#!/bin/bash
#set -x 


# The original genesis for this came from
# http://stubbisms.wordpress.com/2009/07/10/git-script-to-show-largest-pack-objects-and-trim-your-waist-line/ Antony Stubbs

# Modified to use numfmt and only call rev-list once by Don Armstrong

# list all objects with size first; sort, and take the top 10.

NUM="${1:-10}"

ALL_OBJECTS=`git rev-list --all --objects`
OBJECTS=`git cat-file --batch-check='%(objectsize:disk) %(objectsize) %(objectname) %(objecttype)' --batch-all-objects|grep blob| sort -nr | head -n $NUM`
IFS=$'\n'
OUTPUT="Size,Packed,SHA1,Name";
for y in $OBJECTS
do
	# the size is in the 5th field; convert to IEC
	SIZE=$(echo "$y" | cut -f 2 -d ' '|numfmt --to=iec)
	# the compressed size is in the 6th field; convert to IEC
	COMPRESSED_SIZE=$(echo "$y" | cut -f 1 -d ' '|numfmt --to=iec)
	# extract the SHA
	SHA=`echo $y | cut -f 3 -d ' '`
	# find the objects location in the repository tree
	OTHER=`echo "$ALL_OBJECTS" | grep "$SHA"|cut -f2 -d ' '`
	OUTPUT="${OUTPUT}\n${SIZE},${COMPRESSED_SIZE},${SHA},${OTHER}"
done

echo -e $OUTPUT | column -t -s ', '