Shell script to get N most words given text input and desired number of words. Redirected input and output verbosity supported.

n_most_words.sh

#!/usr/bin/env sh

# script flags
VERBOSE=true

# default inputs
DEFAULT_TEXT="The quick brown fox jumps over the lazy dog. Yes, the quick brown fox."
DEFAULT_COUNT=10

# check if input is redirected
if [ ! -t 0 ]; then
  VERBOSE=${2:-false}                         # do not echo piped text unless requested
  TEXT=$(cat)                                 # piped input from stdin
  COUNT=$(expr 0 + ${1:-} 2>/dev/null || :;)  # null string if first paramter is not a number
else
  # parse command line inputs
  TEXT=${1:-$DEFAULT_TEXT}                    # default text if first parameter is missing
  COUNT=$(expr 0 + ${2:-} 2>/dev/null || :;)  # null string if second paramter is not a number
  VERBOSE=${3:-$VERBOSE}                      # update verbose flag given third parameter
fi
COUNT=${COUNT:-$DEFAULT_COUNT}                # default value if count is unset or invalid

# verbose output
if [ "true" = "${VERBOSE}" ]; then
  echo "TEXT = \"${TEXT}\""
  echo "COUNT = ${COUNT}"
  echo ""
  echo "RESULT"
fi

# process output top words by frequency and alphabetical order
echo "${TEXT}" |                 `# initial TEXT` \
  tr '[:upper:]' '[:lower:]' |   `# use lower case` \
  tr -cd '[:alnum:][:blank:]' |  `# strip non-alphanumerics` \
  tr -s '[:space:]' |            `# squash excess whitespace` \
  tr -c '[:alnum:]' '[\n*]' |    `# put each word on a line` \
  sort |                         `# sort words to facilitate counting` \
  uniq -ci |                     `# get unique counts` \
  sort -k1,1nr -k2 |             `# frequency order, alpha order on tie` \
  head -"${COUNT}"               `# take only requested COUNT results`