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/fetcher/hls.py

324 lines
12 KiB
Python
Raw Normal View History

2013-03-02 21:26:28 +01:00
# ex:ts=4:sw=4:sts=4:et
# -*- tab-width: 4; c-basic-offset: 4; indent-tabs-mode: nil -*-
2019-08-25 00:40:39 +02:00
import binascii
import copy
2013-02-12 19:39:52 +01:00
import os
2018-01-07 20:52:19 +01:00
import time
2019-08-25 00:40:39 +02:00
from datetime import datetime
from datetime import timedelta
2019-03-24 21:04:41 +01:00
from cryptography.hazmat.backends import default_backend
2019-08-25 00:40:39 +02:00
from cryptography.hazmat.primitives.ciphers import algorithms
from cryptography.hazmat.primitives.ciphers import Cipher
from cryptography.hazmat.primitives.ciphers import modes
from svtplay_dl.error import ServiceError
from svtplay_dl.error import UIException
2014-04-21 16:50:24 +02:00
from svtplay_dl.fetcher import VideoRetriever
from svtplay_dl.fetcher.m3u8 import M3U8
from svtplay_dl.subtitle import subtitle_probe
2022-06-04 00:45:55 +02:00
from svtplay_dl.utils.fetcher import filter_files
2019-08-25 00:40:39 +02:00
from svtplay_dl.utils.http import get_full_url
from svtplay_dl.utils.output import ETA
from svtplay_dl.utils.output import formatname
2019-08-25 00:40:39 +02:00
from svtplay_dl.utils.output import progress_stream
from svtplay_dl.utils.output import progressbar
2014-04-21 16:50:24 +02:00
class HLSException(UIException):
def __init__(self, url, message):
self.url = url
2019-08-25 00:33:51 +02:00
super().__init__(message)
class LiveHLSException(HLSException):
def __init__(self, url):
2019-08-25 00:33:51 +02:00
super().__init__(url, "This is a live HLS stream, and they are not supported.")
2013-02-12 19:39:52 +01:00
def hlsparse(config, res, url, output, **kwargs):
if not res:
return
if res.status_code > 400:
yield ServiceError(f"Can't read HLS playlist. {res.status_code}")
return
yield from _hlsparse(config, res.text, url, output, cookies=res.cookies, **kwargs)
def _hlsparse(config, text, url, output, **kwargs):
m3u8 = M3U8(text)
keycookie = kwargs.pop("keycookie", None)
cookies = kwargs.pop("cookies", None)
authorization = kwargs.pop("authorization", None)
loutput = copy.copy(output)
loutput["ext"] = "ts"
2020-07-28 21:26:13 +02:00
channels = kwargs.pop("channels", None)
codec = kwargs.pop("codec", "h264")
2017-10-27 00:08:53 +02:00
media = {}
subtitles = {}
videos = {}
segments = None
2018-02-04 22:59:38 +01:00
if m3u8.master_playlist:
for i in m3u8.master_playlist:
audio_url = None
2020-07-28 21:26:13 +02:00
vcodec = None
chans = None
audio_group = None
language = ""
2021-04-18 14:06:25 +02:00
resolution = ""
if i["TAG"] == "EXT-X-MEDIA":
if i["TYPE"] and i["TYPE"] != "SUBTITLES":
if "URI" in i:
if segments is None:
segments = True
if i["GROUP-ID"] not in media:
media[i["GROUP-ID"]] = []
if "CHANNELS" in i:
if i["CHANNELS"] == "6":
chans = "51"
if "LANGUAGE" in i:
language = i["LANGUAGE"]
if "AUTOSELECT" in i and i["AUTOSELECT"].upper() == "YES":
role = "main"
else:
role = "alt"
media[i["GROUP-ID"]].append([i["URI"], chans, language, role])
else:
segments = False
if i["TYPE"] == "SUBTITLES":
if "URI" in i:
2023-10-09 00:09:27 +02:00
caption = None
if i["GROUP-ID"] not in subtitles:
subtitles[i["GROUP-ID"]] = []
if "LANGUAGE" in i:
lang = i["LANGUAGE"]
else:
lang = "und"
2023-10-09 00:09:27 +02:00
if "CHARACTERISTICS" in i:
caption = True
item = [i["URI"], lang, caption]
if item not in subtitles[i["GROUP-ID"]]:
subtitles[i["GROUP-ID"]].append(item)
continue
elif i["TAG"] == "EXT-X-STREAM-INF":
if "AVERAGE-BANDWIDTH" in i:
bit_rate = float(i["AVERAGE-BANDWIDTH"]) / 1000
else:
bit_rate = float(i["BANDWIDTH"]) / 1000
2021-04-18 14:06:25 +02:00
if "RESOLUTION" in i:
resolution = i["RESOLUTION"]
2020-07-28 21:26:13 +02:00
if "CODECS" in i:
if i["CODECS"][:3] == "hvc":
vcodec = "hevc"
if i["CODECS"][:3] == "avc":
vcodec = "h264"
if "AUDIO" in i:
audio_group = i["AUDIO"]
urls = get_full_url(i["URI"], url)
videos[bit_rate] = [urls, resolution, vcodec, audio_group]
else:
2018-01-30 20:11:37 +01:00
continue # Needs to be changed to utilise other tags.
for bit_rate in list(videos.keys()):
urls, resolution, vcodec, audio_group = videos[bit_rate]
if audio_group and media:
for group in media[audio_group]:
audio_url = get_full_url(group[0], url)
chans = group[1] if audio_url else channels
codec = vcodec if vcodec else codec
yield HLS(
copy.copy(config),
urls,
bit_rate,
cookies=cookies,
keycookie=keycookie,
authorization=authorization,
audio=audio_url,
output=loutput,
segments=bool(segments),
channels=chans,
codec=codec,
resolution=resolution,
language=group[2],
role=group[3],
**kwargs,
)
else:
chans = channels
codec = vcodec if vcodec else codec
yield HLS(
copy.copy(config),
urls,
bit_rate,
cookies=cookies,
keycookie=keycookie,
authorization=authorization,
audio=audio_url,
output=loutput,
segments=bool(segments),
channels=chans,
codec=codec,
resolution=resolution,
**kwargs,
)
if subtitles:
for sub in list(subtitles.keys()):
for n in subtitles[sub]:
2023-10-09 00:09:27 +02:00
subfix = n[2]
if len(subtitles[sub]) > 1:
if subfix:
subfix = f"{n[1]}-caption"
yield from subtitle_probe(
2023-10-09 00:09:27 +02:00
copy.copy(config), get_full_url(n[0], url), output=copy.copy(output), subfix=subfix, cookies=cookies, **kwargs
2019-08-25 00:27:31 +02:00
)
elif m3u8.media_segment:
2018-05-13 13:06:45 +02:00
config.set("segments", False)
yield HLS(
2020-12-26 13:10:56 +01:00
copy.copy(config),
url,
0,
cookies=cookies,
2020-12-26 13:10:56 +01:00
keycookie=keycookie,
authorization=authorization,
output=loutput,
2020-12-26 13:10:56 +01:00
segments=False,
2019-08-25 00:27:31 +02:00
)
else:
yield ServiceError("Can't find HLS playlist in m3u8 file.")
2014-04-21 21:42:49 +02:00
2015-09-15 20:10:32 +02:00
2014-04-21 21:55:39 +02:00
class HLS(VideoRetriever):
2018-05-25 22:47:26 +02:00
@property
def name(self):
return "hls"
2014-04-21 16:50:24 +02:00
def download(self):
self.output_extention = "ts"
2018-05-08 22:46:11 +02:00
if self.segments:
if self.audio and not self.config.get("only_video"):
# self._download(self.audio, file_name=(copy.copy(self.output), "audio.ts"))
self._download(self.audio, True)
if not self.audio or not self.config.get("only_audio"):
self._download(self.url)
2017-10-27 00:08:53 +02:00
else:
# Ignore audio
self.audio = None
self._download(self.url)
2017-10-27 00:08:53 +02:00
def _download(self, url, audio=False):
2018-03-10 11:40:36 +01:00
cookies = self.kwargs.get("cookies", None)
2018-01-07 20:52:19 +01:00
start_time = time.time()
2018-02-04 22:59:38 +01:00
m3u8 = M3U8(self.http.request("get", url, cookies=cookies).text)
2014-04-21 16:50:24 +02:00
key = None
def random_iv():
return os.urandom(16)
2019-08-25 00:27:31 +02:00
if audio:
self.output["ext"] = "audio.ts"
else:
self.output["ext"] = "ts"
filename = formatname(self.output, self.config)
file_d = open(filename, "wb")
2014-04-21 16:50:24 +02:00
2018-05-08 22:46:11 +02:00
hls_time_stamp = self.kwargs.pop("hls_time_stamp", False)
2022-06-03 06:02:34 +02:00
if self.kwargs.get("filter", False):
2022-06-04 00:45:55 +02:00
m3u8 = filter_files(m3u8)
decryptor = None
2018-01-07 20:52:19 +01:00
size_media = len(m3u8.media_segment)
eta = ETA(size_media)
total_duration = 0
2018-01-07 20:52:19 +01:00
duration = 0
max_duration = 0
for index, i in enumerate(m3u8.media_segment):
if "EXTINF" in i and "duration" in i["EXTINF"]:
duration = i["EXTINF"]["duration"]
max_duration = max(max_duration, duration)
total_duration += duration
item = get_full_url(i["URI"], url)
2014-04-21 16:50:24 +02:00
2018-05-08 22:46:11 +02:00
if not self.config.get("silent"):
if self.config.get("live"):
progressbar(size_media, index + 1, "".join(["DU: ", str(timedelta(seconds=int(total_duration)))]))
2018-01-07 20:52:19 +01:00
else:
eta.increment()
2019-08-25 00:27:31 +02:00
progressbar(size_media, index + 1, "".join(["ETA: ", str(eta)]))
2014-04-21 16:50:24 +02:00
headers = {}
if "EXT-X-BYTERANGE" in i:
headers["Range"] = f'bytes={i["EXT-X-BYTERANGE"]["o"]}-{i["EXT-X-BYTERANGE"]["o"] + i["EXT-X-BYTERANGE"]["n"] - 1}'
data = self.http.request("get", item, cookies=cookies, headers=headers)
if data.status_code == 404:
break
data = data.content
2021-03-10 14:11:49 +01:00
if m3u8.encrypted:
headers = {}
if self.keycookie:
keycookies = self.keycookie
else:
keycookies = cookies
if self.authorization:
headers["authorization"] = self.authorization
# Update key/decryptor
if "EXT-X-KEY" in i:
keyurl = get_full_url(i["EXT-X-KEY"]["URI"], url)
if keyurl and keyurl[:4] == "skd:":
raise HLSException(keyurl, "Can't decrypt beacuse of DRM")
key = self.http.request("get", keyurl, cookies=keycookies, headers=headers).content
iv = binascii.unhexlify(i["EXT-X-KEY"]["IV"][2:].zfill(32)) if "IV" in i["EXT-X-KEY"] else random_iv()
backend = default_backend()
cipher = Cipher(algorithms.AES(key), modes.CBC(iv), backend=backend)
decryptor = cipher.decryptor()
# In some cases the playlist say its encrypted but the files is not.
# This happen on svtplay 5.1ch stream where it started with ID3..
# Adding the other ones is header for mpeg-ts files. third byte is 10 or 11..
2021-03-10 14:11:49 +01:00
if data[:3] != b"ID3" and data[:3] != b"\x47\x40\x11" and data[:3] != b"\x47\x40\x10" and data[4:12] != b"ftypisom":
if decryptor:
data = decryptor.update(data)
else:
raise ValueError("No decryptor found for encrypted hls steam.")
2014-04-21 16:50:24 +02:00
file_d.write(data)
2013-02-12 19:39:52 +01:00
if self.config.get("capture_time") > 0 and total_duration >= self.config.get("capture_time") * 60:
2018-01-07 21:49:50 +01:00
break
2018-05-08 22:46:11 +02:00
if (size_media == (index + 1)) and self.config.get("live"):
sleep_int = (start_time + max_duration * 2) - time.time()
if sleep_int > 0:
time.sleep(sleep_int)
2018-01-07 20:52:19 +01:00
size_media_old = size_media
while size_media_old == size_media:
start_time = time.time()
2018-05-08 22:46:11 +02:00
if hls_time_stamp:
end_time_stamp = (datetime.utcnow() - timedelta(minutes=1, seconds=max_duration * 2)).replace(microsecond=0)
start_time_stamp = end_time_stamp - timedelta(minutes=1)
2018-02-04 22:59:38 +01:00
base_url = url.split(".m3u8")[0]
2021-02-28 22:05:15 +01:00
url = f"{base_url}.m3u8?in={start_time_stamp.isoformat()}&out={end_time_stamp.isoformat()}?"
2018-02-04 22:59:38 +01:00
new_m3u8 = M3U8(self.http.request("get", url, cookies=cookies).text)
for n_m3u in new_m3u8.media_segment:
if not any(d["URI"] == n_m3u["URI"] for d in m3u8.media_segment):
m3u8.media_segment.append(n_m3u)
2018-01-07 20:52:19 +01:00
size_media = len(m3u8.media_segment)
if size_media_old == size_media:
time.sleep(max_duration)
2018-01-07 20:52:19 +01:00
file_d.close()
2018-05-08 22:46:11 +02:00
if not self.config.get("silent"):
2019-08-25 00:27:31 +02:00
progress_stream.write("\n")
self.finished = True