Mastodon cache files by byte count and extension

cache_by_ext_bytes.sh

#!/bin/sh

# Probably needs GNU AWK. Come to think of it, some OSes
# won't have a `find` with -printf like GNU does.
# Usage:
# ./cache_by_ext_bytes.sh /path/to/mastodon/system/cache

cache_dir=${1-/opt/mastodon/web/system/cache}

# This will handle spaces in filenames, but not newlines in filenames…
find "$cache_dir" -type f -printf "%s %p\n" \
    | awk '{
f_bytes=$1; $1 = "" # Nuke byte count
f_path=$0 # Everything else, so paths with spaces are included
ext_pos=match($f_path, /\.[^\.]+$/)
f_ext=substr($f_path, ext_pos+1) # +1 to skip over . itself

# Consider .jpeg to be .jpg.
if (f_ext == "jpeg") { f_ext="jpg" }

byte_count[f_ext]+=f_bytes
total_bytes+=f_bytes
}
END {
    for (key in byte_count) {
        printf "%-7s %12lu %14s\n",
            key, byte_count[key],
            "(" sprintf("%.1f", byte_count[key] / 1024 / 1024) "MiB)" \
        | "sort -k 2 -rn"
    }
    close("sort -k 2 -rn")
    print "-----"
    printf "Total: %13lu %14s\n",
        total_bytes, "(" sprintf("%.1f", total_bytes / 1024 / 1024) "MiB)"
}'