--- loncom/homework/cleanxml/post_xml.pm 2015/12/03 20:40:31 1.1
+++ loncom/homework/cleanxml/post_xml.pm 2017/01/17 20:29:06 1.12
@@ -1,7 +1,7 @@
# The LearningOnline Network
# Third step to clean a file.
#
-# $Id: post_xml.pm,v 1.1 2015/12/03 20:40:31 damieng Exp $
+# $Id: post_xml.pm,v 1.12 2017/01/17 20:29:06 damieng Exp $
#
# Copyright Michigan State University Board of Trustees
#
@@ -41,18 +41,18 @@ use Cwd 'abs_path';
use XML::LibXML;
use HTML::TokeParser; # used to parse sty files
use Tie::IxHash; # for ordered hashes
-
-use Env qw(RES_DIR); # path of res directory parent (without the / at the end)
+use tth;
+use Apache::html_to_xml;
no warnings 'recursion'; # yes, fix_paragraph is using heavy recursion, I know
# these are constants
-my @block_elements = ('parameter','location','answer','foil','image','polygon','rectangle','text','conceptgroup','itemgroup','item','label','data','function','array','unit','answergroup','functionplotresponse','functionplotruleset','functionplotelements','functionplotcustomrule','essayresponse','hintpart','formulahint','numericalhint','reactionhint','organichint','optionhint','radiobuttonhint','stringhint','customhint','mathhint','formulahintcondition','numericalhintcondition','reactionhintcondition','organichintcondition','optionhintcondition','radiobuttonhintcondition','stringhintcondition','customhintcondition','mathhintcondition','imageresponse','foilgroup','datasubmission','textfield','hiddensubmission','radiobuttonresponse','rankresponse','matchresponse','import','style','script','window','block','library','notsolved','part','postanswerdate','preduedate','problem','problemtype','randomlabel','bgimg','labelgroup','randomlist','solved','while','tex','print','web','gnuplot','curve','Task','IntroParagraph','ClosingParagraph','Question','QuestionText','Setup','Instance','InstanceText','Criteria','CriteriaText','GraderNote','languageblock','translated','lang','instructorcomment','dataresponse','togglebox','standalone','comment','drawimage','allow','displayduedate','displaytitle','responseparam','organicstructure','scriptlib','parserlib','drawoptionlist','spline','backgroundplot','plotobject','plotvector','drawvectorsum','functionplotrule','functionplotvectorrule','functionplotvectorsumrule','axis','key','xtics','ytics','title','xlabel','ylabel','hiddenline','dtm');
-my @inline_like_block = ('stringresponse','optionresponse','numericalresponse','formularesponse','mathresponse','organicresponse','reactionresponse','customresponse','externalresponse', 'hint', 'hintgroup'); # inline elements treated like blocks for pretty print and some other things
+my @block_elements = ('parameter','location','answer','foil','image','polygon','rectangle','text','conceptgroup','itemgroup','item','label','data','function','array','unit','answergroup','functionplotresponse','functionplotruleset','functionplotelements','functionplotcustomrule','essayresponse','hintpart','formulahint','numericalhint','reactionhint','organichint','optionhint','radiobuttonhint','stringhint','customhint','mathhint','formulahintcondition','numericalhintcondition','reactionhintcondition','organichintcondition','optionhintcondition','radiobuttonhintcondition','stringhintcondition','customhintcondition','mathhintcondition','imageresponse','foilgroup','datasubmission','textfield','hiddensubmission','radiobuttonresponse','rankresponse','matchresponse','import','style','script','window','block','library','notsolved','part','postanswerdate','preduedate','problem','problemtype','randomlabel','bgimg','labelgroup','randomlist','solved','while','tex','print','web','gnuplot','curve','Task','IntroParagraph','ClosingParagraph','Question','QuestionText','Setup','Instance','InstanceText','Criteria','CriteriaText','GraderNote','languageblock','instructorcomment','dataresponse','togglebox','standalone','comment','drawimage','allow','displayduedate','displaytitle','responseparam','organicstructure','scriptlib','parserlib','drawoptionlist','spline','backgroundplot','plotobject','plotvector','drawvectorsum','functionplotrule','functionplotvectorrule','functionplotvectorsumrule','axis','key','xtics','ytics','title','xlabel','ylabel','hiddenline','dtm');
+my @inline_like_block = ('stringresponse','optionresponse','numericalresponse','formularesponse','mathresponse','organicresponse','reactionresponse','customresponse','externalresponse', 'hint', 'hintgroup','translated','lang'); # inline elements treated like blocks for pretty print and some other things
my @responses = ('stringresponse','optionresponse','numericalresponse','formularesponse','mathresponse','organicresponse','reactionresponse','customresponse','externalresponse','essayresponse','radiobuttonresponse','matchresponse','rankresponse','imageresponse','functionplotresponse');
my @block_html = ('html','head','body','section','h1','h2','h3','h4','h5','h6','div','p','ul','ol','li','table','tbody','tr','td','th','dl','dt','dd','pre','noscript','hr','address','blockquote','object','applet','embed','map','form','fieldset','iframe','center','frameset');
my @no_newline_inside = ('import','parserlib','scriptlib','data','function','label','xlabel','ylabel','tic','text','rectangle','image','title','h1','h2','h3','h4','h5','h6','li','td','p');
-my @preserve_elements = ('script','answer','pre');
+my @preserve_elements = ('script','answer','pre','style');
my @accepting_style = ('section','h1','h2','h3','h4','h5','h6','div','p','li','td','th','dt','dd','pre','blockquote');
my @latex_math = ('\alpha', '\theta', '\omicron', '\tau', '\beta', '\vartheta', '\pi', '\upsilon', '\gamma', '\gamma', '\varpi', '\phi', '\delta', '\kappa', '\rho', '\varphi', '\epsilon', '\lambda', '\varrho', '\chi', '\varepsilon', '\mu', '\sigma', '\psi', '\zeta', '\nu', '\varsigma', '\omega', '\eta', '\xi',
'\Gamma', '\Lambda', '\Sigma', '\Psi', '\Delta', '\Xi', '\Upsilon', '\Omega', '\Theta', '\Pi', '\Phi',
@@ -77,9 +77,13 @@ my $warnings; # 1 = print warnings
# Parses the XML document and fixes many things to turn it into a document matching the schema.
-# Returns the text of the document as a byte string.
+# @param {reference} textref - reference to the text of the document
+# @param {string} file_path - path of the file being processed (we only extract the directory path from it)
+# @param {string} res_dir - path of res directory parent (without the / at the end)
+# @param {boolean} warn - 1 to print warnings, 0 otherwise
+# @returns the text of the document as a byte string.
sub post_xml {
- my ($textref, $file_path, $warn) = @_;
+ my ($textref, $file_path, $res_dir, $warn) = @_;
$warnings = $warn;
my $dom_doc = XML::LibXML->load_xml(string => $textref);
@@ -92,10 +96,10 @@ sub post_xml {
fix_attribute_case($root);
- my $fix_by_hand = replace_m($root);
+ replace_m($root);
my @all_block = (@block_elements, @block_html);
- add_sty_blocks($file_path, $root, \@all_block); # must come before the subs using @all_block
+ add_sty_blocks($file_path, $res_dir, $root, \@all_block); # must come before the subs using @all_block
fix_block_styles($root, \@all_block);
$root->normalize();
@@ -124,12 +128,16 @@ sub post_xml {
remove_useless_notsolved($root);
+ fix_comments($root);
+
fix_paragraphs_inside($root, \@all_block);
remove_empty_style($root);
fix_empty_lc_elements($root);
+ reduce_empty_p($root);
+
lowercase_attribute_values($root);
replace_numericalresponse_unit_attribute($root);
@@ -140,9 +148,6 @@ sub post_xml {
replace_tm_dtm($root);
- if ($fix_by_hand) {
- die "The file has been converted but it should be fixed by hand.";
- }
return $dom_doc->toString(); # byte string !
}
@@ -151,7 +156,7 @@ sub fix_structure {
# the root element has already been added in pre_xml
my $root = $doc->documentElement;
# inside the root, replace html, problem and library elements by their content
- my @toreplace = ('html','problem','library');
+ my @toreplace = ('html','problem','library','Task');
foreach my $name (@toreplace) {
my @elements = $root->getElementsByTagName($name);
foreach my $element (@elements) {
@@ -353,11 +358,9 @@ sub fix_attribute_case {
# Replaces m by HTML, tm and/or dtm (which will be replaced by <m> later, but they are useful
# to know if the element is a block element or not).
# m might contain non-math LaTeX, while tm and dtm may only contain math.
-# Returns 1 if the file should be fixed by hand, 0 otherwise.
sub replace_m {
my ($root) = @_;
my $doc = $root->ownerDocument;
- my $fix_by_hand = 0;
# search for variable declarations
my @variables = ();
my @scripts = $root->getElementsByTagName('script');
@@ -406,7 +409,6 @@ sub replace_m {
if ($warnings) {
print "WARNING: <m> is used in a script, it should be converted by hand\n";
}
- $fix_by_hand = 1;
}
}
}
@@ -421,7 +423,6 @@ sub replace_m {
if ($warnings) {
print "WARNING: m value is not simple text\n";
}
- $fix_by_hand = 1;
next;
}
my $text = $m->firstChild->nodeValue;
@@ -429,6 +430,15 @@ sub replace_m {
my $var_key1 = 'dfhg3df54hg65hg4';
my $var_key2 = 'dfhg654d6f5g4h5f';
my $eval = defined $m->getAttribute('eval') && $m->getAttribute('eval') eq 'on';
+ my $display = $m->getAttribute('display');
+ if (defined $display) {
+ if ($display eq '') {
+ $display = undef;
+ }
+ if (lc($display) eq 'jsmath') {
+ $display = 'mathjax';
+ }
+ }
if ($eval) {
# replace variables
foreach my $variable (@variables) {
@@ -466,6 +476,9 @@ sub replace_m {
if ($eval) {
$new_node->setAttribute('eval', 'on');
}
+ if (defined $display) {
+ $new_node->setAttribute('display', $display);
+ }
$new_node->appendChild($doc->createTextNode($new_text));
$m->parentNode->replaceChild($new_node, $m);
next;
@@ -496,7 +509,7 @@ sub replace_m {
# there are math separators inside, even after hiding variables, or there was no math symbol
# hide math parts inside before running tth
- my $math_key1 = '#ghjgdh5hg45gf';
+ my $math_key1 = '#5752398247516385';
my $math_key2 = '#';
my @maths = ();
my @separators = (['$$','$$'], ['\\(','\\)'], ['\\[','\\]'], ['$','$']);
@@ -527,14 +540,30 @@ sub replace_m {
$math =~ s/&/&/g;
$math =~ s/</</g;
$math =~ s/>/>/g;
+ my ($mel, $inside);
if ($math =~ /^\$\$(.*)\$\$$/s) {
- $math = '<dtm>'.$1.'</dtm>';
+ $mel = 'dtm';
+ $inside = $1;
} elsif ($math =~ /^\\\[(.*)\\\]$/s) {
- $math = '<dtm>'.$1.'</dtm>';
+ $mel = 'dtm';
+ $inside = $1;
} elsif ($math =~ /^\\\((.*)\\\)$/s) {
- $math = '<tm>'.$1.'</tm>';
+ $mel = 'tm';
+ $inside = $1;
} elsif ($math =~ /^\$(.*)\$$/s) {
- $math = '<tm>'.$1.'</tm>';
+ $mel = 'tm';
+ $inside = $1;
+ }
+ if (defined $inside) {
+ if ($inside =~ /^\s*$/) {
+ $math = '';
+ } else {
+ $math = '<'.$mel;
+ if ($eval && $inside =~ /$var_key1/) {
+ $math .= ' eval="on"';
+ }
+ $math .= '>'.$inside.'</'.$mel.'>';
+ }
}
my $replace = $math_key1.($i+1).$math_key2;
$html_text =~ s/$replace/$math/;
@@ -552,17 +581,16 @@ sub replace_m {
$m->parentNode->replaceChild($fragment, $m);
}
- return $fix_by_hand;
}
# Returns the HTML equivalent of LaTeX input, using tth
sub tth {
my ($text) = @_;
- my ($fh, $tmp_path) = tempfile();
- binmode($fh, ':utf8');
- print $fh $text;
- close $fh;
- my $output = `tth -r -w2 -u -y0 < $tmp_path 2>/dev/null`;
+ my $output = &tth::tth($text);
+ my $errorstring = &tth::ttherror();
+ if ($errorstring) {
+ die $errorstring;
+ }
# hopefully the temp file will not be removed before this point (otherwise we should use unlink_on_destroy 0)
$output =~ s/^\s*|\s*$//;
$output =~ s/<div class="p"><!----><\/div>/<br\/>/; # why is tth using such ugly markup for \newline ?
@@ -573,7 +601,7 @@ sub tth {
sub html_to_dom {
my ($text) = @_;
$text = '<root>'.$text.'</root>';
- my $textref = html_to_xml::html_to_xml(\$text);
+ my $textref = Apache::html_to_xml::html_to_xml(\$text);
utf8::upgrade($$textref); # otherwise the XML parser fails when the HTML parser turns into a character
my $dom_doc = XML::LibXML->load_xml(string => $textref);
my $root = $dom_doc->documentElement;
@@ -591,8 +619,10 @@ sub html_to_dom {
# Use the linked sty files to guess which newly defined elements should be considered blocks.
# Also adds to @containing_styles the sty elements that contain styles.
# @param {string} fn - the file path (we only extract the directory path from it)
+# @param {string} res_dir - path of res directory parent (without the / at the end)
+# @param {Element} root - the root element
sub add_sty_blocks {
- my ($fn, $root, $all_block) = @_;
+ my ($fn, $res_dir, $root, $all_block) = @_;
my $doc = $root->ownerDocument;
my @parserlibs = $doc->getElementsByTagName('parserlib');
my @libs = ();
@@ -608,7 +638,7 @@ sub add_sty_blocks {
my ($name, $path, $suffix) = fileparse($fn);
foreach my $sty (@libs) {
if (substr($sty, 0, 1) eq '/') {
- $sty = $RES_DIR.$sty;
+ $sty = $res_dir.$sty;
} else {
$sty = $path.$sty;
}
@@ -1812,12 +1842,31 @@ sub remove_useless_notsolved {
}
}
+# Use <pre> for multi-line comments without elements.
+sub fix_comments {
+ my ($root) = @_;
+ my $doc = $root->ownerDocument;
+ my @comments = $root->getElementsByTagName('comment');
+ foreach my $comment (@comments) {
+ my $first = $comment->firstChild;
+ if (defined $first) {
+ if ($first->nodeType == XML_TEXT_NODE && $first->nodeValue =~ /\n/ &&
+ !defined $first->nextSibling) {
+ my $pre = $doc->createElement('pre');
+ $comment->removeChild($first);
+ $comment->appendChild($pre);
+ $pre->appendChild($first);
+ }
+ }
+ }
+}
+
# adds a paragraph inside if needed and calls fix_paragraph for all paragraphs (including new ones)
sub fix_paragraphs_inside {
my ($node, $all_block) = @_;
# blocks in which paragrahs will be added:
- my @blocks_with_p = ('loncapa','library','problem','part','problemtype','window','block','while','postanswerdate','preduedate','solved','notsolved','languageblock','translated','lang','instructorcomment','togglebox','standalone','form');
- my @fix_p_if_br_or_p = (@responses,'foil','item','text','label','hintgroup','hintpart','hint','web','windowlink','div','li','dd','td','th','blockquote');
+ my @blocks_with_p = ('loncapa','library','problem','part','problemtype','window','block','while','postanswerdate','preduedate','languageblock','instructorcomment','togglebox','standalone','body','form');
+ my @fix_p_if_br_or_p = (@responses,'foil','item','text','label','hintgroup','hintpart','hint','web','windowlink','div','li','dd','td','th','blockquote','solved','notsolved');
if ((string_in_array(\@blocks_with_p, $node->nodeName) && paragraph_needed($node)) ||
(string_in_array(\@fix_p_if_br_or_p, $node->nodeName) && paragraph_inside($node))) {
# if non-empty, add paragraphs where needed between all br and remove br
@@ -1837,6 +1886,18 @@ sub fix_paragraphs_inside {
push(@new_children, $doc->createElement('p'));
}
$p = undef;
+ # ignore the next node if it is a br (the paragraph default margin will take as much space)
+ # (ignoring whitespace)
+ while (defined $next && $next->nodeType == XML_TEXT_NODE && $next->nodeValue =~ /^[ \t\f\n\r]*$/) {
+ my $next2 = $next->nextSibling;
+ $node->removeChild($next);
+ $next = $next2;
+ }
+ if (defined $next && $next->nodeType == XML_ELEMENT_NODE && $next->nodeName eq 'br') {
+ my $next2 = $next->nextSibling;
+ $node->removeChild($next);
+ $next = $next2;
+ }
} elsif ($child->nodeType == XML_ELEMENT_NODE && string_in_array(\@inline_like_block, $child->nodeName)) {
# inline_like_block: use the paragraph if there is one, otherwise do not create one
if (defined $p) {
@@ -2013,6 +2074,20 @@ sub fix_paragraph {
if (!defined $left || !$left_needs_p) {
$replacement->appendChild($middle);
}
+ # ignore the next node if it is a br (the paragraph default margin will take as much space)
+ my $first_right;
+ if (defined $right) {
+ $first_right = $right->firstChild;
+ # ignore non-nbsp whitespace
+ while (defined $first_right && $first_right->nodeType == XML_TEXT_NODE &&
+ $first_right->nodeValue =~ /^[ \t\f\n\r]*$/) {
+ $first_right = $first_right->nextSibling;
+ }
+ }
+ if (defined $first_right && $first_right->nodeType == XML_ELEMENT_NODE &&
+ $first_right->nodeName eq 'br') {
+ $right->removeChild($first_right);
+ }
} else {
fix_paragraphs_inside($n, $all_block);
$replacement->appendChild($n);
@@ -2250,6 +2325,33 @@ sub fix_empty_lc_elements {
}
}
+# remove consecutive empty paragraphs (they will not show anyway)
+sub reduce_empty_p {
+ my ($node) = @_;
+ my $next;
+ for (my $child=$node->firstChild; defined $child; $child=$next) {
+ $next = $child->nextSibling;
+ while (defined $next && $next->nodeType == XML_TEXT_NODE && $next->nodeValue =~ /^[ \t\f\n\r]*$/) {
+ $next = $next->nextSibling;
+ }
+ if ($child->nodeType == XML_ELEMENT_NODE && $child->nodeName eq 'p' && defined $next &&
+ $next->nodeType == XML_ELEMENT_NODE && $next->nodeName eq 'p') {
+ my $first = $child->firstChild;
+ if (!defined $first || (!defined $first->nextSibling &&
+ $first->nodeType == XML_TEXT_NODE && $first->nodeValue =~ /^[ \t\f\n\r]*$/)) {
+ $first = $next->firstChild;
+ if (!defined $first || (!defined $first->nextSibling &&
+ $first->nodeType == XML_TEXT_NODE && $first->nodeValue =~ /^[ \t\f\n\r]*$/)) {
+ $node->removeChild($child);
+ }
+ }
+ }
+ if ($child->nodeType == XML_ELEMENT_NODE) {
+ reduce_empty_p($child);
+ }
+ }
+}
+
# turn some attribute values into lowercase when they should be
sub lowercase_attribute_values {
my ($root) = @_;
@@ -2398,8 +2500,29 @@ sub pretty {
my $type = $node->nodeType;
if ($type == XML_ELEMENT_NODE) {
my $name = $node->nodeName;
- if ((string_in_array($all_block, $name) || string_in_array(\@inline_like_block, $name)) &&
- !string_in_array(\@preserve_elements, $name)) {
+ if (string_in_array(\@preserve_elements, $name)) {
+ # remove newlines at the beginning and the end of preserve elements
+ if (defined $node->firstChild && ($node->firstChild->nodeType == XML_TEXT_NODE ||
+ $node->firstChild->nodeType == XML_CDATA_SECTION_NODE)) {
+ my $text = $node->firstChild->nodeValue;
+ $text =~ s/^\n+//;
+ if ($text eq '') {
+ $node->removeChild($node->firstChild);
+ } else {
+ $node->firstChild->setData($text);
+ }
+ }
+ if (defined $node->lastChild && ($node->lastChild->nodeType == XML_TEXT_NODE ||
+ $node->lastChild->nodeType == XML_CDATA_SECTION_NODE)) {
+ my $text = $node->lastChild->nodeValue;
+ $text =~ s/\n+$//;
+ if ($text eq '') {
+ $node->removeChild($node->lastChild);
+ } else {
+ $node->lastChild->setData($text);
+ }
+ }
+ } elsif (string_in_array($all_block, $name) || string_in_array(\@inline_like_block, $name)) {
# make sure there is a newline at the beginning and at the end if there is anything inside
if (defined $node->firstChild && !string_in_array(\@no_newline_inside, $name)) {
my $first = $node->firstChild;
@@ -2482,26 +2605,6 @@ sub pretty {
if ($text eq '') {
$node->removeChild($node->lastChild);
} else {
- $node->lastChild->setData($text);
- }
- }
- } elsif (string_in_array(\@preserve_elements, $name)) {
- # collapse newlines at the beginning and the end of scripts
- if (defined $node->firstChild && $node->firstChild->nodeType == XML_TEXT_NODE) {
- my $text = $node->firstChild->nodeValue;
- $text =~ s/^\n( *\n)+/\n/;
- if ($text eq '') {
- $node->removeChild($node->firstChild);
- } else {
- $node->firstChild->setData($text);
- }
- }
- if (defined $node->lastChild && $node->lastChild->nodeType == XML_TEXT_NODE) {
- my $text = $node->lastChild->nodeValue;
- $text =~ s/\n( *\n)+$/\n/;
- if ($text eq '') {
- $node->removeChild($node->lastChild);
- } else {
$node->lastChild->setData($text);
}
}