Fix "video not available" being shown to the user when when YouTube start asking for captcha resolution due to receiving too many requests from the same IP. Show instead an appropiate message.
To be able to keep making requests, the captcha must be solved in a browser and the browser cookie must be passed to youtube-transcript-api.
This commit is contained in:
Your Name
parent
361986e221
commit
14c70359ba
|
|
@ -5,6 +5,7 @@ from ._errors import (
|
|||
NoTranscriptFound,
|
||||
CouldNotRetrieveTranscript,
|
||||
VideoUnavailable,
|
||||
TooManyRequests,
|
||||
NotTranslatable,
|
||||
TranslationLanguageNotAvailable,
|
||||
NoTranscriptAvailable,
|
||||
|
|
|
|||
|
|
@ -38,6 +38,9 @@ class CouldNotRetrieveTranscript(Exception):
|
|||
class VideoUnavailable(CouldNotRetrieveTranscript):
|
||||
CAUSE_MESSAGE = 'The video is no longer available'
|
||||
|
||||
class TooManyRequests(CouldNotRetrieveTranscript):
|
||||
CAUSE_MESSAGE = ('YouTube is receiving too many requests from this IP,'
|
||||
' and now requires that a captcha must be solved in order to continue.')
|
||||
|
||||
class TranscriptsDisabled(CouldNotRetrieveTranscript):
|
||||
CAUSE_MESSAGE = 'Subtitles are disabled for this video'
|
||||
|
|
|
|||
|
|
@ -14,6 +14,7 @@ import re
|
|||
from ._html_unescaping import unescape
|
||||
from ._errors import (
|
||||
VideoUnavailable,
|
||||
TooManyRequests,
|
||||
NoTranscriptFound,
|
||||
TranscriptsDisabled,
|
||||
NotTranslatable,
|
||||
|
|
@ -38,6 +39,8 @@ class TranscriptListFetcher():
|
|||
splitted_html = html.split('"captions":')
|
||||
|
||||
if len(splitted_html) <= 1:
|
||||
if 'class="g-recaptcha"' in html:
|
||||
raise TooManyRequests(video_id)
|
||||
if '"playabilityStatus":' not in html:
|
||||
raise VideoUnavailable(video_id)
|
||||
|
||||
|
|
|
|||
|
|
@ -0,0 +1,239 @@
|
|||
<!DOCTYPE html>
|
||||
<html>
|
||||
<head>
|
||||
<title>YouTube</title>
|
||||
<script
|
||||
src="https://www.google.com/recaptcha/api.js?hl=es"
|
||||
async
|
||||
defer
|
||||
nonce="upPiqA/fwe4T7vXFBXo5Gw"
|
||||
></script>
|
||||
<style nonce="upPiqA/fwe4T7vXFBXo5Gw">
|
||||
html {
|
||||
height: 100%;
|
||||
margin: 0;
|
||||
padding: 0;
|
||||
border: 0;
|
||||
font-size: 100%;
|
||||
background: transparent;
|
||||
}
|
||||
body {
|
||||
margin: 0;
|
||||
padding: 0;
|
||||
border: 0;
|
||||
font: 12px "YouTube Noto", Roboto, arial, sans-serif;
|
||||
background: #f1f1f1;
|
||||
height: 100%;
|
||||
}
|
||||
#captcha-page {
|
||||
text-align: center;
|
||||
height: 100%;
|
||||
white-space: nowrap;
|
||||
word-break: normal;
|
||||
}
|
||||
.g-recaptcha {
|
||||
display: inline-block;
|
||||
}
|
||||
#captcha-page-content {
|
||||
vertical-align: middle;
|
||||
display: inline-block;
|
||||
}
|
||||
#captcha-page-vertical-align {
|
||||
height: 100%;
|
||||
display: inline-block;
|
||||
vertical-align: middle;
|
||||
}
|
||||
p {
|
||||
margin: auto;
|
||||
color: #333;
|
||||
font-size: 16px;
|
||||
white-space: normal;
|
||||
text-shadow: 0 0 0 transparent, 0 1px 1px #fff;
|
||||
}
|
||||
</style>
|
||||
<link
|
||||
rel="shortcut icon"
|
||||
href="https://www.youtube.com/img/favicon.ico"
|
||||
type="image/x-icon"
|
||||
/>
|
||||
<link
|
||||
rel="icon"
|
||||
href="https://www.youtube.com/img/favicon_32.png"
|
||||
sizes="32x32"
|
||||
/>
|
||||
<link
|
||||
rel="icon"
|
||||
href="https://www.youtube.com/img/favicon_48.png"
|
||||
sizes="48x48"
|
||||
/>
|
||||
<link
|
||||
rel="icon"
|
||||
href="https://www.youtube.com/img/favicon_96.png"
|
||||
sizes="96x96"
|
||||
/>
|
||||
<link
|
||||
rel="icon"
|
||||
href="https://www.youtube.com/img/favicon_144.png"
|
||||
sizes="144x144"
|
||||
/>
|
||||
</head>
|
||||
<body>
|
||||
<div id="captcha-page">
|
||||
<div id="captcha-page-content">
|
||||
<p>
|
||||
Perdón por la interrupción. Hemos recibido un gran número de
|
||||
solicitudes de tu red.
|
||||
</p>
|
||||
<p>
|
||||
Para seguir disfrutando de YouTube, rellena el siguiente formulario.
|
||||
</p>
|
||||
<form action="/das_captcha?fw=1" method="POST">
|
||||
<div
|
||||
class="g-recaptcha"
|
||||
data-sitekey="6Lf39AMTAAAAALPbLZdcrWDa8Ygmgk_fmGmrlRog"
|
||||
></div>
|
||||
<br /><input
|
||||
type="hidden"
|
||||
name="action_recaptcha_verify2"
|
||||
value="1"
|
||||
/><input
|
||||
type="hidden"
|
||||
name="next"
|
||||
value="/watch?v=GJLlxj_dtq8"
|
||||
/><input type="submit" value="Submit" />
|
||||
</form>
|
||||
<style nonce="upPiqA/fwe4T7vXFBXo5Gw">
|
||||
#yt-masthead {
|
||||
margin: 15px auto;
|
||||
width: 440px;
|
||||
margin-top: 25px;
|
||||
}
|
||||
#logo-container {
|
||||
margin-right: 5px;
|
||||
float: left;
|
||||
cursor: pointer;
|
||||
text-decoration: none;
|
||||
}
|
||||
.logo {
|
||||
background: no-repeat
|
||||
url("//www.gstatic.com/youtube/img/branding/youtubelogo/1x/youtubelogo_30.png");
|
||||
width: 125px;
|
||||
height: 30px;
|
||||
cursor: pointer;
|
||||
display: inline-block;
|
||||
}
|
||||
#masthead-search {
|
||||
display: flex;
|
||||
margin-top: 3px;
|
||||
max-width: 650px;
|
||||
overflow: hidden;
|
||||
padding: 0;
|
||||
position: relative;
|
||||
}
|
||||
.search-button {
|
||||
border-left: 0;
|
||||
-moz-border-radius-topleft: 0;
|
||||
border-top-left-radius: 0;
|
||||
-moz-border-radius-bottomleft: 0;
|
||||
border-bottom-left-radius: 0;
|
||||
float: right;
|
||||
height: 29px;
|
||||
padding: 0;
|
||||
border: solid 1px transparent;
|
||||
border-color: #d3d3d3;
|
||||
background: #f8f8f8;
|
||||
color: #333;
|
||||
cursor: pointer;
|
||||
}
|
||||
.search-button:hover {
|
||||
border-color: #c6c6c6;
|
||||
background: #f0f0f0;
|
||||
box-shadow: 0 1px 0 rgba(0, 0, 0, 0.1);
|
||||
}
|
||||
.search-button-content {
|
||||
border: none;
|
||||
display: block;
|
||||
opacity: 0.6;
|
||||
padding: 0;
|
||||
text-indent: -10000px;
|
||||
background: no-repeat
|
||||
url(//www.gstatic.com/youtube/src/web/htdocs/img/search.png);
|
||||
background-size: auto;
|
||||
width: 15px;
|
||||
height: 15px;
|
||||
box-shadow: none;
|
||||
margin: 0 25px;
|
||||
}
|
||||
#masthead-search-terms-border {
|
||||
flex: 1 1 auto;
|
||||
border: 1px solid #ccc;
|
||||
box-shadow: inset 0 1px 2px #eee;
|
||||
background-color: #fff;
|
||||
font-size: 14px;
|
||||
height: 29px;
|
||||
line-height: 30px;
|
||||
margin: 0 0 2px;
|
||||
overflow: hidden;
|
||||
position: relative;
|
||||
-moz-box-sizing: border-box;
|
||||
box-sizing: border-box;
|
||||
-webkit-transition: border-color 0.2s ease;
|
||||
transition: border-color 0.2s ease;
|
||||
}
|
||||
#masthead-search-terms {
|
||||
background: transparent;
|
||||
border: 0;
|
||||
font-size: 16px;
|
||||
height: 100%;
|
||||
left: 0;
|
||||
margin: 0;
|
||||
outline: none;
|
||||
padding: 2px 6px;
|
||||
position: absolute;
|
||||
width: 100%;
|
||||
-moz-box-sizing: border-box;
|
||||
box-sizing: border-box;
|
||||
}
|
||||
</style>
|
||||
<div id="yt-masthead">
|
||||
<a
|
||||
id="logo-container"
|
||||
href="https://www.youtube.com/"
|
||||
title="Página de inicio de YouTube"
|
||||
><span class="logo" title="Página de inicio de YouTube"></span
|
||||
><span class="content-region">ES</span></a
|
||||
>
|
||||
<form
|
||||
id="masthead-search"
|
||||
class="search-form"
|
||||
action="https://www.youtube.com/results"
|
||||
onsubmit="if (document.getElementById('masthead-search-terms').value == '') return false;"
|
||||
>
|
||||
<div id="masthead-search-terms-border" dir="ltr">
|
||||
<input
|
||||
id="masthead-search-terms"
|
||||
autocomplete="off"
|
||||
onkeydown="if (!this.value && (event.keyCode == 40 || event.keyCode == 32 || event.keyCode == 34)) {this.onkeydown = null; this.blur();}"
|
||||
name="search_query"
|
||||
value=""
|
||||
type="text"
|
||||
placeholder="Buscar"
|
||||
title="Buscar"
|
||||
aria-label="Buscar"
|
||||
/>
|
||||
</div>
|
||||
<button
|
||||
class="search-button"
|
||||
type="submit"
|
||||
onclick="if (document.getElementById('masthead-search-terms').value == '') return false; document.getElementById('masthead-search').submit(); return false;"
|
||||
dir="ltr"
|
||||
>
|
||||
<span class="search-button-content">Buscar</span>
|
||||
</button>
|
||||
</form>
|
||||
</div>
|
||||
</div>
|
||||
<span id="captcha-page-vertical-align"></span>
|
||||
</div>
|
||||
</body>
|
||||
</html>
|
||||
|
|
@ -12,6 +12,7 @@ from youtube_transcript_api import (
|
|||
TranscriptsDisabled,
|
||||
NoTranscriptFound,
|
||||
VideoUnavailable,
|
||||
TooManyRequests,
|
||||
NoTranscriptAvailable,
|
||||
NotTranslatable,
|
||||
TranslationLanguageNotAvailable,
|
||||
|
|
@ -134,6 +135,16 @@ class TestYouTubeTranscriptApi(TestCase):
|
|||
with self.assertRaises(VideoUnavailable):
|
||||
YouTubeTranscriptApi.get_transcript('abc')
|
||||
|
||||
def test_get_transcript__exception_if_video_unavailable(self):
|
||||
httpretty.register_uri(
|
||||
httpretty.GET,
|
||||
'https://www.youtube.com/watch',
|
||||
body=load_asset('youtube_too_many_requests.html.static')
|
||||
)
|
||||
|
||||
with self.assertRaises(TooManyRequests):
|
||||
YouTubeTranscriptApi.get_transcript('abc')
|
||||
|
||||
def test_get_transcript__exception_if_transcripts_disabled(self):
|
||||
httpretty.register_uri(
|
||||
httpretty.GET,
|
||||
|
|
|
|||
Loading…
Reference in New Issue