->[2]->{'name'})) { 
                       $unikey.='_'.$token->[2]->{'name'}; 
                   }
                   if ($metacache{$uri.'keys'}) {
                       $metacache{$uri.'keys'}.=','.$unikey;
                   } else {
                       $metacache{$uri.'keys'}=$unikey;
                   }
                   map {
                       $metacache{$uri.''.$unikey.'.'.$_}=$token->[2]->{$_};
                   } @{$token->[3]};
                   unless (
                           $metacache{$uri.''.$unikey}=$parser->get_text('/'.$entry)
                           ) { $metacache{$uri.''.$unikey}=
                                   $metacache{$uri.''.$unikey.'.default'};
                           }
               }
           }
       }
       return \%metacache;
   }
   
   # ------------------------------------------------------------ Serves up a file
   # returns either the contents of the file or a -1
   sub getfile {
       my $file=shift;
       if (! -e $file ) { return -1; };
       my $fh=IO::File->new($file);
       my $a='';
       while (<$fh>) { $a .=$_; }
       return $a;
   }
   
   # ------------------------------------------------------------- Declutters URLs
   sub declutter {
       my $thisfn=shift;
       $thisfn=~s/^$perlvar{'lonDocRoot'}//;
       $thisfn=~s/^\///;
       $thisfn=~s/^res\///;
       return $thisfn;
   }
   
   # --------------------------------------- Is this the home server of an author?
   # (copied from lond, modification of the return value)
   sub ishome {
       my $author=shift;
       $author=~s/\/home\/httpd\/html\/res\/([^\/]*)\/([^\/]*).*/$1\/$2/;
       my ($udom,$uname)=split(/\//,$author);
       my $proname=propath($udom,$uname);
       if (-e $proname) {
    return 1;
       } else {
           return 0;
       }
   }
   
   # -------------------------------------------- Return path to profile directory
   # (copied from lond)
   sub propath {
       my ($udom,$uname)=@_;
       $udom=~s/\W//g;
       $uname=~s/\W//g;
       my $subdir=$uname.'__';
       $subdir =~ s/(.)(.)(.).*/$1\/$2\/$3/;
       my $proname="$perlvar{'lonUsersDir'}/$udom/$subdir/$uname";
       return $proname;
   } 
   
   # ---------------------------- convert 'time' format into a datetime sql format
   sub sqltime {
       my $time=&unsqltime(@_[0]);
       unless ($time) { return 'NULL'; }
     my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =      my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
  localtime(&sanitize_time(shift(@_)));   localtime($time);
       $mon++; $year+=1900;
     # Convert month from (0..11) to (1..12).      return "$year-$mon-$mday $hour:$min:$sec";
     $mon += 1;  }
   
     # Make the year compatible with A.D. specification.  sub maketime {
     $year += 1900;      my %th=@_;
       return POSIX::mktime(($th{'seconds'},$th{'minutes'},$th{'hours'},
     # Return a date which is compatible with MySQL's "DATETIME" format.                            $th{'day'},$th{'month'}-1,
     return(join('-',($year,$mon,$mday)).                            $th{'year'}-1900,0,0,$th{'dlsav'}));
    ' '.  }
    join(':',($hour,$min,$sec))  
    );  
   }  
   
   
 # ==================================== The following two subroutines are needed  
 #                 for accommodating incorrect time formats inside the metadata.  
   
 =pod  #########################################
   #
 B<make_seconds_since_epoch> (@) - turns time metadata into seconds since epoch.  # Retro-fixing of un-backward-compatible time format
   
 =over 4  
   
 Parameters:  
   
 =item I<%time_metadata> - a key-value listing characterizing month, year, etc.  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<integer> - seconds since epoch.  
   
 =back  
   
 =cut  
   
 sub make_seconds_since_epoch (@)  
   {  
     # Keytable of time metadata.  
     my %time_metadata = @_;  
   
     # Return seconds since the epoch (January 1, 1970, 00:00:00 UTC).  
     return(POSIX::mktime(  
  ($time_metadata{'seconds'},  
   $time_metadata{'minutes'},  
   $time_metadata{'hours'},  
   $time_metadata{'day'},  
   $time_metadata{'month'}-1,  
   $time_metadata{'year'}-1900,  
   0,  
   0,  
   $time_metadata{'dlsav'})  
  )  
    );  
   }  
   
 =pod  
   
 B<sanitize_time> - if time looks sql-formatted, make it seconds since epoch.  
   
 Somebody described this subroutine as  
 "retro-fixing of un-backward-compatible time format".  
   
 What this means, is that a part of this code expects to get UTC seconds  
 since the epoch (beginning of 1970).  Yet, some of the .meta files have  
 sql-formatted time strings (2001-04-01, etc.) instead of seconds-since-epoch  
 integers (e.g. 1044147435).  These time strings do not encode the timezone  
 and, in this sense, can be considered "un-backwards-compatible".  
   
 =over 4  
   
 Parameters:  
   
 =item I<$potentially_badformat_string> - string to "retro-fix".  
   
 =back  
   
 =over 4  
   
 Returns:  
   
 =item C<integer> - seconds since epoch.  
   
 =back  
   
 =cut  
   
 sub sanitize_time ($)  
   {  
     my $timestamp = shift(@_);  
     # If timestamp is in this unexpected format....  
     if ($timestamp =~ /^(\d+)\-(\d+)\-(\d+)\s+(\d+)\:(\d+)\:(\d+)$/)  
       {  
  # then convert into seconds since epoch (the expected format).  
  $timestamp = &make_seconds_since_epoch(  
        'year' => $1,  
        'month' => $2,  
        'day' => $3,  
        'hours' => $4,  
        'minutes' => $5,  
        'seconds' => $6  
        );  
       }  
     # Otherwise we assume timestamp to be as expected.  
     return($timestamp);  
   }  
   
 =pod  
   
 =head1 AUTHOR  
   
 Written to help the loncapa project.  sub unsqltime {
       my $timestamp=shift;
       if ($timestamp=~/^(\d+)\-(\d+)\-(\d+)\s+(\d+)\:(\d+)\:(\d+)$/) {
           $timestamp=&maketime('year'=>$1,'month'=>$2,'day'=>$3,
                                'hours'=>$4,'minutes'=>$5,'seconds'=>$6);
       }
       return $timestamp;
   }
   
 Scott Harrison, sharrison@users.sourceforge.net  # ----------------- Code to enable 'find' subroutine listing of the .meta files
   
 This is distributed under the same terms as loncapa (i.e. "freeware").  no strict "vars";
   
 =cut  sub wanted {
       (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&
            -f _ &&
            /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&
            push(@metalist,"$dir/$_");
   }

Removed from v.1.29  
changed lines
  Added in v.1.54


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