1: # The LearningOnline Network with CAPA
2: # TeX Conversion Module
3: #
4: # $Id: lontexconvert.pm,v 1.112.2.7 2018/04/23 19:33:17 raeburn Exp $
5: #
6: # Copyright Michigan State University Board of Trustees
7: #
8: # This file is part of the LearningOnline Network with CAPA (LON-CAPA).
9: #
10: # LON-CAPA is free software; you can redistribute it and/or modify
11: # it under the terms of the GNU General Public License as published by
12: # the Free Software Foundation; either version 2 of the License, or
13: # (at your option) any later version.
14: #
15: # LON-CAPA is distributed in the hope that it will be useful,
16: # but WITHOUT ANY WARRANTY; without even the implied warranty of
17: # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18: # GNU General Public License for more details.
19: #
20: # You should have received a copy of the GNU General Public License
21: # along with LON-CAPA; if not, write to the Free Software
22: # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
23: #
24: # /home/httpd/html/adm/gpl.txt
25: #
26: # http://www.lon-capa.org/
27: #
28: # Copyright for TtHfunc and TtMfunc by Ian Hutchinson.
29: # TtHfunc and TtMfunc (the "Code") may be compiled and linked into
30: # binary executable programs or libraries distributed by the
31: # Michigan State University (the "Licensee"), but any binaries so
32: # distributed are hereby licensed only for use in the context
33: # of a program or computational system for which the Licensee is the
34: # primary author or distributor, and which performs substantial
35: # additional tasks beyond the translation of (La)TeX into HTML.
36: # The C source of the Code may not be distributed by the Licensee
37: # to any other parties under any circumstances.
38: #
39:
40: package Apache::lontexconvert;
41:
42: use strict;
43: use tth();
44: use vars qw($errorstring);
45: #use Apache::lonxml();
46: use Apache::lonlocal;
47: use Apache::lonnet;
48: use lib '/home/httpd/lib/perl/';
49: use LONCAPA;
50: use URI::Escape;
51: use IO::Socket::INET;
52:
53:
54: #
55: # Table of substitutions to unicode characters.
56: #
57:
58: my %unicode_harpoons = (
59: '\rightleftharpoons' => 0x21cc,
60: );
61:
62: my %unicode_translations = (
63:
64: # Brackets - unicode for browsers/OS which support it.
65:
66: '' => 0x23a1,
67: '' => 0x23a2,
68: '' => 0x23a3,
69: '' => 0x23a4,
70: '' => 0x23a5,
71: '' => 0x23a6,
72:
73: # Parens - unicode for browsers/OS which support it
74:
75: '' => 0x239b,
76: '' => 0x239c,
77: '' => 0x239d,
78: '' => 0x239e,
79: '' => 0x239f,
80: '' => 0x23a0,
81:
82: );
83:
84: my %ascii_8bit_translations = (
85:
86: # Brackets - pure 8-bit ascii ugliness for browsers/OS which can't handle unicode
87:
88: '' => 0x5b,
89: '' => 0x5b, # '['
90: '' => 0x5b,
91: '' => 0x5d, # ']'
92: '' => 0x5d,
93: '' => 0x5d,
94:
95: # Parens - pure 8-bit ascii ugliness for browsers/OS which can't handle unicode
96:
97: '' => 0x28,
98: '' => 0x28, # '('
99: '' => 0x28,
100: '' => 0x29,
101: '' => 0x29, # '('
102: '' => 0x29,
103:
104: );
105:
106: ##
107: # Utility to convert elements of a string to unicode:
108: #
109: # @param input - Input string
110: # @param pattern - Pattern to convert
111: # @param unicode - Unicode to substitute for pattern.
112: #
113: # @return string - resulting string.
114: #
115: sub unicode_subst {
116: my ($input, $pattern, $unicode) = @_;
117:
118: my $char = pack('U', $unicode);
119:
120: $input =~ s/$pattern/$char/g;
121:
122: return $input;
123: }
124:
125: # ====================================================================== Header
126:
127: sub init_tth {
128: my $options=$env{'course.'.$env{'request.course.id'}.'.tthoptions'};
129: if ($options =~ /\S/) {
130: $options = ' '.$options;
131: } else {
132: undef($options);
133: }
134: if ($env{'browser.mathml'}) {
135: &tth::ttminit();
136: if ($env{'browser.unicode'}) {
137: &tth::ttmoptions('-L -u1'.$options);
138: } else {
139: &tth::ttmoptions('-L -u0'.$options);
140: }
141: } else {
142: &tth::tthinit();
143: if ($env{'browser.unicode'}) {
144: &tth::tthoptions('-L -u1'.$options);
145: } else {
146: &tth::tthoptions('-L -u0'.$options);
147: }
148: }
149: }
150:
151: # ================================================================== Conversion
152:
153: $Apache::lontexconvert::messedup=0;
154:
155:
156: sub convert_real {
157: my ($texstring)=@_;
158: my ($xmlstring,$errorstring);
159: local $SIG{SEGV}=sub { $Apache::lontexconvert::messedup=1; die; };
160: local $SIG{ALRM}=sub {
161: &Apache::lonnet::logthis("ALRM");
162: $xmlstring='['.&mt("TeX unconverted due to errors").']';
163: $Apache::lontexconvert::messedup=1;
164: die &mt("TeX unconverted due to errors");
165: };
166: &Apache::lonxml::start_alarm();
167: if ($env{'browser.mathml'}) {
168: $xmlstring=&tth::ttm($$texstring);
169: $xmlstring=~s/\<math\>/\<math xmlns=\"\&mathns\;\"\>/g;
170: $xmlstring=~s/\<br\>/\<br\/\>/g;
171: $xmlstring=~s/\<p\>/\<p\>\<\/p\>/g;
172: $errorstring.=&tth::ttmerror();
173: } else {
174: $xmlstring=&tth::tth($$texstring);
175: $errorstring.=&tth::ttherror();
176: $xmlstring=~s-</font(\s*)>-</font>-g;
177: }
178: $xmlstring=~s/^\s*\<br clear\=\"all\"/\<br/s;
179: $xmlstring=~s/^\s*//;
180: $xmlstring=~s/\s*$//;
181: &Apache::lonxml::end_alarm();
182:
183: #
184: # Several strings produced by tth require
185: # transliteration -> unicode equivalents to render reliably
186: # in browsers. %unicode_translations and %unicode_harpoons are tables of
187: # string->substitution which we now apply. (%ascii_8bit_translations used
188: # instead for Windows XP and mobile devices.
189:
190: my $use_ascii;
191: if ($env{'browser.os'} eq 'win') {
192: if (($env{'browser.osversion'}) && ($env{'browser.osversion'} < 6.0)) {
193: $use_ascii = 1;
194: }
195: }
196: if ($env{'browser.mobile'}) {
197: $use_ascii = 1;
198: }
199:
200: foreach my $pattern (keys(%unicode_translations)) {
201: my $unicode = $unicode_translations{$pattern};
202: if ($use_ascii) {
203: $unicode = $ascii_8bit_translations{$pattern};
204: }
205: $xmlstring = &unicode_subst($xmlstring, $pattern, $unicode);
206: }
207:
208: foreach my $pattern (keys(%unicode_harpoons)) {
209: $xmlstring = &unicode_subst($xmlstring, $pattern, $unicode_harpoons{$pattern});
210: }
211:
212: return ($xmlstring,$errorstring);
213: }
214:
215: sub tth_converted {
216: my $texstring=shift;
217: my $xmlstring='['.&mt('UNDISPLAYABLE').']';
218: if ($Apache::lontexconvert::messedup) {
219: return '['.&mt('TeX unconverted due to previous errors').']';
220: }
221: $$texstring ='\\documentstyle{article}'.$$texstring;
222:
223: eval(<<'ENDCONV');
224: ($xmlstring,$errorstring)=&convert_real($texstring)
225: ENDCONV
226: if ($@) {
227: $errorstring.=&mt("Evaluation Error: ").$@;
228: $Apache::lontexconvert::messedup=1;
229: }
230: if ($Apache::lontexconvert::messedup || &tth::tthmessedup() ||
231: $errorstring) {
232: &Apache::lonnet::logthis("Trying to kill myself");
233: $Apache::lontexconvert::messedup=1;
234: if (ref($Apache::lonxml::request)) {
235: $Apache::lonxml::request->child_terminate();
236: } else {
237: my $request;
238: eval { $request=Apache->request; };
239: if (!$request) {
240: eval { $request=Apache2::RequestUtil->request; };
241: }
242: if ($request) {
243: $request->child_terminate();
244: } else {
245: &Apache::lonnet::logthis("Unable to find a request to do child_terminate on");
246: }
247: }
248: }
249: return $xmlstring;
250: }
251:
252: sub clean_out_math_mode {
253: my ($texstring)=@_;
254: $$texstring=~s/(?<!\\)\$//g;
255: $$texstring=~s/\\[\)\(\]\[]//g;
256: $$texstring=~s/\\ensuremath//g;
257: return '';
258: }
259:
260: sub displaystyle {
261: my ($texstring)=@_;
262: #has a $$ or \[ or \displaystyle or eqnarray in it, guessinng it's display mode
263: if ($$texstring=~/[^\\]\$\$/ ||
264: $$texstring=~/\\\[/ ||
265: $$texstring=~/\\displaystyle/ ||
266: $$texstring=~/eqnarray/
267: ) { return 1; }
268: return 0;
269: }
270:
271: sub MathJax_converted {
272: my $texstring=shift;
273: my ($tag,$startspan,$endspan);
274: $tag='math/tex;';
275: if (&displaystyle($texstring)) {
276: $tag='math/tex; mode=display';
277: $startspan='';
278: $endspan='';
279: } else {
280: $startspan='<span style="display:inline-block;">';
281: $endspan='</span>';
282: }
283: &clean_out_math_mode($texstring);
284: return &MathJax_header().$startspan.
285: '<script type="'.$tag.'">'.$$texstring.'</script>'.$endspan;
286: }
287:
288: {
289: #Relies heavily on the previous jsMath installation
290: my @MathJax_sent_header;
291: sub MathJax_reset {
292: undef(@MathJax_sent_header);
293: }
294: sub MathJax_push {
295: push(@MathJax_sent_header,0);
296: }
297: sub MathJax_header {
298: if (!@MathJax_sent_header) {
299: &Apache::lonnet::logthis("mismatched calls of MathJax_header and MathJax_process");
300: return '';
301: }
302: return '' if $MathJax_sent_header[-1];
303: $MathJax_sent_header[-1]=1;
304: return
305: '<script type="text/javascript" src="/adm/MathJax/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>'."\n";
306: }
307: }
308:
309: sub tex_engine {
310: if (exists($env{'form.texengine'})) {
311: if ($env{'form.texengine'} ne '') {
312: if (lc($env{'form.texengine'}) eq 'jsmath') {
313: return 'MathJax';
314: }
315: return $env{'form.texengine'};
316: }
317: }
318: if ($env{'request.course.id'}
319: && exists($env{'course.'.$env{'request.course.id'}.'.texengine'})) {
320: if (lc($env{'course.'.$env{'request.course.id'}.'.texengine'}) eq 'jsmath') {
321: return 'MathJax';
322: }
323: return $env{'course.'.$env{'request.course.id'}.'.texengine'};
324: }
325: if (exists($env{'environment.texengine'})) {
326: if (lc($env{'environment.texengine'}) eq 'jsmath') {
327: return 'MathJax';
328: }
329: return $env{'environment.texengine'};
330: }
331: return 'tth';
332: }
333:
334: sub init_math_support {
335: &init_tth();
336: &Apache::lontexconvert::MathJax_push();
337: if (lc(&tex_engine()) eq 'mathjax') {
338: return &Apache::lontexconvert::MathJax_header();
339: }
340: return;
341: }
342:
343: sub mimetex_valign {
344: my ($esc_texstring)=@_;
345: my $valign = 0;
346: my $path = '/cgi-bin/mimetex.cgi?'.$esc_texstring;
347: my $socket;
348: &Apache::lonxml::start_alarm();
349: $socket = IO::Socket::INET->new(PeerAddr => 'localhost',
350: PeerPort => 'http(80)',
351: Proto => 'tcp');
352: if ($socket) {
353: my $headreq = "HEAD $path HTTP/1.0\r\n\r\n";
354: print $socket $headreq;
355: while (<$socket>) {
356: if (/Vertical\-Align\:\s*?([\-\d]+)/) {
357: $valign = $1;
358: }
359: }
360: $socket->close();
361: }
362: &Apache::lonxml::end_alarm();
363: return $valign;
364: }
365:
366: sub mimetex_converted {
367: my $texstring=shift;
368:
369: # Alt-Argument for screen readers
370: my $alt_string=$$texstring;
371: $alt_string=~s/\"/\'\'/g;
372:
373: # Is this displaystyle?
374:
375: my $displaystyle=&displaystyle($texstring);
376:
377: # Remove math environment delimiters
378:
379: &clean_out_math_mode($texstring);
380:
381: if ($displaystyle) {
382: $$texstring='\\displaystyle \\Large '.$$texstring;
383: }
384: my $esc_texstring = &uri_escape($$texstring);
385: my $valign = &mimetex_valign($esc_texstring);
386: my $result='<img src="/cgi-bin/mimetex.cgi?'.$esc_texstring.'" style="vertical-align:'.$valign.'px" alt="'.$alt_string.'" />';
387: if ($displaystyle) {
388: $result='<div style="text-align:center">'.$result.'</div>';
389: }
390: return $result;
391: }
392:
393: sub converted {
394: my ($string,$mode)=@_;
395: if ($mode eq '') { $mode = &tex_engine(); }
396: if ($mode =~ /tth/i) {
397: return &tth_converted($string);
398: } elsif ($mode =~ /jsmath/i) {
399: return &MathJax_converted($string);
400: } elsif ($mode =~ /mathjax/i) {
401: return &MathJax_converted($string);
402: } elsif ($mode =~ /mimetex/i) {
403: return &mimetex_converted($string);
404: } elsif ($mode =~ /raw/i) {
405: return $$string;
406: }
407: return &tth_converted($string);
408: }
409:
410: # ------------------------------------------------------------ Message display
411:
412: sub to_convert {
413: my ($string) = @_;
414: &init_tth();
415: $string=~s/\<br\s*\/?\>/ /gs;
416: # $string=~s/\s/ /gs;
417: $string=&HTML::Entities::decode($string);
418: return &converted(\$string);
419: }
420:
421: sub smiley {
422: my $expression=shift;
423: my %smileys=(
424: '\:\-*\)' => 'face-smile.png',
425: '8\-\)' => 'face-cool.png',
426: '8\-(I|\|)' => 'face-glasses.png',
427: '\:\-(I|\|)' => 'face-plain.png',
428: '\:\-(o|O|\(\))' => 'face-surprise.png',
429: ':\-\(' => 'face-sad.png',
430: '\;\-\)' => 'face-wink.png',
431: '\:\-(P|p)' => 'face-raspberry.png',
432: '\:\-(\\\|\\/)' => 'face-uncertain.png',
433: '\:\-D' => 'face-smile-big.png',
434: '\:\-(C|\@)' => 'face-angry.png',
435: '\:(\'|\`)\-*\(' => 'face-crying.png',
436: '\:\-(X|x|\#)' => 'face-quiet.png',
437: '\:\-(s|S)' => 'face-uncertain.png',
438: '\:\-\$' => 'face-embarrassed.png',
439: '\:\-\*' => 'face-kiss.png',
440: '\+O\(' => 'face-sick.png',
441: '(\<\;3|\(heart\))' => 'heart.png',
442: '\(rose\)' => 'rose.png',
443: '\(pizza\)' => 'food-pizza.png',
444: '\(cake\)' => 'food-cake.png',
445: '\(ninja\)' => 'face-ninja.png',
446: '\(pirate\)' => 'face-pirate.png',
447: '\((agree|yes)\)' => 'opinion-agree.png',
448: '\((disagree|nay)\)' => 'opinion-disagree.png',
449: '(o|O)\-\)' => 'face-angel.png',
450: );
451: my $iconpath=$Apache::lonnet::perlvar{'lonIconsURL'};
452: foreach my $smiley (keys(%smileys)) {
453: $expression=~s/$smiley/\<img src="$iconpath\/$smileys{$smiley}" \/\>/gs;
454: }
455: return $expression;
456: }
457:
458: sub msgtexconverted {
459: my ($message,$email) = @_;
460: $errorstring='';
461: my $outmessage='';
462: my $tex=0;
463: foreach my $fragment (split(/(?:\<\;|\<)\/*m\s*(?:\>\;|\>)/i,$message)) {
464: if ($tex) {
465: if ($email) {
466: $outmessage.='</pre><tt>'.&to_convert($fragment).'</tt><pre>';
467: $tex=0;
468: } else {
469: $outmessage.=&to_convert($fragment);
470: $tex=0;
471: }
472: } else {
473: $outmessage.=&smiley($fragment);
474: $tex=1;
475: }
476: }
477: $message=$outmessage; $outmessage=''; $tex=0;
478: foreach my $fragment (split(/(?:\<\;|\<)\/*algebra\s*(?:\>\;|\>)/i,
479: $message)) {
480: if ($tex) {
481: my $algebra = &algebra($fragment, 'web', undef, undef, undef, 'tth');
482: if ($email) {
483: $outmessage.='</pre><tt>'.$algebra.'</tt><pre>';
484: $tex=0;
485: } else {
486: $outmessage.=$algebra;
487: $tex=0;
488: }
489: } else {
490: $outmessage.=$fragment;
491: $tex=1;
492: }
493: }
494: if (wantarray) {
495: return ($outmessage,$errorstring);
496: } else {
497: return $outmessage.$errorstring;
498: }
499: }
500:
501: sub algebra {
502: use AlgParser;
503: my ($string,$target,$style,$parstack,$safeeval,$tth)=@_;
504: my $parser = new AlgParserWithImplicitExpand;
505: if ($tth eq 'tth') {&init_tth();}
506: $string=&prepare_algebra($string);
507: my $ret = $parser->parse($string);
508: my $result='['.&mt('Algebra unconverted due to previous errors').']';
509: if ( ref($ret) ) {
510: #$parser->tostring();
511: $parser->normalize();
512: my $latex=$parser->tolatex();
513: $latex=&postprocess_algebra($latex);
514: if ($style eq 'display') {
515: $latex='$$'.$latex.'$$x';
516: } else {
517: # style is 'inline'
518: $latex='\\ensuremath{'.$latex.'}';
519: }
520: if ($target eq 'web' || $target eq 'analyze') {
521: my $display=&Apache::lonxml::get_param('display',$parstack,$safeeval);
522: $result = &converted(\$latex,$display);
523: # $result = &converted(\$latex);
524: } else {
525: $result = $latex;
526: }
527: } else {
528: &Apache::lonxml::error($parser->{'htmlerror'});
529: }
530: }
531:
532: sub prepare_algebra {
533: my ($string)=@_;
534:
535: # makes the decision about what is a minus sign easier supposedly
536: $string =~ s/(\<\>|\<\=|\>\=[\=\>\<] *)-/$1 zeroplace -/g;
537:
538: return $string;
539: }
540:
541: sub postprocess_algebra {
542: my ($string)=@_;
543:
544: # moodle had these and I don't know why, ignoring them for now
545: # $string =~s/\\fun/ /g;
546:
547: # sqrt(3,4) means the 4 root of 3
548: $string =~s/\\sqrt\{([^,]+),([^\}]+)}/\\sqrt[$2]{$1}/gs;
549:
550: # log(3,4) means the log base 4 of 3
551: $string =~s/\\log\\left\((.+?),(.+?)\\right\)/\\log_{$2}\\left($1\\right)/gs;
552:
553: # log(3,4) means the log base 4 of 3
554: $string =~s/\\((?:sin|cos|tan|sec|csc|cot)(?:h)?)\\left\((.+?),(.+?)\\right\)/\\$1^{$3}\\left($2\\right)/gs;
555:
556: # int(3,a,b) integral from a to b of 3
557: $string =~s/\\int\\left\((.+?),(.+?),(.+?)\\right\)/\\int_{$2}^{$3}\\left($1\\right)/gs;
558:
559: # int( ... dx) -> ...
560: $string =~s/\\int\\left\((.+?)d[a-z]\\right\)/$1/gs;
561:
562: #
563: $string =~s/\\lim\\left\((.+?),(.+?),(.+?)\\right\)/\\lim_{$2\\to $3}$1/gs;
564: return $string;
565: }
566:
567:
568: 1;
569: __END__
570:
571:
572: =pod
573:
574: =head1 NAME
575:
576: Apache::lontexconvert;
577:
578: =head1 SYNOPSIS
579:
580: Access to tth/ttm
581:
582: This is part of the LearningOnline Network with CAPA project
583: described at http://www.lon-capa.org.
584:
585:
586: =head1 SUBROUTINES
587:
588: =over
589:
590: =item init_tth()
591:
592: Header
593:
594: =item convert_real()
595:
596: we need this routine because &converted can get called from inside
597: of the safespace (through &xmlparse('<m>stuff</m>') which doesn't
598: allow the opcode for alarm, so we need to compile this before we get
599: into the safe space since opcode checks only occur at compile time
600:
601: =item tth_converted()
602:
603:
604: =item clean_out_math_mode()
605:
606:
607: =item displaystyle()
608:
609:
610: =item MathJax_converted()
611:
612: =item tex_engine()
613:
614: =item init_math_support()
615:
616: =item mimetex_valign()
617:
618: Makes a HEAD call to /cgi-bin/mimetex.cgi via IO:: to retrieve the
619: vertical alignment, before the subsequent call to mimetex_converted()
620: which generates the <img> tag and the corresponding image.
621:
622: Input: 1. $esc_texstring (escaped TeX to be rendered by mimetex).
623: Output: 1. $valign - number of pixels: positive or negative integer
624: which will be included in <img> tag for mimetex image to
625: support vertical alignment of image within a line of text.
626:
627: If a server is running SSL, and Apache rewrite rules are in place
628: to rewrite requests for http to https, modification will most likely
629: be needed for pass through for HEAD requests for /cgi-bin/mimetex.cgi.
630:
631: Example rewrite rules which rewrite all http traffic to https,
632: except HEAD requests for /cgi-bin/mimetex.cgi are:
633:
634: <IfModule mod_rewrite.c>
635: RewriteEngine On
636: RewriteLogLevel 0
637:
638: RewriteCond %{HTTPS} off
639: RewriteCond %{HTTP:Host} (.*)
640: RewriteCond %{REQUEST_METHOD} !HEAD
641: RewriteRule ^/(.*) https://%1/$1 [R=301,L]
642:
643: RewriteCond %{HTTPS} off
644: RewriteCond %{HTTP:Host} (.*)
645: RewriteCond %{REQUEST_METHOD} HEAD
646: RewriteCond %{REQUEST_URI} !^/cgi-bin/mimetex.cgi
647: RewriteRule ^/(.*) https://%1/$1 [R=301,L]
648: </IfModule>
649:
650: =item mimetex_converted()
651:
652:
653: =item converted()
654:
655:
656: =item to_convert()
657:
658: message display
659:
660: =item smiley()
661:
662: ???
663:
664: =item msgtexconverted()
665:
666: =item algebra()
667:
668: =item prepare_algebra()
669:
670: =item postprocess_algebra()
671:
672: =back
673:
674: =cut
675:
676:
677:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>