1
0
mirror of https://github.com/spaam/svtplay-dl.git synced 2024-11-24 12:15:40 +01:00
svtplay-dl/lib/svtplay_dl/service/urplay.py

73 lines
2.9 KiB
Python

# ex:ts=4:sw=4:sts=4:et
# -*- tab-width: 4; c-basic-offset: 4; indent-tabs-mode: nil -*-
import copy
import json
import logging
import re
from html import unescape
from svtplay_dl.error import ServiceError
from svtplay_dl.fetcher.hls import hlsparse
from svtplay_dl.service import OpenGraphThumbMixin
from svtplay_dl.service import Service
from svtplay_dl.subtitle import subtitle
class Urplay(Service, OpenGraphThumbMixin):
supported_domains = ["urplay.se", "ur.se", "betaplay.ur.se", "urskola.se"]
def get(self):
match = re.search(r'/Player/Player" data-react-props="([^\"]+)\"', self.get_urldata())
if not match:
yield ServiceError("Can't find json info")
return
data = unescape(match.group(1))
jsondata = json.loads(data)
res = self.http.get("https://streaming-loadbalancer.ur.se/loadbalancer.json")
loadbalancer = res.json()["redirect"]
for streaminfo in jsondata["currentProduct"]["streamingInfo"].keys():
stream = jsondata["currentProduct"]["streamingInfo"][streaminfo]
if streaminfo == "raw":
url = "https://{}/{}playlist.m3u8".format(loadbalancer, stream["sd"]["location"])
streams = hlsparse(self.config, self.http.request("get", url), url, output=self.output)
for n in list(streams.keys()):
yield streams[n]
url = "https://{}/{}playlist.m3u8".format(loadbalancer, stream["hd"]["location"])
streams = hlsparse(self.config, self.http.request("get", url), url, output=self.output)
for n in list(streams.keys()):
yield streams[n]
if not (self.config.get("get_all_subtitles")) and (stream["default"]):
yield subtitle(copy.copy(self.config), "tt", stream["tt"]["location"], output=self.output)
if self.config.get("get_all_subtitles") and "tt" in stream:
label = stream["tt"]["language"]
if stream["tt"]["scope"] != "complete":
label = "{}-{}".format(label, stream["tt"]["scope"])
yield subtitle(copy.copy(self.config), "tt", stream["tt"]["location"], label, output=self.output)
def find_all_episodes(self, config):
episodes = []
match = re.search(r'/Player/Player" data-react-props="([^\"]+)\"', self.get_urldata())
if not match:
logging.error("Can't find json info")
return
data = unescape(match.group(1))
jsondata = json.loads(data)
for episode in jsondata["accessibleEpisodes"]:
episodes.append("https://urplay.se/program/{}".format(episode["slug"]))
episodes_new = []
n = 0
for i in episodes:
if n == config.get("all_last"):
break
if i not in episodes_new:
episodes_new.append(i)
n += 1
return episodes_new