searxng/searx/engines/bing_images.py

126 lines
3.5 KiB
Python
Raw Normal View History

"""
Bing (Images)
@website https://www.bing.com/images
@provide-api yes (http://datamarket.azure.com/dataset/bing/search),
max. 5000 query/month
@using-api no (because of query limit)
@results HTML (using search portal)
@stable no (HTML can change)
@parse url, title, img_src
@todo currently there are up to 35 images receive per page,
because bing does not parse count=10.
limited response to 10 images
"""
2014-09-01 12:39:18 +00:00
from lxml import html
2016-02-01 20:40:28 +00:00
from json import loads
2014-09-01 12:39:18 +00:00
import re
2016-11-30 17:43:03 +00:00
from searx.url_utils import urlencode
from searx.utils import match_language
2014-09-01 12:39:18 +00:00
# engine dependent config
categories = ['images']
paging = True
safesearch = True
2016-10-30 19:58:34 +00:00
time_range_support = True
language_support = True
supported_languages_url = 'https://www.bing.com/account/general'
2014-09-01 12:39:18 +00:00
# search-url
base_url = 'https://www.bing.com/'
search_string = 'images/search?{query}&count=10&first={offset}'
2016-10-30 19:58:34 +00:00
time_range_string = '&qft=+filterui:age-lt{interval}'
time_range_dict = {'day': '1440',
'week': '10080',
'month': '43200',
'year': '525600'}
2014-09-01 12:39:18 +00:00
# safesearch definitions
safesearch_types = {2: 'STRICT',
1: 'DEMOTE',
0: 'OFF'}
2014-09-02 15:13:44 +00:00
2015-02-08 21:01:24 +00:00
2016-02-01 20:40:28 +00:00
_quote_keys_regex = re.compile('({|,)([a-z][a-z0-9]*):(")', re.I | re.U)
2014-09-01 12:39:18 +00:00
# do search-request
def request(query, params):
offset = (params['pageno'] - 1) * 10 + 1
search_path = search_string.format(
query=urlencode({'q': query}),
offset=offset)
language = match_language(params['language'], supported_languages, language_aliases).lower()
2014-09-01 12:39:18 +00:00
params['cookies']['SRCHHPGUSR'] = \
'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')
params['cookies']['_EDGE_S'] = 'mkt=' + language +\
'&ui=' + language + '&F=1'
2014-09-01 12:39:18 +00:00
params['url'] = base_url + search_path
2016-10-30 19:58:34 +00:00
if params['time_range'] in time_range_dict:
params['url'] += time_range_string.format(interval=time_range_dict[params['time_range']])
2014-09-01 12:39:18 +00:00
return params
# get response from search-request
def response(resp):
results = []
dom = html.fromstring(resp.text)
2014-09-01 12:39:18 +00:00
# parse results
2017-05-13 07:22:27 +00:00
for result in dom.xpath('//div[@id="mmComponent_images_1"]/ul/li/div/div[@class="imgpt"]'):
2014-09-01 12:39:18 +00:00
link = result.xpath('./a')[0]
2017-05-13 07:22:27 +00:00
# TODO find actual title
title = link.xpath('.//img/@alt')[0]
2016-02-01 20:40:28 +00:00
# parse json-data (it is required to add a space, to make it parsable)
json_data = loads(_quote_keys_regex.sub(r'\1"\2": \3', link.attrib.get('m')))
2017-05-13 07:22:27 +00:00
url = json_data.get('purl')
img_src = json_data.get('murl')
thumbnail = json_data.get('turl')
2014-09-01 12:39:18 +00:00
# append result
results.append({'template': 'images.html',
'url': url,
'title': title,
'content': '',
2017-05-13 07:22:27 +00:00
'thumbnail_src': thumbnail,
2014-09-01 12:39:18 +00:00
'img_src': img_src})
# TODO stop parsing if 10 images are found
2017-05-13 07:22:27 +00:00
# if len(results) >= 10:
# break
2014-09-01 12:39:18 +00:00
# return results
return results
# get supported languages from their site
def _fetch_supported_languages(resp):
supported_languages = []
dom = html.fromstring(resp.text)
regions_xpath = '//div[@id="region-section-content"]' \
+ '//ul[@class="b_vList"]/li/a/@href'
regions = dom.xpath(regions_xpath)
for region in regions:
code = re.search('setmkt=[^\&]+', region).group()[7:]
if code == 'nb-NO':
code = 'no-NO'
supported_languages.append(code)
return supported_languages