--- loncom/metadata_database/searchcat.pl 2002/05/17 14:03:04 1.18 +++ loncom/metadata_database/searchcat.pl 2002/10/08 16:59:26 1.22 @@ -2,7 +2,7 @@ # The LearningOnline Network # searchcat.pl "Search Catalog" batch script # -# $Id: searchcat.pl,v 1.18 2002/05/17 14:03:04 matthew Exp $ +# $Id: searchcat.pl,v 1.22 2002/10/08 16:59:26 www Exp $ # # Copyright Michigan State University Board of Trustees # @@ -44,8 +44,112 @@ use LONCAPA::Configuration; use IO::File; use HTML::TokeParser; use DBI; +use GDBM_File; my @metalist; + + +# ----------------------------------------------------- Un-Escape Special Chars + +sub unescape { + my $str=shift; + $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg; + return $str; +} + +# -------------------------------------------------------- Escape Special Chars + +sub escape { + my $str=shift; + $str =~ s/(\W)/"%".unpack('H2',$1)/eg; + return $str; +} + + +# ------------------------------------------- Code to evaluate dynamic metadata + +sub dynamicmeta { +# +# +# Do nothing for now ... +# +# + return; +# +# ..., but stuff below already works +# + my $url=&declutter(shift); + $url=~s/\.meta$//; + my %returnhash=(); + my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//); + my $prodir=&propath($adomain,$aauthor); + if (tie(%evaldata,'GDBM_File', + $prodir.'/nohist_resevaldata.db',&GDBM_WRCREAT,0640)) { + my %sum=(); + my %cnt=(); + my %listitems=('count' => 'add', + 'course' => 'add', + 'avetries' => 'avg', + 'stdno' => 'add', + 'difficulty' => 'avg', + 'clear' => 'avg', + 'technical' => 'avg', + 'helpful' => 'avg', + 'correct' => 'avg', + 'depth' => 'avg', + 'comments' => 'app', + 'usage' => 'cnt' + ); + my $regexp=$url; + $regexp=~s/(\W)/\\$1/g; + $regexp='___'.$regexp.'___([a-z]+)$'; + foreach (keys %evaldata) { + my $key=&unescape($_); + if ($key=~/$regexp/) { + my $ctype=$1; + if (defined($cnt{$ctype})) { + $cnt{$ctype}++; + } else { + $cnt{$ctype}=1; + } + unless ($listitems{$ctype} eq 'app') { + if (defined($sum{$ctype})) { + $sum{$ctype}+=$evaldata{$_}; + } else { + $sum{$ctype}=$evaldata{$_}; + } + } else { + if (defined($sum{$ctype})) { + if ($evaldata{$_}) { + $sum{$ctype}.='
'.$evaldata{$_}; + } + } else { + $sum{$ctype}=''.$evaldata{$_}; + } + } + if ($ctype eq 'count') { + delete($evaldata{$_}); + } + } + } + foreach (keys %cnt) { + if ($listitems{$_} eq 'avg') { + $returnhash{$_}=int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0; + } elsif ($listitems{$_} eq 'cnt') { + $returnhash{$_}=$cnt{$_}; + } else { + $returnhash{$_}=$sum{$_}; + } + } + if ($returnhash{'count'}) { + my $newkey=$$.'_'.time.'_searchcat___'.&escape($url).'___count'; + $evaldata{$newkey}=$returnhash{'count'}; + } + untie(%evaldata); + } + return %returnhash; +} + # ----------------- Code to enable 'find' subroutine listing of the .meta files require "find.pl"; sub wanted { @@ -56,8 +160,7 @@ sub wanted { } # --------------- Read loncapa_apache.conf and loncapa.conf and get variables -my $perlvarref=LONCAPA::Configuration::read_conf('loncapa_apache.conf', - 'loncapa.conf'); +my $perlvarref=LONCAPA::Configuration::read_conf('loncapa.conf'); my %perlvar=%{$perlvarref}; undef $perlvarref; # remove since sensitive and not needed delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed @@ -74,6 +177,19 @@ my $dbh; print "Cannot connect to database!\n"; exit; } + 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, 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"; + # It would sure be nice to have some logging mechanism. + $dbh->do($make_metadata_table); } # ------------------------------------------------------------- get .meta files @@ -93,6 +209,7 @@ foreach my $m (@metalist) { my $ref=&metadata($m); my $m2='/res/'.&declutter($m); $m2=~s/\.meta$//; + &dynamicmeta($m2); my $q2="select * from metadata where url like binary '$m2'"; my $sth = $dbh->prepare($q2); $sth->execute();