Return to searchcat.pl CVS log

Up to [LON-CAPA] / loncom / metadata_database

Annotation of loncom/metadata_database/searchcat.pl, revision 1.14

1.1       harris41    1: #!/usr/bin/perl
                      2: # The LearningOnline Network
                      3: # searchcat.pl "Search Catalog" batch script
                      4: 
                      5: # 04/14/2001 Scott Harrison
                      6: 
                      7: # This script goes through a LON-CAPA resource
                      8: # directory and gathers metadata.
                      9: # The metadata is entered into a SQL database.
                     10: 
                     11: use IO::File;
                     12: use HTML::TokeParser;
1.6       harris41   13: use DBI;
1.1       harris41   14: 
                     15: my @metalist;
                     16: # ----------------- Code to enable 'find' subroutine listing of the .meta files
                     17: require "find.pl";
                     18: sub wanted {
                     19:     (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&
                     20:     -f _ &&
1.10      harris41   21:     /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&
1.1       harris41   22:     push(@metalist,"$dir/$_");
                     23: }
                     24: 
                     25: # ------------------------------------ Read httpd access.conf and get variables
                     26: open (CONFIG,"/etc/httpd/conf/access.conf") || die "Can't read access.conf";
                     27: 
                     28: while ($configline=<CONFIG>) {
                     29:     if ($configline =~ /PerlSetVar/) {
                     30: 	my ($dummy,$varname,$varvalue)=split(/\s+/,$configline);
                     31:         chomp($varvalue);
                     32:         $perlvar{$varname}=$varvalue;
                     33:     }
                     34: }
                     35: close(CONFIG);
                     36: 
1.3       harris41   37: my $dbh;
1.1       harris41   38: # ------------------------------------- Make sure that database can be accessed
                     39: {
                     40:     unless (
                     41: 	    $dbh = DBI->connect("DBI:mysql:loncapa","www",$perlvar{'lonSqlAccess'},{ RaiseError =>0,PrintError=>0})
                     42: 	    ) { 
                     43: 	print "Cannot connect to database!\n";
                     44: 	exit;
                     45:     }
                     46: }
                     47: 
                     48: # ------------------------------------------------------------- get .meta files
1.2       harris41   49: opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}");
                     50: my @homeusers=grep
                     51:           {&ishome("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$_")}
                     52:           grep {!/^\.\.?$/} readdir(RESOURCES);
                     53: closedir RESOURCES;
                     54: foreach my $user (@homeusers) {
                     55:     &find("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$user");
                     56: }
1.1       harris41   57: 
                     58: # -- process each file to get metadata and put into search catalog SQL database
1.9       harris41   59: # Also, check to see if already there.
1.11      harris41   60: # I could just delete (without searching first), but this works for now.
1.1       harris41   61: foreach my $m (@metalist) {
                     62:     my $ref=&metadata($m);
1.11      harris41   63:     my $m2='/res/'.&declutter($m);
1.12      harris41   64:     $m2=~s/\.meta$//;
                     65:     my $q2="select * from metadata where url like binary '$m2'";
1.9       harris41   66:     my $sth = $dbh->prepare($q2);
                     67:     $sth->execute();
                     68:     my $r1=$sth->fetchall_arrayref;
                     69:     if (@$r1) {
1.12      harris41   70: 	$sth=$dbh->prepare("delete from metadata where url like binary '$m2'");
1.9       harris41   71:         $sth->execute();
                     72:     }
                     73:     $sth=$dbh->prepare('insert into metadata values ('.
1.8       harris41   74: 			  '"'.delete($ref->{'title'}).'"'.','.
                     75: 			  '"'.delete($ref->{'author'}).'"'.','.
                     76: 			  '"'.delete($ref->{'subject'}).'"'.','.
1.12      harris41   77: 			  '"'.$m2.'"'.','.
1.8       harris41   78: 			  '"'.delete($ref->{'keywords'}).'"'.','.
1.9       harris41   79: 			  '"'.'current'.'"'.','.
1.8       harris41   80: 			  '"'.delete($ref->{'notes'}).'"'.','.
                     81: 			  '"'.delete($ref->{'abstract'}).'"'.','.
                     82: 			  '"'.delete($ref->{'mime'}).'"'.','.
                     83: 			  '"'.delete($ref->{'language'}).'"'.','.
1.13      harris41   84: 			  '"'.sqltime(delete($ref->{'creationdate'})).'"'.','.
                     85: 			  '"'.sqltime(delete($ref->{'lastrevisiondate'})).'"'.','.
1.8       harris41   86: 			  '"'.delete($ref->{'owner'}).'"'.','.
                     87: 			  '"'.delete($ref->{'copyright'}).'"'.')');
1.1       harris41   88:     $sth->execute();
                     89: }
                     90: 
                     91: # ----------------------------------------------------------- Clean up database
                     92: # Need to, perhaps, remove stale SQL database records.
                     93: # ... not yet implemented
                     94: 
                     95: # --------------------------------------------------- Close database connection
                     96: $dbh->disconnect;
                     97: 
                     98: # ---------------------------------------------------------------- Get metadata
                     99: # significantly altered from subroutine present in lonnet
                    100: sub metadata {
                    101:     my ($uri,$what)=@_;
                    102:     my %metacache;
                    103:     $uri=&declutter($uri);
                    104:     my $filename=$uri;
                    105:     $uri=~s/\.meta$//;
                    106:     $uri='';
                    107:     unless ($metacache{$uri.'keys'}) {
                    108:         unless ($filename=~/\.meta$/) { $filename.='.meta'; }
                    109: 	my $metastring=&getfile($perlvar{'lonDocRoot'}.'/res/'.$filename);
                    110:         my $parser=HTML::TokeParser->new(\$metastring);
                    111:         my $token;
                    112:         while ($token=$parser->get_token) {
                    113:            if ($token->[0] eq 'S') {
                    114: 	      my $entry=$token->[1];
                    115:               my $unikey=$entry;
                    116:               if (defined($token->[2]->{'part'})) { 
                    117:                  $unikey.='_'.$token->[2]->{'part'}; 
                    118: 	      }
                    119:               if (defined($token->[2]->{'name'})) { 
                    120:                  $unikey.='_'.$token->[2]->{'name'}; 
                    121: 	      }
                    122:               if ($metacache{$uri.'keys'}) {
                    123:                  $metacache{$uri.'keys'}.=','.$unikey;
                    124:               } else {
                    125:                  $metacache{$uri.'keys'}=$unikey;
                    126: 	      }
                    127:               map {
                    128: 		  $metacache{$uri.''.$unikey.'.'.$_}=$token->[2]->{$_};
                    129:               } @{$token->[3]};
                    130:               unless (
                    131:                  $metacache{$uri.''.$unikey}=$parser->get_text('/'.$entry)
                    132: 		      ) { $metacache{$uri.''.$unikey}=
                    133: 			      $metacache{$uri.''.$unikey.'.default'};
                    134: 		      }
                    135:           }
                    136:        }
                    137:     }
                    138:     return \%metacache;
                    139: }
                    140: 
                    141: # ------------------------------------------------------------ Serves up a file
                    142: # returns either the contents of the file or a -1
                    143: sub getfile {
                    144:   my $file=shift;
                    145:   if (! -e $file ) { return -1; };
                    146:   my $fh=IO::File->new($file);
                    147:   my $a='';
                    148:   while (<$fh>) { $a .=$_; }
                    149:   return $a
                    150: }
                    151: 
                    152: # ------------------------------------------------------------- Declutters URLs
                    153: sub declutter {
                    154:     my $thisfn=shift;
                    155:     $thisfn=~s/^$perlvar{'lonDocRoot'}//;
                    156:     $thisfn=~s/^\///;
                    157:     $thisfn=~s/^res\///;
                    158:     return $thisfn;
                    159: }
1.2       harris41  160: 
                    161: # --------------------------------------- Is this the home server of an author?
                    162: # (copied from lond, modification of the return value)
                    163: sub ishome {
                    164:     my $author=shift;
                    165:     $author=~s/\/home\/httpd\/html\/res\/([^\/]*)\/([^\/]*).*/$1\/$2/;
                    166:     my ($udom,$uname)=split(/\//,$author);
                    167:     my $proname=propath($udom,$uname);
                    168:     if (-e $proname) {
                    169: 	return 1;
                    170:     } else {
                    171:         return 0;
                    172:     }
                    173: }
                    174: 
                    175: # -------------------------------------------- Return path to profile directory
                    176: # (copied from lond)
                    177: sub propath {
                    178:     my ($udom,$uname)=@_;
                    179:     $udom=~s/\W//g;
                    180:     $uname=~s/\W//g;
                    181:     my $subdir=$uname.'__';
                    182:     $subdir =~ s/(.)(.)(.).*/$1\/$2\/$3/;
                    183:     my $proname="$perlvar{'lonUsersDir'}/$udom/$subdir/$uname";
                    184:     return $proname;
                    185: } 
1.13      harris41  186: 
                    187: # ---------------------------- convert 'time' format into a datetime sql format
                    188: sub sqltime {
                    189:     my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
                    190: 	localtime(@_[0]);
1.14    ! harris41  191:     $mon++; $year+=1900;
1.13      harris41  192:     return "$year-$mon-$mday $hour:$min:$sec";
                    193: }