Source code
Revision control
Copy as Markdown
Other Tools
#!/usr/bin/perl
# -*- Mode: Perl; tab-width: 2; indent-tabs-mode: nil; -*-
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
use XML::LibXSLT;
use XML::LibXML;
use LWP::Simple;
# output files
$FILE_UNICODE = "unicode.xml";
$FILE_DICTIONARY = "dictionary.xml";
$FILE_DIFFERENCES = "differences.txt";
$FILE_NEW_DICTIONARY = "new_dictionary.txt";
$FILE_SYNTAX_ERRORS = "syntax_errors.txt";
# our dictionary (property file)
$MOZ_DICTIONARY = "mathfont.properties";
# dictionary provided by the W3C in "XML Entity Definitions for Characters"
# XSL stylesheet to extract relevant data from the dictionary
$DICTIONARY_XSL = "operatorDictionary.xsl";
# dictionary provided by the W3C transformed with operatorDictionary.xsl
$WG_DICTIONARY = $FILE_DICTIONARY;
if (!($#ARGV >= 0 &&
((($ARGV[0] eq "download") && $#ARGV <= 1) ||
(($ARGV[0] eq "compare") && $#ARGV <= 1) ||
(($ARGV[0] eq "check") && $#ARGV <= 0) ||
(($ARGV[0] eq "clean") && $#ARGV <= 0)))) {
&usage;
}
if ($ARGV[0] eq "download") {
if ($#ARGV == 1) {
$WG_DICTIONARY_URL = $ARGV[1];
}
print "Downloading $WG_DICTIONARY_URL...\n";
getstore($WG_DICTIONARY_URL, $FILE_UNICODE);
print "Converting $FILE_UNICODE into $FILE_DICTIONARY...\n";
my $xslt = XML::LibXSLT->new();
my $source = XML::LibXML->load_xml(location => $FILE_UNICODE);
my $style_doc = XML::LibXML->load_xml(location => $DICTIONARY_XSL,
no_cdata=>1);
my $stylesheet = $xslt->parse_stylesheet($style_doc);
my $results = $stylesheet->transform($source);
open($file, ">$FILE_DICTIONARY") || die ("Couldn't open $FILE_DICTIONARY!");
print $file $stylesheet->output_as_bytes($results);
close($file);
exit 0;
}
if ($ARGV[0] eq "clean") {
unlink($FILE_UNICODE,
$FILE_DICTIONARY,
$FILE_DIFFERENCES,
$FILE_NEW_DICTIONARY,
$FILE_SYNTAX_ERRORS);
exit 0;
}
if ($ARGV[0] eq "compare" && $#ARGV == 1) {
$WG_DICTIONARY = $ARGV[1];
}
################################################################################
# structure of the dictionary used by this script:
# - key: same as in mathfont.properties
# - table:
# index | value
# 0 | description
# 1 | lspace
# 2 | rspace
# 4 | largeop
# 5 | movablelimits
# 6 | stretchy
# 7 | separator
# 8 | accent
# 9 | fence
# 10 | symmetric
# 13 | direction
# 1) build %moz_hash from $MOZ_DICTIONARY
print "loading $MOZ_DICTIONARY...\n";
open($file, $MOZ_DICTIONARY) || die ("Couldn't open $MOZ_DICTIONARY!");
print "building dictionary...\n";
while (<$file>) {
next unless (m/^operator\.(.*)$/);
(m/^([\w|\.|\\]*)\s=\s(.*)\s#\s(.*)$/);
# 1.1) build the key
$key = $1;
# 1.2) build the array
$_ = $2;
@value = ();
$value[0] = $3;
if (m/^(.*)lspace:(\d)(.*)$/) { $value[1] = $2; } else { $value[1] = "5"; }
if (m/^(.*)rspace:(\d)(.*)$/) { $value[2] = $2; } else { $value[2] = "5"; }
$value[4] = (m/^(.*)largeop(.*)$/);
$value[5] = (m/^(.*)movablelimits(.*)$/);
$value[6] = (m/^(.*)stretchy(.*)$/);
$value[7] = (m/^(.*)separator(.*)$/);
$value[8] = (m/^(.*)accent(.*)$/);
$value[9] = (m/^(.*)fence(.*)$/);
$value[10] = (m/^(.*)symmetric(.*)$/);
if (m/^(.*)direction:([a-z]*)(.*)$/) { $value[13] = $2; }
else { $value[13] = ""; }
# 1.3) save the key and value
$moz_hash{$key} = [ @value ];
}
close($file);
################################################################################
# 2) If mode "check", verify validity of our operator dictionary and quit.
# If mode "compare", go to step 3)
if ($ARGV[0] eq "check") {
print "checking operator dictionary...\n";
open($file_syntax_errors, ">$FILE_SYNTAX_ERRORS") ||
die ("Couldn't open $FILE_SYNTAX_ERRORS!");
$nb_errors = 0;
$nb_warnings = 0;
@moz_keys = (keys %moz_hash);
# check the validity of our private data
while ($key = pop(@moz_keys)) {
if ($key =~ /\\u.+\\u.+\\u.+/) {
$valid = 0;
$nb_errors++;
print $file_syntax_errors "error: \"$key\" has more than 2 characters\n";
}
if ($key =~ /\\u20D2\./ || $key =~ /\\u0338\./) {
$valid = 0;
$nb_errors++;
print $file_syntax_errors "error: \"$key\" ends with character U+20D2 or U+0338\n";
}
@moz = @{ $moz_hash{$key} };
$entry = &generateEntry($key, @moz);
$valid = 1;
if (!(@moz[13] eq "" ||
@moz[13] eq "horizontal" ||
@moz[13] eq "vertical")) {
$valid = 0;
$nb_errors++;
print $file_syntax_errors "error: invalid direction \"$moz[13]\"\n";
}
if (@moz[4] && !(@moz[13] eq "vertical")) {
$valid = 0;
$nb_errors++;
print $file_syntax_errors "error: operator is largeop but does not have vertical direction\n";
}
if (!$valid) {
print $file_syntax_errors $entry;
print $file_syntax_errors "\n";
}
}
# check that all forms have the same direction.
@moz_keys = (keys %moz_hash);
while ($key = pop(@moz_keys)) {
if (@{ $moz_hash{$key} }) {
# the operator has not been removed from the hash table yet.
$_ = $key;
(m/^([\w|\.|\\]*)\.(prefix|infix|postfix)$/);
$key_prefix = "$1.prefix";
$key_infix = "$1.infix";
$key_postfix = "$1.postfix";
@moz_prefix = @{ $moz_hash{$key_prefix} };
@moz_infix = @{ $moz_hash{$key_infix} };
@moz_postfix = @{ $moz_hash{$key_postfix} };
$same_direction = 1;
if (@moz_prefix) {
if (@moz_infix &&
!($moz_infix[13] eq $moz_prefix[13])) {
$same_direction = 0;
}
if (@moz_postfix &&
!($moz_postfix[13] eq $moz_prefix[13])) {
$same_direction = 0;
}
}
if (@moz_infix) {
if (@moz_postfix &&
!($moz_postfix[13] eq $moz_infix[13])) {
$same_direction = 0;
}
}
if (!$same_direction) {
$nb_errors++;
print $file_syntax_errors
"error: operator has a stretchy form, but all forms";
print $file_syntax_errors
" have not the same direction\n";
if (@moz_prefix) {
$_ = &generateEntry($key_prefix, @moz_prefix);
print $file_syntax_errors $_;
}
if (@moz_infix) {
$_ = &generateEntry($key_infix, @moz_infix);
print $file_syntax_errors $_;
}
if (@moz_postfix) {
$_ = &generateEntry($key_postfix, @moz_postfix);
print $file_syntax_errors $_;
}
print $file_syntax_errors "\n";
}
if (@moz_prefix) {
delete $moz_hash{$key.prefix};
}
if (@moz_infix) {
delete $moz_hash{$key_infix};
}
if (@moz_postfix) {
delete $moz_hash{$key_postfix};
}
}
}
close($file_syntax_errors);
print "\n";
if ($nb_errors > 0 || $nb_warnings > 0) {
print "$nb_errors error(s) found\n";
print "$nb_warnings warning(s) found\n";
print "See output file $FILE_SYNTAX_ERRORS.\n\n";
} else {
print "No error found.\n\n";
}
exit 0;
}
################################################################################
# 3) build %wg_hash and @wg_keys from the page $WG_DICTIONARY
print "loading $WG_DICTIONARY...\n";
my $parser = XML::LibXML->new();
my $doc = $parser->parse_file($WG_DICTIONARY);
print "building dictionary...\n";
@wg_keys = ();
foreach my $entry ($doc->findnodes('/root/entry')) {
# 3.1) build the key
$key = "operator.";
$_ = $entry->getAttribute("unicode");
# Skip non-BMP Arabic characters that are handled specially.
if ($_ == "U1EEF0" || $_ == "U1EEF1") {
next;
}
$_ = "$_-";
while (m/^U?0(\w*)-(.*)$/) {
# Concatenate .\uNNNN
$key = "$key\\u$1";
$_ = $2;
}
$_ = $entry->getAttribute("form"); # "Form"
$key = "$key.$_";
# 3.2) build the array
@value = ();
$value[0] = lc($entry->getAttribute("description"));
$value[1] = $entry->getAttribute("lspace");
if ($value[1] eq "") { $value[1] = "5"; }
$value[2] = $entry->getAttribute("rspace");
if ($value[2] eq "") { $value[2] = "5"; }
$_ = $entry->getAttribute("properties");
$value[4] = (m/^(.*)largeop(.*)$/);
$value[5] = (m/^(.*)movablelimits(.*)$/);
$value[6] = (m/^(.*)stretchy(.*)$/);
$value[7] = (m/^(.*)separator(.*)$/);
$value[9] = (m/^(.*)fence(.*)$/);
$value[10] = (m/^(.*)symmetric(.*)$/);
# not stored in the WG dictionary
$value[8] = ""; # accent
$value[13] = ""; # direction
# 3.3) save the key and value
push(@wg_keys, $key);
$wg_hash{$key} = [ @value ];
}
@wg_keys = reverse(@wg_keys);
################################################################################
# 4) Compare the two dictionaries and output the result
print "comparing dictionaries...\n";
open($file_differences, ">$FILE_DIFFERENCES") ||
die ("Couldn't open $FILE_DIFFERENCES!");
open($file_new_dictionary, ">$FILE_NEW_DICTIONARY") ||
die ("Couldn't open $FILE_NEW_DICTIONARY!");
$conflicting = 0; $conflicting_stretching = 0;
$new = 0; $new_stretching = 0;
$obsolete = 0; $obsolete_stretching = 0;
$unchanged = 0;
# 4.1) look to the entries of the WG dictionary
while ($key = pop(@wg_keys)) {
@wg = @{ $wg_hash{$key} };
delete $wg_hash{$key};
$wg_value = &generateCommon(@wg);
if (exists($moz_hash{$key})) {
# entry is in both dictionary
@moz = @{ $moz_hash{$key} };
delete $moz_hash{$key};
$moz_value = &generateCommon(@moz);
if ($moz_value ne $wg_value) {
# conflicting entry
print $file_differences "[conflict]";
$conflicting++;
if ($moz[6] != $wg[6]) {
print $file_differences "[stretching]";
$conflicting_stretching++;
}
print $file_differences " - $key ($wg[0])\n";
print $file_differences "-$moz_value\n+$wg_value\n\n";
$_ = &completeCommon($wg_value, $key, @moz, @wg);
print $file_new_dictionary $_;
} else {
# unchanged entry
$unchanged++;
$_ = &completeCommon($wg_value, $key, @moz, @wg);
print $file_new_dictionary $_;
}
} else {
# we don't have this entry in our dictionary yet
print $file_differences "[new entry]";
$new++;
if ($wg[6]) {
print $file_differences "[stretching]";
$new_stretching++;
}
print $file_differences " - $key ($wg[0])\n";
print $file_differences "-\n+$wg_value\n\n";
$_ = &completeCommon($wg_value, $key, (), @wg);
print $file_new_dictionary $_;
}
}
print $file_new_dictionary
"\n# Entries below are not part of the official MathML dictionary\n\n";
# 4.2) look in our dictionary the remaining entries
@moz_keys = (keys %moz_hash);
@moz_keys = reverse(sort(@moz_keys));
while ($key = pop(@moz_keys)) {
@moz = @{ $moz_hash{$key} };
$moz_value = &generateCommon(@moz);
print $file_differences "[obsolete entry]";
$obsolete++;
if ($moz[6]) {
print $file_differences "[stretching]";
$obsolete_stretching++;
}
print $file_differences " - $key ($moz[0])\n";
print $file_differences "-$moz_value\n+\n\n";
$_ = &completeCommon($moz_value, $key, (), @moz);
print $file_new_dictionary $_;
}
close($file_differences);
close($file_new_dictionary);
print "\n";
print "- $obsolete obsolete entries ";
print "($obsolete_stretching of them are related to stretching)\n";
print "- $unchanged unchanged entries\n";
print "- $conflicting conflicting entries ";
print "($conflicting_stretching of them are related to stretching)\n";
print "- $new new entries ";
print "($new_stretching of them are related to stretching)\n";
print "\nSee output files $FILE_DIFFERENCES and $FILE_NEW_DICTIONARY.\n\n";
print "After having modified the dictionary, please run";
print "./updateOperatorDictionary check\n\n";
exit 0;
################################################################################
sub usage {
# display the accepted command syntax and quit
print "usage:\n";
print " ./updateOperatorDictionary.pl download [unicode.xml]\n";
print " ./updateOperatorDictionary.pl compare [dictionary.xml]\n";
print " ./updateOperatorDictionary.pl check\n";
print " ./updateOperatorDictionary.pl clean\n";
exit 0;
}
sub generateCommon {
# helper function to generate the string of data shared by both dictionaries
my(@v) = @_;
$entry = "lspace:$v[1] rspace:$v[2]";
if ($v[4]) { $entry = "$entry largeop"; }
if ($v[5]) { $entry = "$entry movablelimits"; }
if ($v[6]) { $entry = "$entry stretchy"; }
if ($v[7]) { $entry = "$entry separator"; }
if ($v[9]) { $entry = "$entry fence"; }
if ($v[10]) { $entry = "$entry symmetric"; }
return $entry;
}
sub completeCommon {
# helper to add key and private data to generateCommon
my($entry, $key, @v_moz, @v_wg) = @_;
$entry = "$key = $entry";
if ($v_moz[8]) { $entry = "$entry accent"; }
if ($v_moz[13]) { $entry = "$entry direction:$v_moz[13]"; }
if ($v_moz[0]) {
# keep our previous comment
$entry = "$entry # $v_moz[0]";
} else {
# otherwise use the description given by the WG
$entry = "$entry # $v_wg[0]";
}
$entry = "$entry\n";
return $entry;
}
sub generateEntry {
# helper function to generate an entry of our operator dictionary
my($key, @moz) = @_;
$entry = &generateCommon(@moz);
$entry = &completeCommon($entry, $key, @moz, @moz);
return $entry;
}