biblicit 1.0 → 2.0.3

data/parscit/lib/ParsCit/CitationContext.pm

@@ -0,0 +1,220 @@
+package ParsCit::CitationContext;
+
+###
+# Utility functions for finding citation contexts after citation
+# parsing.  The idea is to build regular expressions that could match
+# citation references in the body text and scan the body text for
+# occurrences.  The size of the context returned can be configured
+# in ParsCit::Config::context_radius.
+#
+# Isaac Councill, 7/20/07
+###
+use utf8;
+
+use ParsCit::Config;
+
+my $context_radius	= $ParsCit::Config::contextRadius;
+my $max_contexts	= $ParsCit::Config::maxContexts;
+
+sub CountTokens 
+{
+	my ($text) = @_;
+
+  	my $trim_text = $text;
+
+  	$trim_text =~ s/^\s+//;       
+	$trim_text =~ s/\s+$//;
+
+	my @tokens = split(/\s+/, $trim_text);
+  	return scalar(@tokens);
+}
+
+###
+# Build a list of potential regular expressions based on the supplied
+# citation marker and applies the expressions to the specified body text.
+# Returns a reference to a list of strings that match the expressions,
+# expanded to radius $context_radius.
+###
+sub GetCitationContext 
+{
+	my ($rbody_text, $pos_array, $marker) = @_;
+  
+	my ($prioritize, @markers) = GuessPossibleMarkers($marker);
+  
+  	my @matches		= ();
+	# Thang Nov 2009: store the in-text citation strings
+  	my @cit_strs	= (); 	
+	# Thang May 2010: store start word positions of citation markers
+  	my @start_word_positions	= ();
+	# Thang May 2010: store end word positions of citation markers
+  	my @end_word_positions		= (); 
+
+	###
+  	# Modified by Nick Friedrich
+	###
+  	my @positions		= ();
+  	my $position		= undef;  
+	my $contexts_found	= 0;
+
+  	foreach my $mark (@markers) 
+	{
+    	while (($contexts_found < $max_contexts) && $$rbody_text =~ m/(.{$context_radius}($mark).{$context_radius})/gs) 
+		{
+      		# Thang May 2010: check mark length
+      		my $cit_str		= $2;
+      		my $cit_length	= CountTokens($cit_str);
+      
+	  		if($cit_length == 0) { next; }
+
+			push @positions, (pos $$rbody_text) - $context_radius;
+      		push @matches, $1;
+
+      		my $before_mark		= substr($$rbody_text, 0, (pos $$rbody_text) - $context_radius-length($cit_str));
+      		my $before_length	= CountTokens($before_mark);
+
+			if($before_mark =~ / $/ || $cit_str =~ /^ /)
+			{
+				push @start_word_positions, $before_length;
+				push @end_word_positions, $before_length + $cit_length - 1;
+			} 
+			else 
+			{ 
+				push @start_word_positions, $before_length - 1;
+				push @end_word_positions, $before_length - 1 + $cit_length - 1;
+      		}
+
+			# Thang Nov 2009
+      		if($cit_str !~ /\(.+\)$/ && $cit_str =~ /\)$/)			
+			{
+				# Trim ending bracket
+				$cit_str =~ s/\)$//; 
+      		}
+      		push @cit_strs, $cit_str;
+      		# End Thang Nov 2009
+      
+      		$contexts_found++;
+    	}
+
+    	if (($prioritize > 0) && ($#matches >= 0)) { last; }
+  	}
+  	return ( \@matches, \@positions, \@start_word_positions, \@end_word_positions, \@cit_strs);
+}
+
+###
+# Builds a list of regular expressions based on the supplied
+# citation marker that may indicate a citation reference in
+# body text.  The first value returned is a parameter indicating
+# whether the list should be prioritized (i.e., if matches to
+# one element are found, don't try to match subsquent expressions),
+# with 0 indicating no prioritization and 1 indicating otherwise.
+#
+# HISTORY: Nick (v081201)
+#
+# In publications with unmarked citations the citations are not only
+# referenced by markers like (Bottou, 1991). Often it is referenced like
+# "Leon Bottou (1991) discussed two solutions". Getting Parscit to find
+# both styles is easy. Actually ParsCit creates regular expressions like
+# 'Bottou,? 1991' or 'Beel and Gipp,? 2008' . The regular expressions in
+# CitationContext::guessPossibleMarkers (line 75, 78, 79 91 - 79, 82,
+# 83, 95 in my modifies CitationContext.pm) need only to be modified to
+# match parenthesis around the year (e.g. 'Bottou,? \(?1991\)?' ).
+###
+sub GuessPossibleMarkers 
+{
+    my $marker = shift;
+
+    if ($marker =~ m/^([\[\(])([\p{IsUpper}\p{IsLower}\-\d]+)([\]\)])/) 
+	{
+		my $open	= MakeSafe($1);
+		my $mark	= MakeSafe($2);
+		my $close	= MakeSafe($3);
+
+		my $ref_indicator = "$open([\\p{IsUpper}\\p{IsLower}\\-\\d]+[;,] *)*$mark([;,] *[\\p{IsUpper}\\p{IsLower}\\-\\d]+)*$close";
+		return (0, $ref_indicator);
+    }
+
+    if ($marker =~ m/^(\d+)\.?/) 
+	{
+		my $square	= "\\[(\\d+[,;] *)*$1([;,] *\\d+)*\\]";
+		my $paren	= "\\((\\d+[,;] *)*$1([,;] *\\d+)*\\)";
+
+		return (1, $square, $paren);
+    }
+
+    if ($marker =~ m/^([\p{IsUpper}\p{IsLower}\-\.\']+(, )*)+\d{4}/) 
+	{
+		my @tokens	= split ", ", $marker;
+		my $year	= $tokens[$#tokens];
+		my @names	= @tokens[0..$#tokens-1];
+
+		my @possible_markers = ();
+	
+		if ($#names == 0) 
+		{
+	    	push @possible_markers, $names[0].",? \\(?$year\\)?";
+		}
+	
+		if ($#names == 1) 
+		{
+	    	push @possible_markers, $names[0]." and ".$names[1].",? \\(?$year\\)?";
+	    	push @possible_markers, $names[0]." & ".$names[1].",? \\(?$year\\)?";
+		}
+		
+		if ($#names > 1) 
+		{
+	    	map { $_ = MakeSafe($_) } $names;
+	    	map { $_ = $_."," } @names;
+	    	
+			my $last_auth1 = "and ".$names[$#names];
+	    	my $last_auth2 = "& ".$names[$#names];
+	    
+			push @possible_markers,
+	      	join " ", @names[0..$#names-1], $last_auth1, $year;
+	    	push @possible_markers,
+	      	join " ", @names[0..$#names-1], $last_auth2, $year;
+
+	    	push @possible_markers, $names[0]."? et al\\\.?,? \\(?$year\\)?";
+		}
+
+		for (my $i = 0; $i <= $#possible_markers; $i++) 
+		{
+	    	my $safe_marker	= $possible_markers[$i];
+	    	$safe_marker	=~ s/\-/\\\-/g;
+	    	$possible_markers[$i] = $safe_marker;
+		}
+	
+		return (0, @possible_markers);
+    }
+
+    return MakeSafe($marker);
+}
+
+###
+# Prepare strings for safe inclusion within a regular expression,
+# escaping control characters.
+###
+sub MakeSafe 
+{
+    my $marker = shift;
+
+	if (! defined $marker) { return ""; }
+
+    $marker =~ s/\\/\\\\/g;
+    $marker =~ s/\-/\\\-/g;
+    $marker =~ s/\[/\\\[/g;
+    $marker =~ s/\]/\\\]/g;
+    $marker =~ s/\(/\\\(/g;
+    $marker =~ s/\)/\\\)/g;
+    $marker =~ s/\'/\\\'/g;
+    $marker =~ s/\"/\\\"/g;
+    $marker =~ s/\+//g;
+    $marker =~ s/\?//g;
+    $marker =~ s/\*//g;
+    $marker =~ s/\^//g;
+    $marker =~ s/\$//g;
+    $marker =~ s/\./\\\./g;
+
+    return $marker;
+}
+
+1;

data/parscit/lib/ParsCit/Config.pm

@@ -0,0 +1,35 @@
+package ParsCit::Config;
+
+## Global
+$algorithmName		= "ParsCit";
+$algorithmVersion	= "110505";
+
+## Repository Mappings
+%repositories	= (	'rep1'		=> '/repositories/rep1',
+                 	'example1'	=> '/',
+		 		 	'example2'	=> '/home');
+
+## WS settings
+$serverURL	= '130.203.133.46';
+$serverPort	= 10555;
+$URI		= 'http://citeseerx.org/algorithms/parscit/wsdl';
+
+## Tr2crfpp
+## Paths relative to ParsCit root dir ($FindBin::Bin/..)
+$tmpDir		= "tmp";
+$dictFile	= "resources/parsCitDict.txt";
+$crf_test = "$ENV{'CRFPP_HOME'}/bin/crf_test";
+
+## Tr2crfpp parscit model
+$modelFile		= "resources/parsCit.model";
+## Tr2crfpp parscit split reference model
+$splitModelFile	= "resources/parsCit.split.model";
+
+## Citation Context
+$contextRadius	= 200;
+$maxContexts	= 5;
+
+## Write citation and body file components
+$bWriteSplit	= 1;
+
+1;