dirt.sh :: directory tree -- display tree size structure of directory hierarchy.

dirt.sh

#!/usr/bin/env bash
#              bash 4.2.24(1)     Linux Ubuntu 12.04          Date : 2013-12-23
#
# _______________|  dirt : display tree size structure of directory hierarchy.
#
#           Usage:  dirt  [directory=.]  [indent_character=" "]
#
#    Dependencies:  None
#                   [ Note: full-featured "tree" is an available Ubuntu package:
#                   http://manpages.ubuntu.com/manpages/trusty/man1/tree.1.html  
#                      "dirt" does not depend on a graphical hack like "tree" 
#                      instead provides an easily modifiable outline view.      ]
#  
#         Example:     $ dirt /tmp 
#                      /tmp  :848K
#                           hsperfdata_root  :32K
#                           .ICE-unix  :0
#                           .org.chromium.Chromium.YSC7c2  :272K
#                           plugtmp  :0
#                           pulse-XE2kzDXxubH6  :4.0K
#                           sni-qt_vlc_18126-2DtYmT  :4.0K
#                                icons  :4.0K
#                                     hicolor  :4.0K
#                                          128x128  :4.0K
#                                               apps  :16K
#                           ssh-RDwdlTlB1686  :0
#                           .X11-unix  :0


#  CHANGE LOG  LATEST version available:   https://bitbucket.org/rsvp/gists/src
#  2013-12-23  Add highlight to indentation area. Colorize size info.
#  2013-12-22  Use ; in 2nd sed phrase to avoid escaping /.
#                 Add individual directory sizes. Add pager for output.
#                 Improve error handling using $errf file.
#  2013-12-21  Refine the code. Use spaces for default identation.
#                 Show dot directories.
#  2013-12-20  First version based on one-liner script v2.3 by Dem Pilafian in
#                 http://www.centerkey.com/tree/


#           _____ PREAMBLE_v2: settings, variables, and error handling.
#
LC_ALL=POSIX
#      locale means "ASCII, US English, no special rules, 
#      output per ISO and RFC standards." 
#      Esp. use ASCII encoding for glob and sorting characters. 
shopt -s   extglob
#     ^set extended glob for pattern matching.
shopt -s   failglob
#         ^failed pattern matching signals error.
set -e
#   ^errors checked: immediate exit if a command has non-zero status. 
set -u
#   ^unassigned variables shall be errors.
#    Example of default VARIABLE ASSIGNMENT:  arg1=${1:-'foo'}

arg1=${1:-'.'}
#          ^= current directory as default
idchar=${2:-' '}
#            ^character used for identation.
pager=${PAGER:='less --LONG-PROMPT -i -J -W -c -h1 -y1 -~'}
#              ^favorite way for long outputs (plus search within).
#               Additional options added in DISPLAY section.


program=${0##*/}   #  similar to using basename
memf=$( mktemp /dev/shm/88_${program}_tmp.XXXXXXXXXX )
errf=$( mktemp /dev/shm/88_${program}_tmp.XXXXXXXXXX )


cleanup () {
     #  Delete temporary files, then optionally exit given status.
     local status=${1:-'0'}
     rm -f  $memf  $errf
     [ $status = '-1' ] ||  exit $status      #  thus -1 prevents exit.
} #--------------------------------------------------------------------
warn () {
     #  Message with basename to stderr.          Usage: warn "message"
     echo -e "\n !!  ${program}: $1 "  >&2
} #--------------------------------------------------------------------
die () {
     #  Exit with status of most recent command or custom status, after
     #  cleanup and warn.      Usage: command || die "message" [status]
     local status=${2:-"$?"}
     cleanup -1  &&   warn "$1"  &&  exit $status
} #--------------------------------------------------------------------
trap "die 'SIG disruption, but cleanup finished.' 114" 1 2 3 15
#    Cleanup after INTERRUPT: 1=SIGHUP, 2=SIGINT, 3=SIGQUIT, 15=SIGTERM
#
# _______________     ::  BEGIN  Script ::::::::::::::::::::::::::::::::::::::::


#  Establish the working directory:
cd "$arg1" 2> /dev/null  ||  die "Invalid directory argument: $arg1" 113

#  Then check if there is at least one level of sub-directories:
[ $( ls -F -1 | grep '/' | wc -l )  -eq  0 ]   \
     &&  die "NO sub-directories under $(pwd)"  115


#  MAIN: parse the necessary info into a file.
#  Identify directories which will have colon at the end of the line;
#  dot directories are included using ls -A, 
#  also parse -lh for total human sizes.
ls -AlhR  2> $errf  |  egrep ':$|^total '     \
     |  sed -e '/^\./N;s/\n//'                \
            -e 's;[^/]*/;*****;g'             \
            -e "s/\*/$idchar/g"               \
            -e 's/:total /  :/'               \
            -e "s;^\. ;$(pwd) ;"              >  $memf
     #  1st sed: joins the directory line with total size line.
     #  2nd sed: replace any lower level dir names with five stars.
     #              This is where IDENTATION gets introduced.
     #              Hopefully no directories are named using "*" :-)
     #  3rd sed: replace those stars with specified identation character.
     #  4th sed: beautify line from first sed to indicate size by : marker.
     #  5th sed: replace . with explicit present working directory.
     #     Note: some sed commands use ; instead of usual / as delimiter
     #              to avoid the pain of escaping forward slashes.
     #              Last sed is impossible without this trick.


#              Properly output that main file.
#  [ -t <fd> ] checks if the descriptor <fd> is a terminal or not:
if [ -t  1 ] ; then
     #      COLORIZE SIZE info.
     grep --color=always '  :.*$' $memf  |  $pager -p "^$idchar*" -R 
     #        pattern search HIGHLIGHTS INDENTATION^
     #                          retains colorization for size info^
     #  Output gets piped to pager for a nice DISPLAY.   
     #  TIP: use /M$|G$ within pager to search for large files.
else
     cat $memf
     #   On redirection or piping, we DO NOT want 
     #   ANSI escape codes (for colors) embedded in text output.
fi


#                              ERROR FILE from ls concludes output:
sed -e "s/^ls/ !!  $program/"  $errf  >&2
#  Any errors are shown at this stage since pager might have obscured them.
#  They will usually be Permission Denied types.


cleanup
# _______________ EOS ::  END of Script ::::::::::::::::::::::::::::::::::::::::


#             ANSI escape codes for color can be removed by escoff:
#             sed -e 's/\x1B\[[0-9;]*[JKmsu]//g' 


#  # _______________ 2013-12-21 Sat 19:54  
#       ls -AR  |  grep ":$"                                         \
#            |  sed -e 's/:$//'          -e 's;[^/]*/;*****;g'       \
#                   -e "s/\*/$idchar/g"  -e '/^\.$/d'
#            #  1st sed: remove colon identifers.
#            #  2nd sed: replace any lower level dir names with five stars.
#            #           This is where identation gets introduced.
#            #           Hopefully no directories are named using "*" :-)
#            #  3rd sed: replace those stars with specified identation character.
#            #  4th sed: delete that unnecessary line with single dot.


#  2013-12-20  Retrieved one-liner script v2.3 by Dem Pilafian in
#                 http://www.centerkey.com/tree/
#
#  Quick, what does the following Unix/Linux command do?
#  
#  ls -R | grep ":$" | sed -e 's/:$//' -e 's/[^-][^\/]*\//--/g' -e 's/^/   /' -e 's/-/|/'
#  
#  If you said, "Well, that's obvious; it shows a graphical representation of the
#  current sub-directories," you'd be correct.


#  vim: set fileencoding=utf-8 ff=unix tw=78 ai syn=sh :