Word-frequency calculator. Can be used standalone or in pipelines.

wf.sh

#!/usr/bin/env bash

# Word-frequency.
# Usage:
#
#  $ wf [-ni] file.txt [file1.txt file2.txt ...]
#
# You can choose to display just the first [n] most frequent words.
#
#  $ wf -n2 file.txt
#
# The -i switch enables case insensitivity.
#
# It can be used to count simple files:
#
#  $ wf -n2 -i file.txt
#
# or in pipelines:
#
#  $ find -name '*.rb' | xargs cat | wf -i
#

fail() {
  echo "$1" >&2
  exit 1
}

usage() {
  progname="`basename $0`"
  fail "Usage: $progname [-ni] file1.txt [file2.txt file3.txt ...]"
}

yellow() {
  echo -e "\033[1;33m$1\033[0m"
}

process_stdin() {
  while read line; do
    echo "$line" >>$tmpfile
  done

  process
}

process_files() {
  for file in $1; do
    yellow "$file"

    cp "$file" $tmpfile

    if [[ -n $downcase ]]; then
      tr A-Z a-z <"$file" >$tmpfile
    fi

    process
  done
}

process() {
  sed -e "$sed_src" "$tmpfile" | sort | uniq -c | sort -k1nr | head -n$limit
  echo
}

progname=`basename "$0"`
limit=-0

while getopts 'n:i' option; do
  case $option in
    n) 
      let "limit = $OPTARG"
    ;;

    i) 
      let "downcase = y"
    ;;
  esac
done

shift $((OPTIND - 1))

# Remove punctuations, extra white spaces, and insert line breaks between words
sed_src='
  s/[[:punct:]]/ /g
  s/	/ /g
  s/^  *//
  s/  *$//
  s/  */ /g
  /^ *$/d
  s/ /\n/g
'

tmpfile="/tmp/__wf_$$"

if [[ $# = 0 ]]; then
  process_stdin
else
  process_files $*
fi

rm $tmpfile