#!/bin/sh

#RJ Trenchard 100281657

# Prints a message describing program usage and exit 3 if no argument is given
# Takes a list of arguments representing files
# Tries to compress each file using 4 different compression utilities: gzip, bzip2, zip, 7z
# Use the following file extensions:
# .gz : gzip
# .bz2 : bzip2
# .zip : zip
# .7z : 7z
# Keeps whatever version is smallest, and deletes the other (including the original)
# If the original is smallest, keeps it.
# The kept file should be in the same directory as the original was
# Prints a message saying which file it's keeping
# Does everything in parallel. All of the input files are processed simultaneously. All 4 compressions are done for each file simultaneously.
# Files and paths may contain spaces
# You may assume that there are no naming collisions
# HINT: this as not as complicated as it seems. Do some actions, compare results, choose best result. Repeated for each input.
# HINT: make it work in serial first
# HINT: use the compression programs as filters with file redirection. eg: 7z < fin > fout
# HINT: you way want to use basename and dirname, but it's easier without them
# HINT: you may want to use stat. Dont. Just read the man page for ls.

if [ $# -eq 0 ]; then
    echo "Usage: enter filenames separated by spaces."
else
    for file in "$@"; do
        if [ -f "$file" ]; then
            #quietly compress the file in 4 formats simasldifjeously
            (
                #I don't get 7zip's command line at all
                #and I don't understand it quietly!
                7z a "$file.7Zip" "$file" >/dev/null &  pid_of_7z=$!
                zip <"$file" >"$file.zip" &             pid_of_zip=$!
                gzip <"$file" >"$file.gz" &             pid_of_gz=$!
                bzip2 <"$file" >"$file.bz2" &           pid_of_bz=$!

                #wait on subprocesses to finish, otherwise we'll have a 0byte file
                wait $pid_of_7z $pid_of_zip $pid_of_gz $pid_of_bz

                #list in a single column all the files we just created, sort by size with -S, show only the largest 4 out of 5, and then remove them.
                ls -1 -S --file-type "$file" "$file.7Zip" "$file.zip" "$file.gz" "$file.bz2" | head -n 4 | while read line; do
                    rm $line
                done

                #lazy
                echo "keeping `ls -1 --file-type \"$file\" \"$file.7Zip\" \"$file.zip\" \"$file.gz\" \"$file.bz2\"`"

            ) 2>/dev/null &

            #delete the biggest files
        else
            echo "$file is not a valid file" >&2
        fi
    done
fi

#wait until everything is done for a clean shell.
wait