git-find-large-files

#!/usr/bin/env bash
#
# Print the largest files in a Git repository. The script must be called
# from the root of the Git repository. You can pass a threshold to print
# only files greater than a certain size (compressed size in Git database,
# default is 500kb).
#
# Files that have a large compressed size should usually be stored in
# Git LFS [2].
#
# Based on script from Antony Stubbs [1] and improved with ideas from Peff.
#
# [1] http://stubbisms.wordpress.com/2009/07/10/git-script-to-show-largest-pack-objects-and-trim-your-waist-line/
# [2] https://git-lfs.github.com/
#
# Usage:
# git-find-large-files [size threshold in KB]
#

if [ -z "$1" ]; then
    MIN_SIZE_IN_KB=500
else
    MIN_SIZE_IN_KB=$1
fi

# Use "look" if it is available, otherwise use "grep" (e.g. on Windows)
if look >/dev/null 2>&1; then
    # On Debian the "-b" is available and required to make "look" perform
    # a binary search (see https://unix.stackexchange.com/a/499312/275508 ).
    if look 2>&1 | grep -q .-b; then
        search="look -b"
    else
        search=look
    fi
else
    search=grep
fi

# set the internal field separator to line break,
# so that we can iterate easily over the verify-pack output
IFS=$'\n';

# list all objects including their size, sort by compressed size
OBJECTS=$(
    git cat-file \
        --batch-all-objects \
        --batch-check='%(objectsize:disk) %(objectname)' \
    | sort -nr
)

TMP_DIR=$(mktemp -d "${TMPDIR:-/tmp}/git-find-large-files.XXXXXX") || exit
trap "rm -rf '$TMP_DIR'" EXIT

git rev-list --all --objects | sort > "$TMP_DIR/objects"
git rev-list --all --objects --max-count=1 | sort > "$TMP_DIR/objects.1"

for OBJ in $OBJECTS; do
    # extract the compressed size in kilobytes
    COMPRESSED_SIZE=$(($(echo $OBJ | cut -f 1 -d ' ')/1024))

    if [ $COMPRESSED_SIZE -le $MIN_SIZE_IN_KB ]; then
        break
    fi

    # extract the SHA
    SHA=$(echo $OBJ | cut -f 2 -d ' ')

    # find the objects location in the repository tree
    LOCATION=$($search $SHA "$TMP_DIR/objects" | sed "s/$SHA //")
    if $search $SHA "$TMP_DIR/objects.1" >/dev/null; then
        # Object is in the head revision
        HEAD="Present"
    elif [ -e $LOCATION ]; then
        # Objects path is in the head revision
        HEAD="Changed"
    else
        # Object nor its path is in the head revision
        HEAD="Deleted"
    fi

    echo "$COMPRESSED_SIZE,$HEAD,$LOCATION" >> "$TMP_DIR/output"
done

if [ -f "$TMP_DIR/output" ]; then
    column -t -s ',' < "$TMP_DIR/output"
fi

rm -rf "$TMP_DIR"
exit 0