declone-each-stdin0-file-stem-via-size

#!/bin/sh
set -euf -o pipefail
out () { printf %s\\n "$*" ; }

##
# Declone each stdin 0-delimited file stem within two directories.
#
# This script iterates on each file stem, and looks for the file stem
# within each of the two directories specified on the command line.
#
# If both directories contain a file with the same stem and same size,
# then the second directory's file is called a clone.
#
# We keep the file in the first directory as is, and delete the clone.
#
# This script is useful for deleting duplicate files,
# while also giving the user full control of which files
# are evaluated by using stdin with 0-delimited stems.
#
# This script skips symlinks.
#
# Example:
#
#     echo "a\0b\0c\0" | declone-each-stdin0-file-stem /originals /clones
#
# The example compares these pairs:
#
#     originals/a and clones/a
#     originals/b and clones/b
#     originals/c and clones/c
#
# Example to create file stems by using `find`:
#
#     find /foo/goo -type f -print0 | sed -z 's|/foo/goo/||' > stems
#
# The example does a typical find, then chops off the root directory;
# the example output is the file stems and is 0-delimited as we want.
#
# Contact: Joel Parker Henderson (joel@joelparkerhenderson.com)
# License: GPL
# Updated: 2015-07-13
##

a_root_path="$1"
b_root_path="$2"

file_size(){
  stat -f %z "$1"
}

file_digest(){
  shasum -a 512 "$1" | awk '{print $1}'
}

while read -d $'\0' stem_path; do
  a="$a_root_path/$stem_path"
  b="$b_root_path/$stem_path"
  if [ -f "$a" ] && [ -f "$b" ] && [ ! -L "$a" ] && [ ! -L "$b" ] && [ `file_size "$a"` == `file_size "$b"` ]; then
    out "⊤ $stem_path"
    rm "$b"
  else
    out "⊥ $stem_path"
  fi
done