kamp/searxng
1
0
Fork 0
mirror of https://github.com/searxng/searxng.git synced 2024-11-23 12:31:25 +01:00
Code

Merge pull request #1186 from kvch/fix-bing-videos

Browse Source 
Fix Bing videos engine
This commit is contained in:
Adam Tauber 2018-02-11 01:01:01 +01:00 committed by GitHub
commit 360f8fab97
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 40 additions and 85 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
searx/engines/bing_videos.py
View File

@ -69,22 +69,11 @@ def response(resp):
dom = html.fromstring(resp.text) dom = html.fromstring(resp.text)
for result in dom.xpath('//div[@class="dg_u"]'): for result in dom.xpath('//div[@class="dg_u"]'):
url = result.xpath('./div[@class="mc_vtvc"]/a/@href')[0]
# try to extract the url url = 'https://bing.com' + url
url_container = result.xpath('.//div[@class="sa_wrapper"]/@data-eventpayload') title = extract_text(result.xpath('./div/a/div/div[@class="mc_vtvc_title"]/@title'))
if len(url_container) > 0: content = extract_text(result.xpath('./div/a/div/div/div/div/text()'))
url = loads(url_container[0])['purl'] thumbnail = result.xpath('./div/a/div/div/img/@src')[0]
else:
url = result.xpath('./a/@href')[0]
# discard results that do not return an external url
# very recent results sometimes don't return the video's url
if url.startswith('/videos/search?'):
continue
title = extract_text(result.xpath('./a//div[@class="tl"]'))
content = extract_text(result.xpath('.//div[@class="pubInfo"]'))
thumbnail = result.xpath('.//div[@class="vthumb"]/img/@src')[0]
results.append({'url': url, results.append({'url': url,
'title': title, 'title': title,
@ -92,7 +81,6 @@ def response(resp):
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'template': 'videos.html'}) 'template': 'videos.html'})
# first page ignores requested number of results
if len(results) >= number_of_results: if len(results) >= number_of_results:
break break

103
tests/unit/engines/test_bing_videos.py
View File

@ -47,87 +47,54 @@ class TestBingVideosEngine(SearxTestCase):
self.assertEqual(bing_videos.response(response), []) self.assertEqual(bing_videos.response(response), [])
html = """ html = """
<div> <div class="dg_u">
<div class="dg_u"> <div id="mc_vtvc_1" class="mc_vtvc">
<a class="dv_i" href="/videos/search?abcde"> <a class="mc_vtvc_link" href="/video">
<div class="vthblock"> <div class="mc_vtvc_th">
<div class="vthumb"> <div class="cico">
<img src="thumb_1.jpg" /> <img src="thumb_1.jpg" />
</div> </div>
<div> <div class="mc_vtvc_ban_lo">
<div class="tl"> <div class="vtbc">
Title 1 <div class="mc_bc_w b_smText">
<div class="mc_bc pivot bpi_2">
<span title="">
<span class="mv_vtvc_play cipg "></span>
</span>
</div>
<div class="mc_bc items">10:06</div>
</div>
</div> </div>
</div> </div>
</div> </div>
<div class="videoInfoPanel"> <div class="mc_vtvc_meta">
<div class="pubInfo"> <div class="mc_vtvc_title" title="Title 1"></div>
<div>Content 1</div> <div class="mc_vtvc_meta_block_area">
<div class="mc_vtvc_meta_block">
<div class="mc_vtvc_meta_row">
<span>65,696,000+ views</span>
<span>1 year ago</span>
</div>
<div class="mc_vtvc_meta_row mc_vtvc_meta_row_channel">Content 1</div>
<div class="mc_vtvc_meta_row"><span>
<div class="cico mc_vtvc_src_ico">
<div></div>
</div>
<span>YouTube</span>
</span></div>
</div>
</div> </div>
</div> </div>
</a> <div class="vrhdata"></div>
<div class="sa_wrapper" </a>
data-eventpayload="{&quot;purl&quot;: &quot;https://url.com/1&quot;}">
</div> </div>
</div> </div>
</div>
""" """
response = mock.Mock(text=html) response = mock.Mock(text=html)
results = bing_videos.response(response) results = bing_videos.response(response)
self.assertEqual(type(results), list) self.assertEqual(type(results), list)
self.assertEqual(len(results), 1) self.assertEqual(len(results), 1)
self.assertEqual(results[0]['title'], 'Title 1') self.assertEqual(results[0]['title'], 'Title 1')
self.assertEqual(results[0]['url'], 'https://url.com/1') self.assertEqual(results[0]['url'], 'https://bing.com/video')
self.assertEqual(results[0]['content'], 'Content 1')
self.assertEqual(results[0]['thumbnail'], 'thumb_1.jpg')
html = """
<div>
<div class="dg_u">
<a class="dv_i" href="https://url.com/1">
<div class="vthblock">
<div class="vthumb">
<img src="thumb_1.jpg" />
</div>
<div>
<div class="tl">
Title 1
</div>
</div>
</div>
<div class="videoInfoPanel">
<div class="pubInfo">
<div>Content 1</div>
</div>
</div>
</a>
</div>
<div class="dg_u">
<a class="dv_i" href="/videos/search?abcde">
<div class="vthblock">
<div class="vthumb">
<img src="thumb_2.jpg" />
</div>
<div>
<div class="tl">
Title 2
</div>
</div>
</div>
<div class="videoInfoPanel">
<div class="pubInfo">
<div>Content 2</div>
</div>
</div>
</a>
</div>
</div>
"""
response = mock.Mock(text=html)
results = bing_videos.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 1)
self.assertEqual(results[0]['title'], 'Title 1')
self.assertEqual(results[0]['url'], 'https://url.com/1')
self.assertEqual(results[0]['content'], 'Content 1') self.assertEqual(results[0]['content'], 'Content 1')
self.assertEqual(results[0]['thumbnail'], 'thumb_1.jpg') self.assertEqual(results[0]['thumbnail'], 'thumb_1.jpg')