#!/bin/bash

# Download files one-by-one of a single directory from GitHub

# Note: If repo is not that large, this will be much slower than the whole repo tarball.

# Usage: github_download.sh URL_TO_DIRECTORY

url=$1

user=$( cut -d/ -f4 <<< $url )

repo=$( cut -d/ -f5 <<< $url )

branch=$( cut -d/ -f7 <<< $url )

path=$( cut -d/ -f8- <<< $url )

urlcontent=https://api.github.com/repos/$user/$repo/contents/$path?ref=$branch

subdir=$( basename $path )

dirtarget=$PWD

fileGet() {

dirtarget+=/$1

mkdir -p $dirtarget

lines=$( curl -s $urlcontent )

types=$( grep '"type":' <<< $lines | cut -d'"' -f4 )

names=( $( grep '"name":' <<< $lines | cut -d'"' -f4 ) )

readarray -t download_url <<< $( grep '"download_url":' <<< $lines | cut -d'"' -f4 )

dirs=

i=0

for t in $types; do

name=${names[i]}

if [[ $t == file ]]; then

echo File: $name

curl -LO ${download_url[i]} --output-dir $dirtarget

else

dirs+="$name "

fi

(( i++ ))

done

if [[ $dirs ]]; then

for d in $dirs; do

echo Subdirectory: $d

urlcontent=${urlcontent/\?*}/$d?ref=$branch

fileGet $d

done

fi

}

fileGet $subdir

echo "

Download successfully to: $subdir

"