Treat environment content as string, replace substrings and interpret afterwards

Question

I would like to define an environment which "parses" its content and changes it before displaying it.

An additional difficulty comes because I also need to replace whitespaces. The more flexibility in the replacement tools (regex would be ideal) the better.

I'm aware or xstring's StrSubstitute and this solution here: How to replace text. Yet they do not help so far.

Example:

\begin{changecontent}
some text 
 some text 
  some text 
\end{changecontent}

Simplified, what I would like to get from that is something like:

\command\\
~\command\\
~~\command\\

which then would compile into something where \command would be expanded to whatever it defined to before. This is an arbitrary example. It is not really crucial what this command is. It could be defined in the preamble, e.g. it could draw a tikz square or just expand to simple text.

The crucial steps are:

1. keep spaces and newlines until the content is parsed, so that they can be used for the replacing part

2. do replacements: in regex that would be in this example:

   • ^.* => \command\\
   • ^\s.* => ~\command\\
   • ^\s\s.* => ~~\command\\

3. expand the result and output

There are no restrictions in terms of how this should be implemented.

Is this asking to much of *TeX?

Answer 1

Here's a LuaLaTeX-based solution. It defines an environment called changecontent. Within instances of this environment (and only within instances of this environment), all instances of the string "some text" are changed to \somecommand\par; and, all singleton or doubleton spaces that precede this string are converted to ~ (unbreakable spaces). If there are more than two spaces ahead of an instance of "some text", only the final two spaces are converted to ~. The Lua function that performs the substitutions is assigned to the process_input_buffer callback, which operates on the input stream at a very early stage, before the "eyes", "mouth", and "stomach" of TeX get to do anything.

For the code below, I've set up a dummy definition for \somecommand. You, should, of course, provide the real definition of \somecommand in the preamble of your document.

\documentclass{article}
\usepackage{luacode,luatexbase}

%% Lua-side code
\begin{luacode}
function sometexttocommand ( line )
    line = string.gsub ( line , "%s%ssome text", "~~\\somecommand\\par" )
    line = string.gsub ( line , "%ssome text",   "~\\somecommand\\par" )    
    line = string.gsub ( line , "some text",     "\\somecommand\\par" )
    return ( line )
end 
\end{luacode}

%% TeX-side code
\newenvironment{changecontext}{%
   \directlua{luatexbase.add_to_callback(
              "process_input_buffer",
              sometexttocommand, "sometexttocommand")}}{%
   \directlua{luatexbase.remove_from_callback(
              "process_input_buffer",
              "sometexttocommand")}}

% provide a dummy definition of "\somecommand"
\providecommand\somecommand{Hello World} 

\begin{document}
some text
 some text
  some text
   some text

\begin{changecontext}
some text
 some text
  some text
   some text
\end{changecontext}

some text
 some text
  some text
   some text

\end{document}