unixify

Updated 7/8/2010.

unixify.sh is a simple shell script I use to sanitize files people send me, often from other operating systems.

It cleans up file names by removing most punctuation and converting spaces to underscores and upper case to lower case. For example, it converts Foo, Bar, & Baz (Special).JPEG to foo_bar_and_baz_special.jpg.

It also re-compresses images with ImageMagick, removes carriage returns from DOS and Windows style CRLF line endings, and converts Microsoft Word documents to plain text with antiword.

It supports dry runs via the -n flag, which just prints what would be done instead of actually doing it.

Here’s the script:

#!/bin/bash

# parse args
while getopts "n" options; do
  case $options in
    n ) DRYRUN="-n";;
    * ) echo 'Usage: unixify.sh FILES...'
        exit 1;;
  esac
done

# getops updates OPTIND to point to the arg it stopped at. shift $@ to that point.
shift $((OPTIND-1))

for file in "$@"; do
  # clean filename. (careful with the quoting and line break continuations!)
  newfile=`rename -v $DRYRUN \
      "y/ /_/;
      s/[!?':\",\[\]()#]//g; "'
      s/&/and/g;
      y/A-Z/a-z/;
      s/\.\.\./_/g;
      s/_+/_/g;
      s/_(\.[^.]+$)/$1/;
      s/\.jpeg$/.jpg/' \
      "$file"`

  if [[ $DRYRUN != "" ]]; then
    if [[ $newfile != "" ]]; then
      echo $newfile
    fi
    continue
  fi

  if [[ $newfile =~ ' renamed as ' ]]; then
    file=${newfile/* renamed as /}
  fi

  if [[ $file =~ \.(gif|jpg|png)$ ]]; then
    # optimize image
    convert $file $file
  elif [[ $file =~ \.doc$ ]]; then
    # convert word doc to text
    antiword $file > `basename $file .doc`.txt
  fi

  if [[ `file -b $file` =~ text,.*with\ CRLF ]]; then
    # remove any carriage returns
    sed --in-place 's/\r//g' $file
  fi
done