--- loncom/homework/cleanxml/post_xml.pm 2016/01/08 20:32:42 1.6 +++ loncom/homework/cleanxml/post_xml.pm 2016/11/10 21:53:56 1.11 @@ -1,7 +1,7 @@ # The LearningOnline Network # Third step to clean a file. # -# $Id: post_xml.pm,v 1.6 2016/01/08 20:32:42 damieng Exp $ +# $Id: post_xml.pm,v 1.11 2016/11/10 21:53:56 damieng Exp $ # # Copyright Michigan State University Board of Trustees # @@ -41,6 +41,8 @@ use Cwd 'abs_path'; use XML::LibXML; use HTML::TokeParser; # used to parse sty files use Tie::IxHash; # for ordered hashes +use tth; +use Apache::html_to_xml; no warnings 'recursion'; # yes, fix_paragraph is using heavy recursion, I know @@ -126,12 +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); @@ -552,11 +558,11 @@ sub replace_m { # 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) $output =~ s/^\s*|\s*$//; $output =~ s/
for multi-line comments without elements. +sub fix_comments { + my ($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','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'); + 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 @@ -1833,6 +1858,18 @@ 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) { @@ -2009,6 +2046,20 @@ 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); @@ -2245,6 +2296,33 @@ 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 {