Commit 3e9ce263 authored by Alexandre Flament's avatar Alexandre Flament Committed by GitHub

Merge branch 'master' into patch-1

parents cef77f0d ca09d910
......@@ -8,7 +8,8 @@
# @stable no
# @parse url, title, content, publishedDate, thumbnail, embedded
from lxml import html
from functools import reduce
from json import loads
from searx.engines.xpath import extract_text
from searx.utils import list_get
from searx.url_utils import quote_plus
......@@ -34,20 +35,6 @@ embedded_url = '<iframe width="540" height="304" ' +\
base_youtube_url = 'https://www.youtube.com/watch?v='
# specific xpath variables
results_xpath = "//ol/li/div[contains(@class, 'yt-lockup yt-lockup-tile yt-lockup-video vve-check')]"
url_xpath = './/h3/a/@href'
title_xpath = './/div[@class="yt-lockup-content"]/h3/a'
content_xpath = './/div[@class="yt-lockup-content"]/div[@class="yt-lockup-description yt-ui-ellipsis yt-ui-ellipsis-2"]'
# returns extract_text on the first result selected by the xpath or None
def extract_text_from_dom(result, xpath):
r = result.xpath(xpath)
if len(r) > 0:
return extract_text(r[0])
return None
# do search-request
def request(query, params):
......@@ -63,18 +50,29 @@ def request(query, params):
def response(resp):
results = []
dom = html.fromstring(resp.text)
results_data = resp.text[resp.text.find('ytInitialData'):]
results_data = results_data[results_data.find('{'):results_data.find(';\n')]
# parse results
for result in dom.xpath(results_xpath):
videoid = list_get(result.xpath('@data-context-item-id'), 0)
results_json = loads(results_data) if results_data else {}
sections = results_json.get('contents', {})\
.get('twoColumnSearchResultsRenderer', {})\
.get('primaryContents', {})\
.get('sectionListRenderer', {})\
.get('contents', [])
for section in sections:
for video_container in section.get('itemSectionRenderer', {}).get('contents', []):
video = video_container.get('videoRenderer', {})
videoid = video.get('videoId')
if videoid is not None:
url = base_youtube_url + videoid
thumbnail = 'https://i.ytimg.com/vi/' + videoid + '/hqdefault.jpg'
title = extract_text_from_dom(result, title_xpath) or videoid
content = extract_text_from_dom(result, content_xpath)
title = video.get('title', {}).get('simpleText', videoid)
description_snippet = video.get('descriptionSnippet', {})
if 'runs' in description_snippet:
content = reduce(lambda a, b: a + b.get('text', ''), description_snippet.get('runs'), '')
else:
content = description_snippet.get('simpleText', '')
embedded = embedded_url.format(videoid=videoid)
# append result
......
......@@ -113,17 +113,14 @@ class RawTextQuery(object):
parse_next = True
engine_name = engine_shortcuts[prefix]
if engine_name in engines:
for engine_category in engines[engine_name].categories:
self.engines.append({'category': engine_category,
self.engines.append({'category': 'none',
'name': engine_name,
'from_bang': True})
# check if prefix is equal with engine name
elif prefix in engines:
parse_next = True
if prefix in engines:
for engine_category in engines[prefix].categories:
self.engines.append({'category': engine_category,
self.engines.append({'category': 'none',
'name': prefix,
'from_bang': True})
......
......@@ -46,121 +46,71 @@ class TestYoutubeNoAPIEngine(SearxTestCase):
self.assertEqual(youtube_noapi.response(response), [])
html = """
<ol id="item-section-063864" class="item-section">
<li>
<div class="yt-lockup yt-lockup-tile yt-lockup-video vve-check clearfix yt-uix-tile"
data-context-item-id="DIVZCPfAOeM"
data-visibility-tracking="CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JECx_-GK5uqMpcIB">
<div class="yt-lockup-dismissable"><div class="yt-lockup-thumbnail contains-addto">
<a aria-hidden="true" href="/watch?v=DIVZCPfAOeM" class=" yt-uix-sessionlink pf-link"
data-sessionlink="itct=CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JFIEdGVzdA">
<div class="yt-thumb video-thumb"><img src="//i.ytimg.com/vi/DIVZCPfAOeM/mqdefault.jpg"
width="196" height="110"/></div><span class="video-time" aria-hidden="true">11:35</span></a>
<span class="thumb-menu dark-overflow-action-menu video-actions">
</span>
</div>
<div class="yt-lockup-content">
<h3 class="yt-lockup-title">
<a href="/watch?v=DIVZCPfAOeM"
class="yt-uix-tile-link yt-ui-ellipsis yt-ui-ellipsis-2 yt-uix-sessionlink spf-link"
data-sessionlink="itct=CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JFIEdGVzdA"
title="Top Speed Test Kawasaki Ninja H2 (Thailand) By. MEHAY SUPERBIKE"
aria-describedby="description-id-259079" rel="spf-prefetch" dir="ltr">
Title
</a>
<span class="accessible-description" id="description-id-259079"> - Durée : 11:35.</span>
</h3>
<div class="yt-lockup-byline">de
<a href="/user/mheejapan" class=" yt-uix-sessionlink spf-link g-hovercard"
data-sessionlink="itct=CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JA" data-ytid="UCzEesu54Hjs0uRKmpy66qeA"
data-name="">MEHAY SUPERBIKE</a></div><div class="yt-lockup-meta">
<ul class="yt-lockup-meta-info">
<li>il y a 20 heures</li>
<li>8 424 vues</li>
</ul>
</div>
<div class="yt-lockup-description yt-ui-ellipsis yt-ui-ellipsis-2" dir="ltr">
Description
</div>
<div class="yt-lockup-badges">
<ul class="yt-badge-list ">
<li class="yt-badge-item" >
<span class="yt-badge">Nouveauté</span>
</li>
<li class="yt-badge-item" ><span class="yt-badge " >HD</span></li>
</ul>
</div>
<div class="yt-lockup-action-menu yt-uix-menu-container">
<div class="yt-uix-menu yt-uix-videoactionmenu hide-until-delayloaded"
data-video-id="DIVZCPfAOeM" data-menu-content-id="yt-uix-videoactionmenu-menu">
</div>
</div>
</div>
</div>
</div>
</li>
</ol>
<div></div>
<script>
window["ytInitialData"] = {
"contents": {
"twoColumnSearchResultsRenderer": {
"primaryContents": {
"sectionListRenderer": {
"contents": [
{
"itemSectionRenderer": {
"contents": [
{
"videoRenderer": {
"videoId": "DIVZCPfAOeM",
"title": {
"simpleText": "Title"
},
"descriptionSnippet": {
"runs": [
{
"text": "Des"
},
{
"text": "cription"
}
]
}
}
},
{
"videoRenderer": {
"videoId": "9C_HReR_McQ",
"title": {
"simpleText": "Title"
},
"descriptionSnippet": {
"simpleText": "Description"
}
}
}
]
}
}
]
}
}
}
}
};
</script>
"""
response = mock.Mock(text=html)
results = youtube_noapi.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 1)
self.assertEqual(len(results), 2)
self.assertEqual(results[0]['title'], 'Title')
self.assertEqual(results[0]['url'], 'https://www.youtube.com/watch?v=DIVZCPfAOeM')
self.assertEqual(results[0]['content'], 'Description')
self.assertEqual(results[0]['thumbnail'], 'https://i.ytimg.com/vi/DIVZCPfAOeM/hqdefault.jpg')
self.assertTrue('DIVZCPfAOeM' in results[0]['embedded'])
html = """
<ol id="item-section-063864" class="item-section">
<li>
<div class="yt-lockup yt-lockup-tile yt-lockup-video vve-check clearfix yt-uix-tile"
data-context-item-id="DIVZCPfAOeM"
data-visibility-tracking="CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JECx_-GK5uqMpcIB">
<div class="yt-lockup-dismissable"><div class="yt-lockup-thumbnail contains-addto">
<a aria-hidden="true" href="/watch?v=DIVZCPfAOeM" class=" yt-uix-sessionlink pf-link"
data-sessionlink="itct=CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JFIEdGVzdA">
<div class="yt-thumb video-thumb"><img src="//i.ytimg.com/vi/DIVZCPfAOeM/mqdefault.jpg"
width="196" height="110"/></div><span class="video-time" aria-hidden="true">11:35</span></a>
<span class="thumb-menu dark-overflow-action-menu video-actions">
</span>
</div>
<div class="yt-lockup-content">
<h3 class="yt-lockup-title">
<span class="accessible-description" id="description-id-259079"> - Durée : 11:35.</span>
</h3>
<div class="yt-lockup-byline">de
<a href="/user/mheejapan" class=" yt-uix-sessionlink spf-link g-hovercard"
data-sessionlink="itct=CBgQ3DAYACITCPGXnYau6sUCFZEIHAod-VQASCj0JA" data-ytid="UCzEesu54Hjs0uRKmpy66qeA"
data-name="">MEHAY SUPERBIKE</a></div><div class="yt-lockup-meta">
<ul class="yt-lockup-meta-info">
<li>il y a 20 heures</li>
<li>8 424 vues</li>
</ul>
</div>
<div class="yt-lockup-badges">
<ul class="yt-badge-list ">
<li class="yt-badge-item" >
<span class="yt-badge">Nouveauté</span>
</li>
<li class="yt-badge-item" ><span class="yt-badge " >HD</span></li>
</ul>
</div>
<div class="yt-lockup-action-menu yt-uix-menu-container">
<div class="yt-uix-menu yt-uix-videoactionmenu hide-until-delayloaded"
data-video-id="DIVZCPfAOeM" data-menu-content-id="yt-uix-videoactionmenu-menu">
</div>
</div>
</div>
</div>
</div>
</li>
</ol>
"""
response = mock.Mock(text=html)
results = youtube_noapi.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 1)
self.assertEqual(results[1]['title'], 'Title')
self.assertEqual(results[1]['url'], 'https://www.youtube.com/watch?v=9C_HReR_McQ')
self.assertEqual(results[1]['content'], 'Description')
self.assertEqual(results[1]['thumbnail'], 'https://i.ytimg.com/vi/9C_HReR_McQ/hqdefault.jpg')
self.assertTrue('9C_HReR_McQ' in results[1]['embedded'])
html = """
<ol id="item-section-063864" class="item-section">
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment