apalognorm

#!/bin/sh
 
######################################################################
#
# APALOGNORM : a normalizer for logs of the Apache combined format
#
# Written by Rich Mikan (richmikan[at]richlab.org) at 2017/01/31
#
# Usage  : apalognorm [-s string] <logfile>
#          <logfile> should be written with Apache combine format.
#          -s  Set the substitute string you want to convert from
#              the space character. (default is "_")
#
# * This is a filter convert every space character to the other one.
#   It is very useful from the various Unix command to treat.
# * e.g. the following command give you the referer field.
#   $ apalognorm http-access.log | awk '{print $8}'
#
######################################################################
 
# definition: print the usage and exit
print_usage_and_exit () {
  cat <<-USAGE 1>&2
	Usage   : ${0##*/} [-s string] <logfile>
	          <logfile> should be written with Apache combine format.
	          -s  Set the alternative character you want convert from
	              the space character. (default is "_")
	Version : Mon May  5 11:28:22 JST 2014
	USAGE
  exit 1
}

# initialization
PATH="$(command -p getconf PATH):${PATH:-}"

# parse the arguments
s_opt='_'
file=''
i=0
optmode=''
for arg in "$@"; do
  i=$((i+1))
  if [ -z "$optmode" ]; then
    case "$arg" in
      -s*)
        ret=$(echo "_${arg#-s}" | sed '1s/^_//')
        if [ -n "$ret" ]; then
          s_opt=$ret
        else
          optmode='s'
        fi
        ;;
      *)
        if [ -z "$file" ]; then
          [ $i -eq $# ] || print_usage_and_exit
          file=$arg
        else
          print_usage_and_exit
        fi
        ;;
    esac
  elif [ "$optmode" = 's' ]; then
    s_opt=$arg
    optmode=''
  else
    print_usage_and_exit
  fi
done
if [ ! -f "$file"     ] &&
   [ ! -c "$file"     ] &&
   [ ! -p "$file"     ] &&
   [ "_$file" != '_-' ] &&
   [ ! -z "$file"     ]; then
  echo "${0##*/}: No such file found" 1>&2
  print_usage_and_exit
elif [ -z "$file" ]; then
  file='-'
fi
 
 
# convert the substitute string for setting to the sed command
sub=$(printf '%s\n' "$s_opt"   |
      sed 's/\([\&/]\)/\\\1/g' )
 
# Define some marks for converting
RS=$(printf '\036')             # a mark for the real new lines
LF=$(printf '\\\n_');LF=${LF%_} # LF for the sed command
 
# Convert the file
if [ "_$file" = '_-' ]; then                      #
  sed 's/^\(.*\)$/\1'"$RS"'/' 2>/dev/null         #
else                                              #
  sed 's/^\(.*\)$/\1'"$RS"'/' "$file" 2>/dev/null #
fi                                                |
sed 's/"\([^"]*\)"/'"$LF"'"\1"'"$LF"'/g'          |
sed 's/\[\([^]]*\)\]/'"$LF"'[\1]'"$LF"'/g'        |
sed '/^["[]/s/[[:blank:]]/'"$sub"'/g'             |
tr -d '\n'                                        |
tr "$RS" '\n'