I use lua to make some complex job to prepare arguments for macros in Tex/LaTex.
Part I
Here is a stupid minimal example :
\newcommand{\test}{\luaexec{tex.print("11,12")}}% aim to create 11,12
\def\compare#1,#2.{\ifthenelse{#1<#2}{less}{more}}
\string\compare11,12. : \compare11,12.\\ %answer is less
\string\test : \test\\ % answer is 11,12
\string\compare : \compare\test. % generate an error
The last line creates an error. Obviously, Tex did not detect the "," included in \test.
How can I do so that \test is understood as 11 followed by , followed by 12 and not the string 11,12 and finally used as a correctly formed argument for \compare ?
There are several misunderstandings of how TeX works.
Your \compare macro wants to find something followed by a comma, then something followed by a period. However when you call
\compare\test
no comma is found, so TeX keeps looking for it until finding either the end of file or a \par (or a blank line as well). Note that TeX never expands macros when looking for the arguments to a macro.
You might do
\expandafter\compare\test.
provided that \test immediately expands to tokens in the required format, which however don't, because the expansion of \test is
\luaexec{tex.print("11,12")}
and the comma is hidden by the braces, so it doesn't count. But it wouldn't help nonetheless.
The problem is the same: when you do
\newcommand{\test}{\luaexec{tex.print("11,12")}}
the argument is not expanded. You might use “expanded definition” with \edef, but the problem is that \luaexec is not fully expandable.
If you do
\edef\test{\directlua{tex.sprint("11,12")}}
then
\expandafter\compare\test.
would work.
Related
I am newbie in Lua. I write a script in which there is a line
ISH_activation.getState() = CONST_ACTIVATION_STATE$ON
Then, I need to write another C++ script to read this line from Lua and parse it. Now the problem is that, when I try to debug the lua script itself, at this line it gives me the error "unexpected symbol near '$'".
I don't know why this would happen. Is format like "CONST_ACTIVATION_STATE$ON" not allowed in Lua? And is "$" a special character in Lua that we cannot use it?(But I didn't find anything mentioning this)....
Well, did you find anything saying that you can use $? :)
Yes, you cannot use dollar signs in identifiers. From the manual:
Names (also called identifiers) in Lua can be any string of letters, digits, and underscores, not beginning with a digit. Identifiers are used to name variables, table fields, and labels.
If you really want to keep the dollar sign you could use strings instead
"CONST_ACTIVATION_STATE$ON"
I have been coding a program in Lua that automatically formats IRC logs from a roleplay. In the roleplay logs there is a specific guideline for "Out of character" conversation, which we use double parentheses for. For example: ((<Things unrelated to roleplay go here>)). I have been trying to have my program remove text between double brackets (and including both brackets). The code is:
ofile = io.open("Output.txt", "w")
rfile = io.open("Input.txt", "r")
p = rfile:read("*all")
w = string.gsub(p, "%(%(.*?%)%)", "")
ofile:write(w)
The pattern here is > "%(%(.*?%)%)" I've tried multiple variations of the pattern. All resulted in fruitless results:
1. %(%(.*?%)%) --Wouldn't do anything.
2. %(%(.*%)%) --Would remove *everything* after the first OOC message.
Then, my friend told me that prepending the brackets with percentages wouldn't work, and that I had to use backslashes to 'escape' the parentheses.
3. \(\(.*\)\) --resulted in the output file being completely empty.
4. (\(\(.*\)\)) --Same result as above.
5. (\(\(.*?\)\) --would for some reason, remove large parts of the text for no apparent reason.
6. \(\(.*?\)\) --would just remove all the text except for the last line.
The short, absolute question:
What pattern would I need to use to remove all text between double parentheses, and remove the double parentheses themselves too?
You're friend is thinking of regular expressions. Lua patterns are similar, but different. % is the correct escape character.
Your pattern should be %(%(.-%)%). The - is similar to * in that it matches any number of the preceding sequence, but while * tries to match as many characters as it can (it's greedy), - matches the least amount of characters possible (it's non-greedy). It won't go overboard and match extra double-close-parenthesis.
I need to remove all \text generated by TeXForm in Mathematica.
What I am doing now is this:
MyTeXForm[a_]:=StringReplace[ToString[TeXForm[a]], "\\text" -> ""]
But the result keeps the braces, for example:
for a=fx,
the result of TeXForm[a] is \text{fx}
the result of MyTeXForm[a] is {fx}
But what I would like is it to be just fx
You should be able to use string patterns. Based on http://reference.wolfram.com/mathematica/tutorial/StringPatterns.html, something like the following should work:
MyTeXForm[a_]:=StringReplace[ToString[TeXForm[a]], "\\text{"~~s___~~"}"->s]
I don't have Mathematica handy right now, but this should say 'Match "\text{" followed by zero or more characters that are stored in the variable s, followed by "}", then replace all of that with whatever is stored in s.'
UPDATE:
The above works in the simplest case of there being a single "\text{...}" element, but the pattern s___ is greedy, so on input a+bb+xx+y, which Mathematica's TeXForm renders as "a+\text{bb}+\text{xx}+y", it matches everything between the first "\text{" and last "}" --- so, "bb}+\text{xx" --- leading to the output
In[1]:= MyTeXForm[a+bb+xx+y]
Out[1]= a+bb}+\text{xx+y
A fix for this is to wrap the pattern with Shortest[], leading to a second definition
In[2]:= MyTeXForm2[a_] := StringReplace[
ToString[TeXForm[a]],
Shortest["\\text{" ~~ s___ ~~ "}"] -> s
]
which yields the output
In[3]:= MyTeXForm2[a+bb+xx+y]
Out[3]= a+bb+xx+y
as desired.
Unfortunately this still won't work when the text itself contains a closing brace. For example, the input f["a}b","c}d"] (for some reason...) would give
In[4]:= MyTeXForm2[f["a}b","c}d"]]
Out[4]= f(a$\$b},c$\$d})
instead of "f(a$\}$b,c$\}$d)", which would be the proper processing of the TeXForm output "f(\text{a$\}$b},\text{c$\}$d})".
This is what I did (works fine for me):
MyTeXForm[a_] := ToString[ToExpression[StringReplace[ToString[TeXForm[a]], "\\text" -> ""]][[1]]]
This is a really late reply, but I just came up against the same issue and discovered a simple solution. Put a space between the variables in the Mathematica expression that you wish to convert using TexForm.
For the original poster's example, the following code works great:
a=f x
TeXForm[a]
The output is as desired: f x
Since LaTeX will ignore that space in math mode, things will format correctly.
(As an aside, I was having the same issue with subscripted expressions that have two side-by-side variables in the subscript. Inserting a space between them solved the issue.)
I am using TeXnicCenter to edit a LaTeX document.
I now want to remove a certain tag (say, emph{blabla}} which occurs multiple times in my document , but not tag's content (so in this example, I want to remove all emphasization).
What is the easiest way to do so?
May also be using another program easily available on Windows 7.
Edit: In response to regex suggestions, it is important that it can deal with nested tags.
Edit 2: I really want to remove the tag from the text file, not just disable it.
Using a regular expression do something like s/\\emph\{([^\}]*)\}/\1/g. If you are not familiar with regular expressions this says:
s -- replace
/ -- begin match section
\\emph\{ -- match \emph{
( -- begin capture
[^\}]* -- match any characters except (meaning up until) a close brace because:
[] a group of characters
^ means not or "everything except"
\} -- the close brace
and * means 0 or more times
) -- end capture, because this is the first (in this case only) capture, it is number 1
\} -- match end brace
/ -- begin replace section
\1 -- replace with captured section number 1
/ -- end regular expression, begin extra flags
g -- global flag, meaning do this every time the match is found not just the first time
This is with Perl syntax, as that is what I am familiar with. The following perl "one-liners" will accomplish two tasks
perl -pe 's/\\emph\{([^\}]*)\}/\1/g' filename will "test" printing the file to the command line
perl -pi -e 's/\\emph\{([^\}]*)\}/\1/g' filename will change the file in place.
Similar commands may be available in your editor, but if not this will (should) work.
Crowley should have added this as an answer, but I will do that for him, if you replace all \emph{ with { you should be able to do this without disturbing the other content. It will still be in braces, but unless you have done some odd stuff it shouldn't matter.
The regex would be a simple s/\\emph\{/\{/g but the search and replace in your editor will do that one too.
Edit: Sorry, used the wrong brace in the regex, fixed now.
\renewcommand{\emph}[1]{#1}
any reasonably advanced editor should let you do a search/replace using regular expressions, replacing emph{bla} by bla etc.
I want to redefine the \part* command so that it automatically adds a contents line. This proves difficult since I want to reuse the original \part* command inside my starred version.
Normally (i.e. for unstarred commands) I would do it like this:
\let\old#part\part
\renewcommand\part[2][]{
\old#part[#1]{#2}
… rest of definition}
That is, I would save the original definition of \part in \old#part and use that.
However, this doesn’t work for starred commands since they don’t define a single lexeme (unlike the \part command in the example above). This boils down to the following question: How can I save a starred command?
Notice that I already know how to redefine a starred command itself, using the \WithSuffix command from the suffix package. This isn’t the problem.
There is no \part* command. What happens is the \part command takes a look at the next character after it (with \#ifstar) and dispatches to one of two other routines that does the actual work based on whether there's an asterisk there or not.
Reference: TeX FAQ entry Commands defined with * options
Thanks to #smg’s answer, I’ve cobbled together a solution that works perfectly. Here’s the complete source, along with explanatory comments:
% If this is in *.tex file, uncomment the following line.
%\makeatletter
% Save the original \part declaration
\let\old#part\part
% To that definition, add a new special starred version.
\WithSuffix\def\part*{
% Handle the optional parameter.
\ifx\next[%
\let\next\thesis#part#star%
\else
\def\next{\thesis#part#star[]}%
\fi
\next}
% The actual macro definition.
\def\thesis#part#star[#1]#2{
\ifthenelse{\equal{#1}{}}
{% If the first argument isn’t given, default to the second one.
\def\thesis#part#short{#2}
% Insert the actual (unnumbered) \part header.
\old#part*{#2}}
{% Short name is given.
\def\thesis#part#short{#1}
% Insert the actual (unnumbered) \part header with short name.
\old#part*[#1]{#2}}
% Last, add the part to the table of contents. Use the short name, if provided.
\addcontentsline{toc}{part}{\thesis#part#short}
}
% If this is in *.tex file, uncomment the following line.
%\makeatother
(This needs the packages suffix and ifthen.)
Now, we can use it:
\part*{Example 1}
This will be an unnumbered part that appears in the TOC.
\part{Example 2}
Yes, the unstarred version of \verb/\part/ still works, too.