Ignoring all but certain text

Question

I'm wondering if there's a way to define a command so that, when a flag is specified, I can ignore everything but text within the tag. For instance, if I had

...
This is a \dontignore{bird}.

I would want the output to be either "This is a bird." or "bird" depending on whether or not I switch this flag.

Is this doable? Obviously surrounding everything but the text I want in some tag would be easy, but that seems quite inelegant.

Edit: Here's exactly what I have in mind:

\documentclass{article}
\newcommand{\dontignore}[1]{...}
\begin{document}
Hello world.  This is my full document with a lot of extra bells and whistles and lines.  But in the end, \dontignore{this is all that matters}.
\end{document}

With the flag off, the output should be

Hello world. This is my full document with a lot of extra bells and whistles and lines. But in the end, this is all that matters.

Otherwise,

this is all that matters

Optimally, I would also be able to tell LaTeX not to ignore things like section and enumerate/itemize/item commands as well, as follows

\documentclass{article}
\newcommand{\dontignore}[1]{...}
\begin{document}
\section{First part}
Hello world.  This is my full document with a lot of extra bells and whistles and lines.  But in the end, \dontignore{this is all that matters}.
\begin{enumerate}
  \item \dontignore{Item 1}
  \item Something like \dontignore{Item 2}
\end{enumerate}
\end{document}

to produce either everything (eg, with dontignore defined as \newcommand{\dontignore}[1]{#1}) or otherwise (with some flag specifying section, enumerate, etc.) as follows

1 First part

this is all that matters

1. Item 1
2. Item 2

Answer 1

---

---

In the following you get one or the other output shown above depending on whether this line is commented out or not

%\let\ignoreflag\relax

---

\documentclass{article}

\long\def\dontignore#1{#1}
\makeatletter

\long\def\ignoreflag{%
\@makeother\{%
\@makeother\}%
\xignore}
\long\def\xignore#1\dontignore#2{%
\catcode`\{\@ne
\catcode`\}\tw@
\afterassignment\xxdontignore\toks@\bgroup}
\long\def\xxdontignore{%
\the\toks@\ignoreflag}
\makeatother

%\let\ignoreflag\relax
\begin{document}
\ignoreflag

This is some text 

\begin{itemize}
\item aaa
\item bbb
\item ccc \dontignore{hello }
\end{itemize}

more text 
\dontignore{
\begin{enumerate}
\item this
\item that
\end{enumerate}
}

more stuff

\dontignore{\end{document}}

Answer 2

\documentclass{article}
\newif\ifToggle \Togglefalse % change to \Toggletrue
\makeatletter
\def\myText#1{\expandafter\myText@i#1\@nil}
\def\myText@i#1\dontignore#2\END#3\@nil{\ifToggle#1\fi#2\ifToggle#3\fi}
\makeatother
\begin{document}
\myText{%
Hello world.  This is my full document with a lot of extra 
bells and whistles and lines.  But in the end, \dontignore this 
is all that matters \END some more text.}
\end{document}

with \Toggletrue all text is printed.

Answer 3

Here is a LuaTeX based solution (in ConTeXt)

• I assume that the content is surrounding in an environment \startignore ... \stopignore.

• Ignoring is enabled or disabled using a mode (and, therefore, can be controlled via the command line)

• The gobbling of everything not inside \dontignore{...} tags is done in Lua using an lpeg parser that captures the contents \dontignore. The parser handles nested {...} correctly.

I don't know the LuaLaTeX equivalent of ConTeXt buffers, so I am not sure how to translate this to LuaLaTeX.

\let\dontignore\firstofoneargument

\startluacode
    local lpegmatch, patterns = lpeg.match, lpeg.patterns
    local P, V, Cs, C = lpeg.P, lpeg.V, lpeg.Cs, lpeg.C
    local format = string.format


    local backslash    = P("\\")
    local csname       = backslash * P("dontignore") 
    local whitespace   = lpeg.patterns.whitespace 
    local leftbrace    = P("{")
    local rightbrace   = P("}")
    local nonbrace     = 1 - leftbrace - rightbrace
    local nested       = P { leftbrace * (nonbrace + V(1))^0 * rightbrace }
    -- I am not sure if it makes sense to add spaces around the captured text or
    -- not. Below I assume that you want spaces around the next. If not, then
    -- you can just use
    --      local value        = P(csname * whitespace^0 * leftbrace * C((nested + nonbrace)^0) * rightbrace) / context
    local value        = P(csname * whitespace^0 * leftbrace * C((nested + nonbrace)^0) * rightbrace) / (function(s) context(" %s ", s) end)
    local nonvalue     = (1 - value)
    local parser       = ( (nonvalue^0 * value * nonvalue^0)^0 )

    thirddata = thirddata or {}

    local getcontent = buffers.getcontent

    -- Function that ignores all the data except that in \dontignore
    thirddata.ignoredata = function (name) 
        lpegmatch(parser, getcontent(name))
    end
\stopluacode

\unexpanded\def\startignoring
    {\grabbufferdata[ignorebuffer][startignoring][stopignoring]}

\unexpanded\def\stopignoring
    {\doifmodeelse{ignore}
        {\ctxlua{thirddata.ignoredata('ignorebuffer')}}
        {\getbuffer[ignorebuffer]}}

This can be used as follows

\starttext

\enablemode[ignore]

\startignoring
  \startitemize[n]
    \item This is \dontignore{item 1}
    \item And this is \dontignore{item 2}
  \stopitemize
\stopignoring

\disablemode[ignore]

\startignoring
  \startitemize[n]
    \item This is \dontignore{item 1}
    \item And this is \dontignore{item 2}
  \stopitemize
\stopignoring

\stoptext

which gives

The advantage of parsing content in Lua is that the anything outside \dontignore does not need to be valid TeX code. Therefore, even something like

\startignoring
  \undefined \dontignore{Does this work} 
\stopignoring

works correctly (of course, this will fail when ignore mode is disabled).