NAME

mathgrep - grep through a LaTeX document looking only in math sections

SYNOPSIS

mathgrep [OPTION]... CODE [FILE]...
Apply CODE to mathematics in each FILE or standard input.
Example: mathgrep '/\\hbar/' quantum.tex classical.tex

Options:
  -e, --code CODE  Use CODE
  -i, --[no]inline  Apply CODE to inline math
  -d, --[no]display  Apply CODE to display math
  -E, --environments ENVS  Apply CODE to ENVS
  -c, --count  Only print a count of matching lines per FILE
  -l, --files-with-matches  Only print FILE names containing matches
  -L, --files-without-matches  Only print FILE names containing no match
  -n, --line-number  Print line number with output lines
  -R, --preamble  Skip preamble (before \begin{document})
  -O, --postamble  Skip postamble (after \end{document})
  -r, --replace [EXT] Print all lines; if EXT is specified then backup
                     FILE to FILEEXT and redirect output to FILE
  -S, --start NUM  Start applying CODE at line NUM
  -F, --finish NUM  Stop applying CODE at line NUM
  -U, --colo[u]r [COLOUR]  Highlight matches in COLOUR (default: red)
  -P, --[no]separate  Print a separator between matches
  -C, --context NUM  Print NUM lines of output context
  -B, --before-context NUM  Print NUM lines of leading context
  -A, --after-context NUM  Print NUM lines of trailing context
  -h, -?, --help  Print help message
  --man  Print man page
  --gpl  Print GPL
  -v, --verbose  Increase verbosity
  -s, --quiet  Suppress verbosity
  -V, --version  Print version and exit

DESCRIPTION

mathgrep goes through the specified input FILEs (or standard input if no FILEs are given) applying CODE to any mathematics it finds. By default, mathgrep prints any lines containing mathematics on which CODE evaluated successfully.

The CODE is a bit of perl code which does something to $_. For grep-like performace try m/regexp/. mathgrep can also do search-and-replace.

The CODE is evaluated on the whole of a section of mathematics in one go, even if the mathematics splits across several lines. All the lines containing the section of mathematics are printed if CODE evaluated successfully.

The options and their behaviour are as like to those of grep as I could make them.

OPTIONS

-A NUM, --after-context NUM

Print NUM lines of trailing context after a line on which CODE executes successfully.
-B, --before-context NUM

Print NUM lines of leading context after a line on which CODE executes successfully.
-C, --context NUM

Print NUM lines of context either side of a line on which CODE executes successfully.
-c, --count

Suppress normal output, instead print a count of the number of sections of mathematics on which CODE evaluated successfully. Note: this may not be the exact number of matches as CODE is applied once to each section of mathematics.
-d, --[no]display

Toggles whether to look in display math environments.
-E, --environments ENV[,ENV]...

Allows the user to specify a list of extra environments to look in. These are always checked. Multiple instances of this option are allowed, as are comma-separated lists.
-e, --code CODE

Specifies the CODE to apply. This gives a way to ensure that the code is not swallowed up in another option. Another way is to finish the list of options with -- before specifying the CODE.
-F, --finish NUM

Finish checking the input after NUM lines. Subsequent lines may still be printed depending on the status of the --after-context, --context, or --replace options.
-h, -?, --help

Print a summary of all the options and exit.
-i, --[no]inline

Toggles whether to look in inline math environments.
-l, --files-with-matches

Suppress normal output, instead just print the FILE name of a FILE on which CODE evaluated successfully at least once. Can be meaningfully combined with --count.
-L, --files-without-matches

Suppress normal output, instead just print the FILE name of a FILE on which CODE did not evaluate successfully at all. Cannot be meaningfully combined with --count.
--man

Print this manpage and exit.
--gpl

Display the GNU Public License and exit (note: it actually displays the GPL that comes with Perl so if your installation of Perl is not set up correctly then this might not work).
-n, --line-number

Prefix each line of output with its line number within its input file. Note: if CODE has mucked about with the number of lines in a section of mathematics then the line count may go a little awry within that section but it will correct itself again for the next section.
-O, --postamble

Stop applying CODE to maths when \end{document} has been reached.
-P, --[no]separate

Toggle whether to print a separator between matches. Default is on, unless --replace has been specified. The separator is %---% so won't mess up running LaTeX on the output.
-R, --preamble

Don't start applying CODE to maths until \begin{document} has been reached.
-r, --replace [EXT]

Prints all lines whether CODE successfully ran or not. If EXT is specified then it backups up FILE to FILEEXT and redirects the output to FILE. Similar to the -i option of perl.
-s, --quiet

Suppress verbosity.
-S, --start NUM

Only apply CODE to maths found on or after line NUM. Note: depending on --before-context, --context, or --replace then lines before NUM may still be printed.
-U, --colo[u]r [COLOUR]

If CODE is of the form m/PATTERN/[cgimosx] or one of its simple variants then this option colours the matching text; red if COLOUR is not specified.
-v, --verbose

Increase the verbosity level.
-V, --version

Print version number and exit.

EXAMPLES

Suppose that quantum.tex is

\documentclass{article}

\newcommand{\texthbar}{\(\hbar\)}

\begin{document}

The number \(\hbar\) is very small.  The product, \(\hbar\hbar\) is
miniscule.   On the other hand, \(c\), the speed of light is very big.
\[
\hbar = \frac{h}{2\pi}
\]
\begin{myenv}
Think of hbar as your friend.
\end{myenv}

\end{document}

Some junk involving \(\hbar\).

Then the following examples produce the specified output

mathgrep '/hbar/' quantum.tex

%---%
\newcommand{\texthbar}{\(\hbar\)}
%---%
The number \(\hbar\) is very small.  The product, \(\hbar\hbar\) is
%---%
\[
\hbar = \frac{h}{2\pi}
\]
%---%
Some junk involving \(\hbar\).

mathgrep --replace -- 's/hbar/hfoo/g' quantum.tex

\documentclass{article}

\newcommand{\texthbar}{\(\hfoo\)}

\begin{document}

The number \(\hfoo\) is very small.  The product, \(\hfoo\hfoo\) is
miniscule.  On the other hand, \(c\), the speed of light is very big.
\[
\hfoo = \frac{h}{2\pi}
\]
\begin{myenv}
Think of hbar as your friend.
\end{myenv}

\end{document}

Some junk involving \(\hfoo\).

Note the g option on the s///, otherwise the line "The number..." would have been:

The number \(\hfoo\) is very small.  The product, \(\hfoo\hbar\) is

mathgrep --replace .orig 's/hbar/hfoo/g' quantum.tex

Produces no output, but quantum.tex is now the altered text above and quantum.tex.orig is the original.

mathgrep --preamble '/hbar/' quantum.tex

%---%
The number \(\hbar\) is very small.  The product, \(\hbar\hbar\) is
%---%
\[
\hbar = \frac{h}{2\pi}
\]
%---%
Some junk involving \(\hbar\).

mathgrep --postamble '/hbar/' quantum.tex

%---%
\newcommand{\texthbar}{\(\hbar\)}
%---%
The number \(\hbar\) is very small.  The product, \(\hbar\hbar\) is
%---%
\[
\hbar = \frac{h}{2\pi}
\]

mathgrep --noinline --nodisplay --environments myenv '/hbar/' quantum.tex

%---%
\begin{myenv}
Think of hbar as your friend.
\end{myenv}

NOTES

This program determines what is mathematics by looking for a matched pair; for example, $ and $. It does not check that the end of the pair is the correct one, just that it is the next one. Thus on encountering the text:

\(x = y \text{ if \(a = b\), otherwise } x = z\)

the program thinks that x = y \text{ if \(a = b is the next bit of mathematics. Consequently, it won't see the x = z part at all. This sort of thing only happens when the same type of mathematical environment is used inside and outside the \text command. On encountering the text:

\[x = y \text{ if \(a = b\), otherwise } x = z\]

then the whole of the inner part is considered as mathematics. Note that a byproduct of this is that the arguments of \text and \intertext are viewed as being in mathematics.

Another place where this might cause a problem is if you define an alias for a standard math environment:

\let\myendalign\endalign
\begin{align}
  x = y
\end{myalign}

then this program will continue slurping in input until it finds the next \end{align}.

Currently this program does not consider $ and $$ to have anything to do with mathematics. This is partly because the coding is a bit tricky (consider the text $x=y$$a=b$) but also because $ and $$ are not actually LaTeX commands but are inherited from TeX. Whilst $ is synonymous with \(, $$ and \[ are actually slightly different (take a look at latex.ltx if you don't believe me). Maybe if I get bored one day I'll take a look at implementing it.

BUGS

Whad'dya mean, "I've found a bug."? Impossible. There are no bugs, only features. To tell me about features, send an email to

loopspace@mathforge.org

With the following in the subject line:

Howay man, av fund ah borg een maathgrip.

Or open an issue on github at: https://github.com/loopspace/mathgrep

AUTHOR

Andrew Stacey

LICENSE

This program is free software; you can redistribute it and/or mod ify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

loopspace / mathgrep