Ponysearch/searx/engines/bing_images.py

108 lines
3.3 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
"""
2014-09-01 14:39:18 +02:00
from lxml import html
2016-02-01 21:40:28 +01:00
from json import loads
2014-09-01 14:39:18 +02:00
import re
2016-11-30 18:43:03 +01:00
from searx.url_utils import urlencode
from searx.utils import match_language
2014-09-01 14:39:18 +02:00
from searx.engines.bing import _fetch_supported_languages, supported_languages_url, language_aliases
2014-09-01 14:39:18 +02:00
# engine dependent config
categories = ['images']
paging = True
safesearch = True
2016-10-30 20:58:34 +01:00
time_range_support = True
language_support = True
supported_languages_url = 'https://www.bing.com/account/general'
number_of_results = 28
2014-09-01 14:39:18 +02:00
# search-url
base_url = 'https://www.bing.com/'
search_string = 'images/search'\
'?{query}'\
'&count={count}'\
'&first={first}'\
'&FORM=IBASEP'
2016-10-30 20:58:34 +01:00
time_range_string = '&qft=+filterui:age-lt{interval}'
time_range_dict = {'day': '1440',
'week': '10080',
'month': '43200',
'year': '525600'}
2014-09-01 14:39:18 +02:00
# safesearch definitions
safesearch_types = {2: 'STRICT',
1: 'DEMOTE',
0: 'OFF'}
2014-09-02 17:13:44 +02:00
2015-02-08 22:01:24 +01:00
2014-09-01 14:39:18 +02:00
# do search-request
def request(query, params):
offset = ((params['pageno'] - 1) * number_of_results) + 1
2014-09-01 14:39:18 +02:00
search_path = search_string.format(
query=urlencode({'q': query}),
count=number_of_results,
first=offset)
2014-09-01 14:39:18 +02:00
language = match_language(params['language'], supported_languages, language_aliases).lower()
2014-09-01 14:39:18 +02:00
params['cookies']['SRCHHPGUSR'] = \
'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')
params['cookies']['_EDGE_S'] = 'mkt=' + language +\
'&ui=' + language + '&F=1'
2014-09-01 14:39:18 +02:00
params['url'] = base_url + search_path
2016-10-30 20:58:34 +01:00
if params['time_range'] in time_range_dict:
params['url'] += time_range_string.format(interval=time_range_dict[params['time_range']])
2014-09-01 14:39:18 +02:00
return params
# get response from search-request
def response(resp):
results = []
dom = html.fromstring(resp.text)
2014-09-01 14:39:18 +02:00
# parse results
for result in dom.xpath('//div[@class="imgpt"]'):
img_format = result.xpath('./div[contains(@class, "img_info")]/span/text()')[0]
# Microsoft seems to experiment with this code so don't make the path too specific,
# just catch the text section for the first anchor in img_info assuming this to be
# the originating site.
source = result.xpath('./div[contains(@class, "img_info")]//a/text()')[0]
try:
m = loads(result.xpath('./a/@m')[0])
# strip 'Unicode private use area' highlighting, they render to Tux
# the Linux penguin and a standing diamond on my machine...
title = m.get('t', '').replace(u'\ue000', '').replace(u'\ue001', '')
results.append({'template': 'images.html',
'url': m['purl'],
'thumbnail_src': m['turl'],
'img_src': m['murl'],
'content': '',
'title': title,
'source': source,
'img_format': img_format})
except:
continue
2014-09-01 14:39:18 +02:00
return results