1
|
|
|
#!/usr/bin/python3 |
2
|
|
|
|
3
|
|
|
import re |
4
|
|
|
import abc |
5
|
|
|
import subprocess |
6
|
|
|
from time import sleep |
7
|
|
|
|
8
|
|
|
|
9
|
|
|
class YoutubeDownloader: |
10
|
|
|
|
11
|
|
|
_cmd = 'youtube-dl --extract-audio --audio-quality 0 --audio-format mp3 -o "%(title)s.%(ext)s" "{url}"' |
12
|
|
|
_args = ['youtube-dl', '--extract-audio', '--audio-quality', '0', '--audio-format', 'mp3', '-o', '%(title)s.%(ext)s'] |
13
|
|
|
_capture_stdout = {True: {'stdout': subprocess.PIPE}, # captures std out ie: print(str(ro.stdout, encoding='utf-8')) |
14
|
|
|
False: {}} # does not capture stdout and it streams it in the terminal} |
15
|
|
|
|
16
|
|
|
update_command_args = ('sudo', 'pip3', 'install', '--upgrade', 'youtube-dl') |
17
|
|
|
update_backend_command = ' '.join(update_command_args) |
18
|
|
|
|
19
|
|
|
@classmethod |
20
|
|
|
def download_times(cls, video_url, directory, times=10, spawn=True, verbose=True, supress_stdout=False, suppress_certificate_validation=False, delay=1): |
21
|
|
|
i = 0 |
22
|
|
|
while i < times - 1: |
23
|
|
|
try: |
24
|
|
|
cls.download(video_url, directory, spawn=spawn, verbose=verbose, supress_stdout=supress_stdout, suppress_certificate_validation=suppress_certificate_validation) |
25
|
|
|
return |
26
|
|
|
except TooManyRequestsError as e: |
27
|
|
|
print(e) |
28
|
|
|
i += 1 |
29
|
|
|
sleep(delay) |
30
|
|
|
cls.download(video_url, directory, spawn=spawn, verbose=verbose, supress_stdout=supress_stdout, suppress_certificate_validation=suppress_certificate_validation) |
31
|
|
|
|
32
|
|
|
@classmethod |
33
|
|
|
def download(cls, video_url, directory, spawn=True, verbose=True, supress_stdout=False, suppress_certificate_validation=False): |
34
|
|
|
""" |
35
|
|
|
Downloads a video from youtube given a url, converts it to mp3 and stores in input directory.\n |
36
|
|
|
:param str video_url: |
37
|
|
|
:param str directory: |
38
|
|
|
:param bool spawn: |
39
|
|
|
:param bool verbose: |
40
|
|
|
:param bool supress_stdout: |
41
|
|
|
:param bool suppress_certificate_validation: |
42
|
|
|
:return: |
43
|
|
|
""" |
44
|
|
|
cls.__download(video_url, directory, spawn=spawn, verbose=verbose, supress_stdout=supress_stdout, suppress_certificate_validation=suppress_certificate_validation) |
45
|
|
|
|
46
|
|
|
@classmethod |
47
|
|
|
def __download(cls, video_url, directory, **kwargs): |
48
|
|
|
args = cls._args[:-1] + ['{}/{}'.format(directory, cls._args[-1])] + [video_url] |
49
|
|
|
# If suppress HTTPS certificate validation |
50
|
|
|
if kwargs.get('suppress_certificate_validation', False): |
51
|
|
|
args.insert(1, '--no-check-certificate') |
52
|
|
|
if kwargs.get('verbose', False): |
53
|
|
|
print("Downloading stream '{}' and converting to mp3 ..".format(video_url)) |
54
|
|
|
if kwargs.get('spawn', True): |
55
|
|
|
_ = subprocess.Popen(args, stderr=subprocess.STDOUT, **cls._capture_stdout[kwargs.get('supress_stdout', True)]) |
56
|
|
|
else: |
57
|
|
|
ro = subprocess.run(args, stderr=subprocess.PIPE, **cls._capture_stdout[kwargs.get('supress_stdout', True)]) |
58
|
|
|
if ro.returncode != 0: |
59
|
|
|
stderr = str(ro.stderr, encoding='utf-8') |
60
|
|
|
raise YoutubeDownloaderErrorFactory.create_from_stderr(stderr, video_url) |
61
|
|
|
|
62
|
|
|
@classmethod |
63
|
|
|
def update_backend(cls): |
64
|
|
|
ro = subprocess.run(cls.update_command_args, stderr=subprocess.PIPE, **cls._capture_stdout[False]) |
65
|
|
|
return ro |
66
|
|
|
|
67
|
|
|
@classmethod |
68
|
|
|
def cmd(cls, video_url): |
69
|
|
|
return cls._cmd.format(url=video_url) |
70
|
|
|
|
71
|
|
|
|
72
|
|
|
class YoutubeDownloaderErrorFactory: |
73
|
|
|
@staticmethod |
74
|
|
|
def create_with_message(msg): |
75
|
|
|
return Exception(msg) |
76
|
|
|
|
77
|
|
|
@staticmethod |
78
|
|
|
def create_from_stderr(stderror, video_url): |
79
|
|
|
exception_classes = (TokenParameterNotInVideoInfoError, InvalidUrlError, UnavailableVideoError, TooManyRequestsError, CertificateVerificationError) |
80
|
|
|
for subclass in exception_classes: |
81
|
|
|
if subclass.reg.search(stderror): |
82
|
|
|
return subclass(video_url, stderror) |
83
|
|
|
s = "NOTE: None of the predesinged exceptions' regexs [{}] matched. Perhaps you want to derive a new subclass from AbstractYoutubeDownloaderError to account for this youtube-dl exception with string to parse <S>{}</S>'".format(', '.join(['"{}"'.format(_.reg) for _ in exception_classes]), stderror) |
84
|
|
|
return Exception(AbstractYoutubeDownloaderError(video_url, stderror)._msg + '\n' + s) |
85
|
|
|
|
86
|
|
|
|
87
|
|
|
#### EXCEPTIONS |
88
|
|
|
|
89
|
|
|
class AbstractYoutubeDownloaderError(metaclass=abc.ABCMeta): |
90
|
|
|
def __init__(self, *args, **kwargs): |
91
|
|
|
super().__init__() |
92
|
|
|
if len(args) > 1: |
93
|
|
|
self.video_url = args[0] |
94
|
|
|
self.stderr = args[1] |
95
|
|
|
elif len(args) > 0: |
96
|
|
|
self.video_url = args[0] |
97
|
|
|
self._msg = "YoutubeDownloader generic error." |
98
|
|
|
self._short_msg = kwargs.get('msg') |
99
|
|
|
if args or 'msg' in kwargs: |
100
|
|
|
self._msg = '\n'.join([_ for _ in [kwargs.get('msg', ''), getattr(self, 'stderr', '')] if _]) |
101
|
|
|
|
102
|
|
|
|
103
|
|
|
class TokenParameterNotInVideoInfoError(Exception, AbstractYoutubeDownloaderError): |
104
|
|
|
"""Token error""" |
105
|
|
|
reg = re.compile('"token" parameter not in video info for unknown reason') |
106
|
|
|
|
107
|
|
|
def __init__(self, video_url, stderror): |
108
|
|
|
AbstractYoutubeDownloaderError.__init__(self, video_url, stderror) |
109
|
|
|
Exception.__init__(self, self._msg) |
110
|
|
|
|
111
|
|
|
class InvalidUrlError(Exception, AbstractYoutubeDownloaderError): |
112
|
|
|
"""Invalid url error""" |
113
|
|
|
reg = re.compile(r'is not a valid URL\.') |
114
|
|
|
|
115
|
|
|
def __init__(self, video_url, stderror): |
116
|
|
|
AbstractYoutubeDownloaderError.__init__(self, video_url, stderror, msg="Invalid url '{}'.".format(video_url)) |
117
|
|
|
Exception.__init__(self, self._short_msg) |
118
|
|
|
|
119
|
|
|
class UnavailableVideoError(Exception, AbstractYoutubeDownloaderError): |
120
|
|
|
"""Wrong url error""" |
121
|
|
|
reg = re.compile(r'ERROR: This video is unavailable\.') |
122
|
|
|
|
123
|
|
|
def __init__(self, video_url, stderror): |
124
|
|
|
AbstractYoutubeDownloaderError.__init__(self, video_url, stderror, msg="Unavailable video at '{}'.".format(video_url)) |
125
|
|
|
Exception.__init__(self, self._msg) |
126
|
|
|
|
127
|
|
|
class TooManyRequestsError(Exception, AbstractYoutubeDownloaderError): |
128
|
|
|
"""Too many requests (for youtube) to serve""" |
129
|
|
|
reg = re.compile(r"(?:ERROR: Unable to download webpage: HTTP Error 429: Too Many Requests|WARNING: unable to download video info webpage: HTTP Error 429)") |
130
|
|
|
|
131
|
|
|
def __init__(self, video_url, stderror): |
132
|
|
|
AbstractYoutubeDownloaderError.__init__(self, video_url, stderror, msg="Too many requests for youtube at the moment.".format(video_url)) |
133
|
|
|
Exception.__init__(self, self._msg) |
134
|
|
|
|
135
|
|
|
class CertificateVerificationError(Exception, AbstractYoutubeDownloaderError): |
136
|
|
|
"""This can happen when downloading is requested from a server like scrutinizer.io\n |
137
|
|
|
ERROR: Unable to download webpage: <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get |
138
|
|
|
local issuer certificate (_ssl.c:1056)> (caused by URLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] |
139
|
|
|
certificate verify failed: unable to get local issuer certificate (_ssl.c:1056)'))) |
140
|
|
|
""" |
141
|
|
|
reg = re.compile(r"ERROR: Unable to download webpage: <urlopen error \[SSL: CERTIFICATE_VERIFY_FAILED\]") |
142
|
|
|
|
143
|
|
|
def __init__(self, video_url, stderror): |
144
|
|
|
AbstractYoutubeDownloaderError.__init__(self, video_url, stderror, |
145
|
|
|
msg="Unable to download webpage because ssl certificate verification failed:\n[SSL: CERTIFICATE_VERIFY_FAILED] certificate " |
146
|
|
|
"verify failed: unable to get local issuer certificate (_ssl.c:1056)> (caused by " |
147
|
|
|
"URLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: " |
148
|
|
|
"unable to get local issuer certificate (_ssl.c:1056)')))") |
149
|
|
|
Exception.__init__(self, self._msg) |
150
|
|
|
|