witchard / mog

A different take on the UNIX tool cat

mog

A different take on the UNIX tool cat.

What is this?

The man page for cat says that it can: concatenate files and print on the standard output. Often (at least for me) its main use is for the latter... mog tries to help you "print on the standard output" in a more intelligent way. For example, it can be configured to:

• Syntax highlight scripts
• Print a hex dump of binary files
• Show details of image files
• Perform objdump on executables
• List a directory

Installation

The simplest way is to install via pip: [sudo] pip install mog.

Required dependencies are the file command (e.g. sudo apt install file).

If you plan on using the default configuration file, then you will also want poppler-utils, binutils and mediainfo installed (e.g. sudo apt install poppler-utils binutils mediainfo on a debian based machine). You'll also need pygmentize - this should be installed automatically with pip install mog, but some users have reported this not to be the case (e.g. issue #18) - if this happens a separate pip install pygments will hopefully do the trick.

@gregf has kindly provided an Arch Linux package: https://aur.archlinux.org/packages/mog-git/. There is also a FreeBSD port: https://www.freebsd.org/cgi/ports.cgi?query=%5Emog&stype=name&sektion=sysutils.

For the latest development version:

• [sudo] pip install git+https://github.com/witchard/mog
• Or clone this repository and then run [sudo] python ./setup.py install

Default Config

If you don't give mog a configuration file, it will use the defaults. Here is what you will get (prioritised in the order below - i.e. the first thing to match is done)

• File extension is recognised by pygments - Format with pygmentize (http://pygments.org/)
• File mime type is recognisd by pygments - Format with pygmentize (http://pygments.org/)
• File type is ELF - Parse with objdump -ft
• File is not a file (i.e. directory / symlink / fifo) - List it using ls -lh
• File is of a video or image mime type - Summarise it with mediainfo (http://mediainfo.sourceforge.net)
• File is a PDF document - Print it as text using pdftotext (https://poppler.freedesktop.org/)
• File is a tar archive - List contents of tar using tar --list
• File extension is .deb - Show information using dpkg -I
• File extension is .rpm - Show information using rpm -qip
• File extension is .csv - Format it using column -xt -s,
• File extension is .tsv - Format it using column -xt
• File contains ASCII text - Print using cat
• Anything else - Assumed to be binary, print using xxd

How does it work?

mog reads the $HOME/.mogrc config file which describes a series of operations it can do in an ordered manner. The config file can be overridden with the --config argument. Each operation has a match command and an action command. For each file you give to mog it will test each match command in turn, when one matches it will perform the action. A reasonably useful config file is generated when you first run it.

Matches

Currently the following match commands are supported:

• name=<regex> - Check if the file name matches the regex
• file=<regex> - Check if the output of file matches the regex
• file_mime=<regex> - Check if the output of file --mime matches the regex
• pygmentize=<regex> - Check if the output of pygmentize -N matches the regex
• pygmentsmime=<regex> - Check if the pygments lexer for the files mimetype matches the regex. Always failes when there is no pygments lexer for the specified mimetype

Note, one can specify invert_match, you can use this to cause a match when the regex does not match.

Actions

The following actions are supported:

• arg=<program> - Pass the file name as an argument to the program
• argreplace=<program> - Replace %F in <program> with the filename. Replace %N (where N is an integer) in <program> with matching capture group from match regex. Execute the result.

Config file format

The config file is an ini style format defined as follows:

[settings]
showname=yes
showsection=no
viewinless=no
toponly=no
toplines=10
followsymlinks=no
recursive=no

[name-of-match-action-1]
match=arg
action=arg
invert_match=boolean

[name-of-match-action-2]
match=arg
action=arg
invert_match=boolean

The settings section may contain the following:

• showname - default: True. Show the name of each file before performing the action.
• showsection - default: False. Show config file section where match occurred next to file name. showname must be True for this to work.
• viewinless - default: False. Output everything in a pager (less -S).
• toponly - default: False. Output only the top few lines of each file.
• toplines - default: 10. Number of lines to output in toponly mode.
• followsymlinks - default: False. Follow symbolic links when processing files.
• recursive - default: False. Recurse into directories to process the files within them.

The invert_match value is optional and will cause the match to be inverted - i.e. you can use this to cause a match when the regex does not match.

Matches and actions will be processed in the order found in the file.

It should be noted that mog uses the name of the script to determine what config file to read. So for example one can ln -s mog feline and then it would use the $HOME/.felinerc as the config file. This means you can have multiple configurations for different names.

mog's Issues

Do something useful with videos

Support macOS

Test on and fix any bugs when running on macOS

Is pygmentize a dependency?

Pipe output through less

Add a config option / command line arg to force all output to go into less.

Don't process symlinks that are broken

With recursive and followsymlink on its possible to try and process broken symlinks which doesn't work.

Add recursion for directories

Add support for recursion into directories to run on the files inside them

feature request: beautify *.csv or *.tsv files

Hi,

If at any point you have some time, I think it would be very nice to automatically view formatted files like tables with something like column -xt or so

Remove dependency on imagemagick

Use a python tool instead maybe? Then it can be installed automatically when we do #7.

What happens when nothing matches?

Decimal representation in '.sh' files fails

When a BASH script('.sh' file) that has a decimal number in its contents is rendered with the mog command, it tends to produce erroneous output where the mantissa part(numbers after the decimal point) of the number is a different number in itself and not as a whole with the abscissa.

Here is a link for the output.

match on pygmentize supported files

Use pygmentize -N to see if pygments can do highlighting for us.

Document supported python versions

Via setup.py and then a shield.

Support different config files

Command line option for providing a different config

Comand line arguments to override settings

Do something with PDFs

Install default rc file

If no rc file is found, install the example default.

TypeError when specifying regex which may return empty matches

Hi!

With this config in my ~/.mogrc;

[xml]
name=.*(?=(nuget\.config|NuGet\.Config|\.([fc]sproj|ruleset)))
argreplace=pygmentize -l xml -O style=solarized-light %F

I get the following output for files named like the following

• <folder path>/nuget.config
• and <folder path>/NuGet.Config
  (Don't ask my why the mix of capitalization, blame dotnet core on Linux =/).

All the while, this file matches nicely and displays correctly; <folder path>/<Project name>.csproj.

This is the error I get;

-> $ mog <failing Nuget xml file>
==> <failing Nuget xml file> <==
==> Error: "TypeError('replace() argument 2 must be str, not None',)" when processing file <failing Nuget xml file> <==

New match for file --mime

Matching on file --mime gives a more predictable output format.

Add support for following symlinks

By default symlinks are reported as such by file, with the default config this can lead to symlinks being run through ls due to them not matching on higher level things. Following them in the python will mean this doesn't happen. Note, we probably also want to lower the priority of the filesystem section in the default config.

Show the config section that matched

Show in the printed file name header which config section matched.

Do something more intelligent when an action fails

Currently we dont do anything!

Archlinux Package

Just a heads up I created a aur package for mog-git.

https://aur.archlinux.org/packages/mog-git/

Debug mode

Explain what things are being tried as it runs

Release to pip

Simplify the install...

Add command line argument to enable piping the output of each file through head

This will allow a quick mog-ified overview.

Document default config

Document what file types are supported with the default config

Support for zip files

Add some automated tests

Enable capture groups in the regexes to be used as arguments

E.g. to capture script type for syntax highlighting

Improve help

-h is a bit vague as it references config variables. make it clearer

Directory recursion depth

Ability to only recurse so far