From c2cd04758a144be27b99b4dd8fb37075baf470a4 Mon Sep 17 00:00:00 2001
From: Timothy Allen <tim@treehouse.org.za>
Date: Thu, 15 Oct 2015 16:22:03 +0200
Subject: [PATCH] Download Daily Show and Late Show

---
 download_daily_shows.sh | 108 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 108 insertions(+)
 create mode 100755 download_daily_shows.sh

diff --git a/download_daily_shows.sh b/download_daily_shows.sh
new file mode 100755
index 0000000..c25345f
--- /dev/null
+++ b/download_daily_shows.sh
@@ -0,0 +1,108 @@
+#/bin/bash
+
+# TODO videos seem to be downloading multiple times
+# TODO late show
+
+DOWNLOAD_DIR=/media/storage/shared/Videos
+YOUTUBE_DL="python3 /usr/local/bin/youtube-dl"
+#KEEP_DAYS=40
+KEEP_DAYS=100
+ARGS=-qwc
+
+OPTIND=1 # Reset is necessary if getopts was used previously
+while getopts ":u:d" opt; do
+	case "${opt}" in
+		d) 
+			ARGS=-vwc
+			set -x
+			;;
+		u) 
+			url="${OPTARG}"
+			;;
+		h) 
+			echo "Usage: ${0} <URL> <-d>"
+			exit 1
+			;;
+	esac
+done
+shift "$((OPTIND-1))" # Shift off the options and optional --.
+
+# Formats: 
+# python3 /usr/local/bin/youtube-dl -F http://www.cc.com/shows/the-daily-show-with-trevor-noah/full-episodes/
+# [info] Available formats for b2a5ae64-3795-4914-b3a7-644eed83372f:
+# format code  extension  resolution note
+# 200          mp4        384x216    
+# 400          mp4        384x216    
+# 750          mp4        512x288    
+# 1200         mp4        640x360    
+# 1700         mp4        768x432    
+# 2200         mp4        960x540    
+# 3500         mp4        1280x720   (best)
+#FORMAT="vhttp-750"
+#FORMAT="rtmp-750"
+#FORMAT="rtmp-3500/rtmp2200"
+
+function download_show {
+	url="$1"
+	show="$2"
+
+	mkdir -p "${DOWNLOAD_DIR}/${show}/"
+
+	# Get a chosen video format for each site
+	if [[ "${url}" =~ cc.com ]]; then
+		FORMAT=1200 
+	elif [[ "${url}" =~ cbs.com ]]; then
+		FORMAT="rtmp-496-0"
+	fi
+
+	# Try and download twice (resume if partially downloaded)
+	for i in 1 2; do
+		${YOUTUBE_DL} ${ARGS} -f ${FORMAT} \
+			-o "${DOWNLOAD_DIR}/${show}/%(upload_date)s/%(title)s-%(id)s.%(ext)s" \
+			--download-archive "${DOWNLOAD_DIR}/${show}.archive" \
+			${url}
+
+		# Wait 10 minutes before trying again
+		#sleep 600
+	done
+
+	# Move files from ${DOWNLOAD_DIR}/The Daily Show/NA/ to ${DOWNLOAD_DIR}/The Daily Show/$date
+	while IFS=  read -r -d '' file; do 
+		dir=$( stat -c %y "$file" | sed -e 's/\([0-9]*\)-\([0-9]*\)-\([0-9]*\).*/\1\2\3/' );
+		mkdir -p "${DOWNLOAD_DIR}/${show}/${dir}/"
+		# Keep a copy in /NA/ to avoid re-downloading
+		cp -n "$file" "${DOWNLOAD_DIR}/${show}/${dir}/"
+	done < <(find "${DOWNLOAD_DIR}/${show}/NA" -type f -print0)
+	# Remove old copies in /NA/, since we don't need to
+	# worry about re-downloading them any more
+	find "${DOWNLOAD_DIR}/${show}/NA/" -mindepth 1 -mtime +7 -delete
+
+	# Remove old shows
+	find "${DOWNLOAD_DIR}/${show}/" -mindepth 1 -mtime +${KEEP_DAYS} -delete
+}
+
+
+if [[ -n ${url} ]]; then
+	if [[ "${url}" =~ cc.com ]]; then
+		show="The Daily Show"
+	elif [[ "${url}" =~ cbs.com ]]; then
+		show="The Late Show"
+	fi
+	download_show "${url}" "${show}"
+else
+	for show in "The Daily Show" "The Late Show"; do 
+		url=
+		if [[ "${show}" =~ Daily ]]; then
+			url=http://www.cc.com/shows/the-daily-show-with-trevor-noah/full-episodes/
+		elif [[ "${show}" =~ Late ]]; then
+			# Get URL from http://www.cbs.com/shows/the-late-show-with-stephen-colbert/video/
+			# Requires Unotelly DNS:
+			# 197.242.94.51
+			# 197.189.212.164
+			url=$( wget -qO- http://www.cbs.com/shows/the-late-show-with-stephen-colbert/video/ | grep -E "/video/[^/]+/" | grep '"url":' | perl -pe 's/.*?"url":"(.*?)".*/$1/ims' | head -1 )
+		fi
+		download_show "${url}" "${show}"
+	done
+fi
+	
+