diff options
Diffstat (limited to 'tex/context/base/x-asciimath.mkiv')
| -rw-r--r-- | tex/context/base/x-asciimath.mkiv | 421 |
1 files changed, 358 insertions, 63 deletions
diff --git a/tex/context/base/x-asciimath.mkiv b/tex/context/base/x-asciimath.mkiv index b555115ff..1d62fb93d 100644 --- a/tex/context/base/x-asciimath.mkiv +++ b/tex/context/base/x-asciimath.mkiv @@ -1,6 +1,6 @@ %D \module -%D [ file=m-asciimath, -%D version=2006.04.24, % 1999.11.06, +%D [ file=x-asciimath, +%D version=2014.06.01, % 2006.04.24, % 1999.11.06, %D title=\CONTEXT\ Modules, %D subtitle=AsciiMath, %D author=Hans Hagen, @@ -11,88 +11,383 @@ %C therefore copyrighted by \PRAGMA. See mreadme.pdf for %C details. -%D Lua code. - \registerctxluafile{x-asciimath}{} -\def\ctxmoduleasciimath#1{\ctxlua{moduledata.asciimath.#1}} - -%D The following code is not officially supported and is only meant -%D for the Math4All project. +%D When the Math4All project started, we immediately started using content \MATHML. +%D Because in school math there is often a reference to calculator input, we also +%D provided what we called \quote {calcmath}: a predictable expression based way +%D entering math. At some point \OPENMATH\ was also used but that was later +%D abandoned because editing is more cumbersome. %D -%D The following code kind of maps ascii math -%D http://www1.chapman.edu/~jipsen/mathml/asciimath.html onto \TEX. The -%D code was written for the math4all project but in retrospect we -%D could have used just tex code as the web version can handle that -%D as well. Anyhow, as we use \MATHML\ as basis it makes sense to add -%D this to the repertoire as annotation variant, so now we have -%D content \MATHML\ (prefered), presentation \MATHML\ (often messy), -%D \OPENMATH\ (what was which we started with in this project) -%D calcmath (handy for students who are accustomed to calculators), -%D asciimath (to make Frits's live easier) and of course \TEX. Of -%D course all are used mixed. +%D Due to limitations in the web variant (which is independent of rendering for +%D paper but often determines the coding of document, not seldom for the worse) the +%D switch was made to presentational \MATHML. But even that proved to be too complex +%D for rendering on the web, so it got converted to so called \ASCIIMATH\ which +%D can be rendered using some \JAVASCRIPT\ magic. However, all the formulas (and +%D we're talking of tens of thousands of them) were very precisely coded by the main +%D author. Because in intermediate stages of the editing (by additional authors) a +%D mixture of \MATHML\ and \ASCIIMATH\ was used, we wrote the first version of this +%D module. As reference we took \url +%D {http://www1.chapman.edu/~jipsen/mathml/asciimath.html} and. The idea was to +%D stick to \MATHML\ as reference and if needed use \ASCIIMATH\ as annotation. %D -%D We don't support all quirks of asciimath as I am not in the mood to -%D write a complex parser while a bit of sane coding can work as well. +%D Eventually we ended up with supporting several math encodings in \CONTEXT\ that +%D could be used mixed: content \MATHML\ (preferred), presentation \MATHML\ (often +%D messy), \OPENMATH\ (somewhat minimalistic) calcmath (handy for students who are +%D accustomed to calculators), \ASCIIMATH\ (to make web support easier) and of +%D course \TEX. %D +%D The first version had some limitations as we didn't want to support all quirks of +%D \ASCIIMATH\ and also because I was not really in the mood to write a complex parser +%D when a bit of sane coding can work equally well. Some comments from that version: +%D +%D \startnarrower %D \startitemize -%D \item We support only the syntactically clear variants and as long -%D as lpeg does not support left recursion this is as far as we -%D want to go. -%D \item The parser is rather insensitive for spaces but yet the advice is -%D to avoid weird coding like \type {d/dxf(x)} but use \type {d/dx -%D f(x)} instead. After all we're not in a compact coding cq.\ -%D parser challenge. -%D \item We also don't support the somewhat confusing \type {sqrt sqrt 2} -%D nor \type {root3x} (although the second one kind of works). A bit -%D of defensive coding does not hurt. -%D \item We can process \type {a/b/c/d} but it's not compatible with the -%D default behaviour of asciimath. Use grouping instead. Yes, we do -%D support the somewhat nonstandard grouping token mix. -%D \item You should use explicit \type {text(..)} directives as one can -%D never be sure what is a reserved word and not. +%D \item We support only the syntactically clear variants and as long as lpeg does +%D not support left recursion this is as far as we want to go. +%D \item The parser is rather insensitive for spaces but yet the advice is to avoid +%D weird coding like \type {d/dxf(x)} but use \type {d/dx f(x)} instead. After +%D all we're not in a compact coding cq.\ parser challenge. +%D \item We also don't support the somewhat confusing \type {sqrt sqrt 2} nor \type +%D {root3x} (although the second one kind of works). A bit of defensive coding +%D does not hurt. +%D \item We can process \type {a/b/c/d} but it's not compatible with the default +%D behaviour of \ASCIIMATH. Use grouping instead. Yes, we do support the somewhat +%D nonstandard grouping token mix. +%D \item You should use explicit \type {text(..)} directives as one can never be sure +%D what is a reserved word and not. %D \stopitemize %D -%D Actually, as the only parsing sensitive elements of \TEX\ are -%D fractions (\type {\over} and friends, a restricted use of \TEX\ -%D coding is probably as comprehensive and parseble. -%D -%D The webpage with examples served as starting point so anything beyond +%D Actually, as the only parsing sensitive elements of \TEX\ are fractions (\type {\over} +%D and friends, a restricted use of \TEX\ coding is probably as comprehensive and +%D parsable. The webpage with examples served as starting point so anything beyond %D what can be found there isn't supported. +%D \stopnarrower +%D +%D Then in 2014 something bad happened. Following the fashion of minimal encoding +%D (which of course means messy encoding of complex cases and which can make authors +%D sloppy too) the web based support workflow of the mentioned project ran into some +%D limitations and magically one day all carefully coded \MATHML\ was converted into +%D \ASCIIMATH. As there was no way to recover the original thousands of files and +%D tens of thousands of formulas we were suddenly stuck with \ASCIIMATH. Because the +%D conversion had be done automagically, we also saw numerous errors and were forced +%D to come up with some methods to check formulas. Because \MATHML\ poses some +%D restrictions it has predictable rendering; \ASCIIMATH\ on the other hand enforces +%D no structure. Also, because \MATHML\ has to be valid \XML\ it always processes. +%D Of course, during the decade that the project had run we also had to built in +%D some catches for abuse but at least we had a relatively stable and configurable +%D subsystem. So, in order to deal with less predictable cases as well as extensive +%D checking, a new \ASCIIMATH\ parser was written, one that could also be used to +%D trace bad coding. +%D +%D Because the formal description is incomplete, and because some links to resources +%D are broken, and because some testing on the web showed that sequences of characters +%D are interpreted that were not mentioned anywhere (visible), and because we noticed +%D that the parser was dangerously tolerant, the new code is quite different from the +%D old code. +%D +%D One need to keep in mind that because spaces are optional, the only robust way to +%D edit \ASCIIMATH\ is to use a \WYSIWYG\ editor and hope that the parser doesn't +%D change ever. Keys are picked up from spaceless sequences and when not recognized +%D a (sequence) of characters is considered to be variables. So, \type {xsqrtx} is +%D valid and renders as \type {$x\sqrt{x}$}, \type {xx} becomes \type {×} (times) +%D but \type {ac} becomes \type {$a c$} (a times c). We're lucky that \type {AC} is +%D not turned into Alternating Current, but who knows what happens a few years from +%D now. So, we do support this spaceless mess, but users are warned: best use a +%D spacy sequence. The extra amount of spaces (at one byte each) an author has to +%D include in his|/|her active writing time probably stays below the size of one +%D holiday picture. Another complication is that numbers (in Dutch) use commas instead +%D of periods, but vectors use commas as well. We also hav esome different names for +%D functions which then can conflict with the expectations about collapsed variables. +%D +%D It must be noted that simplified encodings (that seem to be the fashion today) +%D can demand from applications to apply fuzzy logic to make something work out +%D well. Because we have sequential data that gets rendered, sometimes wrong input +%D gets obscured simply by the rendering: like the comma's in numbers as well as +%D for separators (depending on space usage), or plain wrong symbols that somehow +%D get a representation anyway. This in itself is more a side effect of trying to +%D use the simplified encoding without applying rules (in the input) or to use it +%D beyong its intended usage, which then of course can lead to adapted parsers and +%D catches that themselves trigger further abuse. Imagine that instead of developing +%D new cars, planes, space ships, mobile phones, computers we would have adapted +%D horse cars, kites, firework, old fashioned phones and mechanical calculators in a +%D similar way: patch upon patch of traditional means for sure would not have +%D worked. So, when you use \ASCIIMATH\ best check immediately how it gets rendered +%D in the browser as well as on paper. And be prepared to check the more complex +%D code in the future again. We don't offer any guarantees but of course will try to +%D keep up. +%D +%D In retrospect I sometimes wonder if the energy put into constantly adapting to +%D the fashion of the day pays off. Probably not. It definitely doesn't pay of. +%D +%D More complex crap: +%D +%D 1: $x + \stackrel{comment}{\stackrel{\utfchar{"23DE}}{yyyyyyyy}} = y$ \blank +%D 2: \asciimath{x + stackrel{\utfchar{"23DE}}{yyyyyyyy} = y} \blank +%D 3: \asciimath{x + stackrel{yyyyyyyy}{\utfchar{"23DE}} = y} \blank +%D 4: \asciimath{x + stackrel{"comment"}{stackrel{\utfchar{"23DE}}{yyyyyyyy}} = y} \blank + +\usemodule[mathml-basics] + +\startmodule[asciimath] \unprotect \writestatus{asciimath}{beware, this is an experimental (m4all only) module} -\unexpanded\def\asciimath#1{\ctxmoduleasciimath{convert(\!!bs\detokenize{#1}\!!es,true)}} +%D Hacks: + +\unexpanded\def\asciimathoptext #1{\ifmmode\mathoptext{#1}\else#1\fi} +\unexpanded\def\asciimathoptexttraced #1{\ifmmode\mathoptext{\color[darkgreen]{#1}}\else\color[darkgreen]{#1}\fi} +\unexpanded\def\asciimathstackrel #1#2{\mathematics{\mathop{\let\limits\relax\mover{#2}{#1}}}} +\unexpanded\def\asciimathroot #1#2{\sqrt[#1]{#2}} +\unexpanded\def\asciimathsqrt #1{\sqrt{#1}} + +%D The core commands: + +% if we need to set + +\installsetuponlycommandhandler {asciimath} {asciimath} + +\appendtoks + \ctxlua{moduledata.asciimath.setup { + splitmethod = "\asciimathparameter\c!splitmethod", + separator = "\asciimathparameter\c!separator", + }}% +\to \everysetupasciimath + +\newtoks\everyascimath + +% \appendtoks +% \ignorediscretionaries +% \to \everyasciimath + +\unexpanded\def\asciimath + {\doifnextoptionalelse\asciimath_yes\asciimath_nop} + +\def\asciimath_yes[#1]#2% + {\mathematics[#1]{\clf_justasciimath{\detokenize\expandafter{\normalexpanded{#2}}}}} + +\def\asciimath_nop#1% + {\mathematics{\clf_justasciimath{\detokenize\expandafter{\normalexpanded{#1}}}}} + +\unexpanded\def\ctxmoduleasciimath#1% + {\ctxlua{moduledata.asciimath.#1}} + +%D Some tracing commands. Using tex commands is 10\% slower that directly piping +%D from \LUA, but this is non|-|critical code. + +\unexpanded\def\ShowAsciiMathLoad [#1]{\ctxlua{moduledata.asciimath.show.load("#1")}} +\unexpanded\def\ShowAsciiMathIgnore[#1]{\ctxlua{moduledata.asciimath.show.ignore("#1")}} +\unexpanded\def\ShowAsciiMathXML #1#2{\ctxlua{moduledata.asciimath.show.filter("#1","#2")}} +\unexpanded\def\ShowAsciiMathStats {\ctxlua{moduledata.asciimath.show.statistics()}} +\unexpanded\def\ShowAsciiMathMax {\ctxlua{moduledata.asciimath.show.max()}} + +\unexpanded\def\ShowAsciiMathResult#1% + {\begingroup + \blank + % if we are in vmode, we don't get positions i.e. a smaller tuc file + \inleft{\ttbf#1\hfill\ctxlua{moduledata.asciimath.show.count(#1,true)}}% + \dontleavehmode + \begingroup + \ttbf + \ctxlua{moduledata.asciimath.show.files(#1)} + \endgroup + \blank[medium,samepage] + \startcolor[darkblue] + \ctxlua{moduledata.asciimath.show.input(#1,true)} + \stopcolor + \blank[medium,samepage] + \doifmode{asciimath:show:dirty} { + \dorecurse{\ctxlua{moduledata.asciimath.show.nofdirty(#1)}} { + \ctxlua{moduledata.asciimath.show.dirty(\recurselevel,true)} + \blank[medium,samepage] + } + } + \ctxlua{moduledata.asciimath.show.result(#1)} + \blank + \endgroup} + +\unexpanded\def\ShowAsciiMathStart + {\begingroup + \let\asciimathoptext\asciimathoptexttraced + \setuptyping[\v!buffer][\c!before=,\c!after=] + \setupmargindata[\v!left][\c!style=]} + +\unexpanded\def\ShowAsciiMathStop + {\endgroup} + +\unexpanded\def\ShowAsciiMath + {\dodoubleempty\doShowAsciiMath} + +\unexpanded\def\doShowAsciiMath[#1][#2]% + {\iffirstargument + \ShowAsciiMathStart + \ShowAsciiMathLoad[#1] + \ifsecondargument + \ShowAsciiMathIgnore[#2] + \fi + \dorecurse{\ShowAsciiMathMax}{\ShowAsciiMathResult\recurselevel} + \page + \ShowAsciiMathStats + \ShowAsciiMathStop + \fi} + +\unexpanded\def\xmlShowAsciiMath#1#2% + {\iffirstargument + \ShowAsciiMathStart + \ShowAsciiMathXML{#1}{#2}% + \dorecurse{\ShowAsciiMathMax}{\ShowAsciiMathResult\recurselevel} + \page + \ShowAsciiMathStats + \ShowAsciiMathStop + \fi} + +\unexpanded\def\ShowAsciiMathSave + {\dosingleempty\doShowAsciiMathSave} + +\unexpanded\def\doShowAsciiMathSave[#1]% + {\ctxlua{moduledata.asciimath.show.save("#1")}} \protect +\startsetups asciimath:layout + + \setupbodyfont + % [pagella,10pt] + [dejavu,10pt] + + \setuplayout + [backspace=35mm, + leftmargin=20mm, + rightmargindistance=0pt, + leftmargindistance=5mm, + cutspace=1cm, + topspace=1cm, + bottomspace=1cm, + width=middle, + height=middle, + header=0cm, + footer=1cm] + + \setupheadertexts + [] + + \setupfootertexts + [\currentdate][\pagenumber] + + \setupalign + [flushleft,verytolerant,stretch] + + \dontcomplain + +\stopsetups + +\stopmodule + \continueifinputfile{x-asciimath.mkiv} -\enabletrackers[modules.asciimath.mapping] +%D This will become an extra. -\starttext +\showframe + +\setups[asciimath:layout] + +% \enabletrackers[modules.asciimath.mapping] +% \enabletrackers[modules.asciimath.detail] +% \starttext +% \enablemode[asciimath:show:dirty] +% \ShowAsciiMath[e:/temporary/asciimath/*.xml] +% % \ShowAsciiMathSave[e:/temporary/asciimath/asciimath.lua] +% \stoptext + +\starttext +\unexpanded\def\MyAsciiMath#1{\startformula\asciimath{#1}\stopformula} \startlines -\asciimath{x^2+y_1+z_12^34} -\asciimath{sin^-1(x)} -\asciimath{d/dx f(x)=lim_(h->0) (f(x+h)-f(x))/h} -\asciimath{f(x)=sum_(n=0)^oo(f^((n))(a))/(n!)(x-a)^n} -\asciimath{int_0^1 f(x)dx} -\asciimath{int^1_0 f(x)dx} -\asciimath{a//b} -\asciimath{(a/b)/(d/c)} -\asciimath{((a*b))/(d/c)} -\asciimath{[[a,b],[c,d]]((n),(k))} -\asciimath{1/x={(1,text{if } x!=0),(text{undefined},if x=0):}} -\asciimath{{ (1,2), (x,(x + text(x))) }} -\asciimath{{(1,2),(x,(x+text(x))),(x,text(x))}} -\asciimath{{(1,2),(x,(x+text(x))),(x,x text(x))}} -\asciimath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,x text(xyz))}} -\asciimath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,text(xyz)+1+text(hans))}} -\asciimath{<<a,b>> text{and} {:(x,y),(u,v):}} -\asciimath{(a,b] = {x text(in) RR | a < x <= b}} +\MyAsciiMath{x^2 / 10 // z_12^34 / 20} +% \MyAsciiMath{{:{:x^2:} / 10:} // {:{:z_12^34 :} / 20:}} +% \MyAsciiMath{x^2+y_1+z_12^34} +% \MyAsciiMath{sin^-1(x)} +% \MyAsciiMath{d/dx f(x)=lim_(h->0) (f(x+h)-f(x))/h} +% \MyAsciiMath{f(x)=sum_(n=0)^oo(f^((n))(a))/(n!)(x-a)^n} +% \MyAsciiMath{int_0^1 f(x)dx} +% \MyAsciiMath{int^1_0 f(x)dx} +% \MyAsciiMath{a//b} +% \MyAsciiMath{a//\alpha} +% \MyAsciiMath{(a/b)/(d/c)} +% \MyAsciiMath{((a*b))/(d/c)} +% \MyAsciiMath{[[a,b],[c,d]]((n),(k))} +% \MyAsciiMath{1/x={(1,text{if } x!=0),(text{undefined},if x=0):}} +% \MyAsciiMath{{ (1,2), (x,(x + text(x))) }} +% \MyAsciiMath{{(1,2),(x,(x+text(x))),(x,text(x))}} +% \MyAsciiMath{{(1,2),(x,(x+text(x))),(x,x text(x))}} +% \MyAsciiMath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,x text(xyz))}} +% \MyAsciiMath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,text(xyz)+1+text(hans))}} +% \MyAsciiMath{<<a,b>> text{and} {:(x,y),(u,v):}} +% \MyAsciiMath{(a,b] = {x text(in) RR | a < x <= b}} +% \MyAsciiMath{a/b / c/d = (a * d) / (b * d) / (b * c) / (b * d) = (a * d) / (b * c)} +% \MyAsciiMath{ (a/b) // (c/d) = ( (a * d) / (b * d) ) // ( (b * c) / (b * d) ) = (a * d) / (b * c)} +% \MyAsciiMath{sin(x+1)_3^2/b / c/d} +% \MyAsciiMath{{:{:sin(x+1)_3^2:}/b:} / {:c/d:}} +% \MyAsciiMath{cos(a) + sin(x+1)_3^2/b / c/d = (a * d) / (b * d) / (b * c) / (b * d) = (a * d) / (b * c)} +% \MyAsciiMath{S_(11)} +% \MyAsciiMath{f(x)} +% \MyAsciiMath{sin(x)} +% \MyAsciiMath{sin(x+1)} +% \MyAsciiMath{sin^-1(x)} +% \MyAsciiMath{sin(2x)} +% \MyAsciiMath{a_2^2} +% \MyAsciiMath{( (S_(11),S_(12),S_(1n)),(vdots,ddots,vdots),(S_(m1),S_(m2),S_(mn)) ]} +% \MyAsciiMath{frac a b} +% \MyAsciiMath{sin(x)/2 // cos(x)/pi} +% \MyAsciiMath{a/13 // c/d} +% \MyAsciiMath{a/b // c/d} +% \MyAsciiMath{x} +% \MyAsciiMath{x^2} +% \MyAsciiMath{sqrt x} +% \MyAsciiMath{sqrt (x)} +% \MyAsciiMath{root 2 x} +% \MyAsciiMath{x+x} +% \MyAsciiMath{x/3} +% \MyAsciiMath{x^2 / 10} +% \MyAsciiMath{x^2 / 10 // z_12^34 / 20} +% \MyAsciiMath{a^23} +% \MyAsciiMath{a^{:b^23:}+3x} +% \MyAsciiMath{a/b / c/d} +% \MyAsciiMath{sin(x)/b / c/d} +% \MyAsciiMath{sin(x)/b // c/d} +% \MyAsciiMath{a/b / c/d = (a * d) / (b * d) / (b * c) / (b * d) = (a * d) / (b * c) } +% \MyAsciiMath{{:{:x^2:} / 10:} // {:{:z_12^34 :} / 20:}} +% \MyAsciiMath{x^2+y_1+z_12^34} +% \MyAsciiMath{sin^-1(x)} +% \MyAsciiMath{d/dx f(x)=lim_(h->0) (f(x+h)-f(x))/h} +% \MyAsciiMath{f(x)=sum_(n=0)^oo(f^((n))(a))/(n!)(x-a)^n} +% \MyAsciiMath{int_0^1 f(x)dx} +% \MyAsciiMath{int^1_0 f(x)dx} +% \MyAsciiMath{2x} +% \MyAsciiMath{a//b} +% \MyAsciiMath{a//\alpha} +% \MyAsciiMath{(a/b)/(d/c)} +% \MyAsciiMath{((a*b))/(d/c)} +% \MyAsciiMath{[[a,b],[c,d]]((n),(k))} +% \MyAsciiMath{1/x={(1,text{if } x!=0),(text{undefined},if x=0):}} +% \MyAsciiMath{{ (1,2), (x,(x + text(x))) }} +% \MyAsciiMath{{(1,2),(x,(x+text(x))),(x,text(x))}} +% \MyAsciiMath{{(1,2),(x,(x+text(x))),(x,x text(x))}} +% \MyAsciiMath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,x text(xyz))}} +% \MyAsciiMath{{(1,2/2),(x,(x+x^22+sqrt(xx))),(x,text(xyz)+1+text(hans))}} +% \MyAsciiMath{<<a,b>> text{and} {:(x,y),(u,v):}} +% \MyAsciiMath{(a,b] = {x text(in) RR | a < x <= b}} +% \MyAsciiMath{x^-2} +% \MyAsciiMath{x^2(x-1/16)=0} +% \MyAsciiMath{y= ((1/4)) ^x} +% \MyAsciiMath{log (0,002) / (log(1/4))} +% \MyAsciiMath{x=ax+b \ oeps} +% \MyAsciiMath{x=\ ^ (1/4) log(x)} +% \MyAsciiMath{x=\ ^ (1/4) log(0 ,002 )= log(0,002) / (log(1/4))} +% \MyAsciiMath{x^ (-1 1/2) =1/x^ (1 1/2)=1/ (x^1*x^ (1/2)) =1/ (xsqrt(x))} +% \MyAsciiMath{x^2(10 -x)>2 x^2} +% \MyAsciiMath{x^4>x} \stoplines - \stoptext |