Youtube channels to RSS

download.py

#!/usr/bin/python2.7

import json
import urllib
import pafy
import pytz
import os
import hashlib
from datetime import datetime
from feedgen.feed import FeedGenerator

ROOT = "https://82.237.11.61/yt/"
RSS_FILENAME = "rss.xml"
MAX_ITEMS = 20
QUIET = False

chans = json.load(open("chans.json"))
urls = []
downloaded = json.load(open("downloads.json"))

for chan in chans :
    feed = json.load(urllib.urlopen("https://gdata.youtube.com/feeds/base/users/" + chan + "/uploads?alt=json"))
    
    for vid in feed["feed"]["entry"] :
        for link in vid["link"] :
            if link["rel"] == "alternate" and link["type"] == "text/html" :
                urls.append({'url' : link["href"], 'published' : vid["published"]["$t"], 'updated' : vid["updated"]["$t"], 'id':vid["id"]["$t"]})

last_update = ""

urls.sort(key = lambda x : int(datetime.strptime(x["published"], '%Y-%m-%dT%H:%M:%S.%fZ').strftime('%s')), reverse=True)

while len(urls) > MAX_ITEMS : urls.pop()

for url in urls :
    videoid = hashlib.md5(url["id"]).hexdigest()
    filename = videoid + ".mp4"
    if not videoid in downloaded or not os.path.isfile(filename) :
        video = pafy.new(url["url"])
        best = video.getbest(preftype="mp4")
        downloaded[videoid] = {'filename': filename, 'title':video.title, 'description':video.description, 'thumb':video.thumb, 'published':url["published"], 'updated':url["updated"], 'id' : videoid, 'link':url["url"], 'author':video.author, 'category':video.category, 'duration': video.duration}
        print ("Downloading " + video.title)
        best.download(quiet=QUIET, filepath = filename)
        print()
        json.dump(downloaded, open("downloads.json", mode="w"))

downloaded_sorted = sorted(downloaded.iteritems(), key=lambda (k,v) :int(datetime.strptime(v["published"], '%Y-%m-%dT%H:%M:%S.%fZ').strftime('%s')), reverse=True)

while len(downloaded_sorted) > MAX_ITEMS :
    to_del = downloaded_sorted.pop()
    os.remove(downloaded[to_del[0]]["filename"])
    del downloaded[to_del[0]]

json.dump(downloaded, open("downloads.json", mode="w"))



fg = FeedGenerator()
fg.title('Youtube Subscriptions')
fg.id(ROOT)
fg.language('en')
fg.updated(datetime.now(tz=pytz.utc))
fg.link({'href':ROOT + RSS_FILENAME, 'rel':'self'})
fg.load_extension('podcast')
fg.description("Youtube subscriptions")

for download in downloaded :
    fe = fg.add_entry()
    fe.id(downloaded[download]["link"])
    fe.title(downloaded[download]["title"])
    fe.description(downloaded[download]["description"])
    fe.published(downloaded[download]["published"])
    fe.updated(downloaded[download]["updated"])
    fe.link({'href':downloaded[download]["link"], 'rel':'alternate', 'type':'text/html'})
    fe.author({'name':downloaded[download]["author"]})
    fe.category({'term':downloaded[download]["category"]})
    fe.enclosure(ROOT + downloaded[download]["filename"], type = "video/mp4", length = os.path.getsize(downloaded[download]["filename"]))
    fe.podcast.itunes_image(downloaded[download]["thumb"])
    fe.podcast.itunes_duration(downloaded[download]["duration"])
    
fg.rss_file(RSS_FILENAME)