Ponysearch/searx/engines/wikipedia.py

113 lines
3.3 KiB
Python
Raw Normal View History

"""
Wikipedia (Web)
@website https://{language}.wikipedia.org
@provide-api yes
@using-api yes
@results JSON
@stable yes
@parse url, infobox
"""
from json import loads
from lxml.html import fromstring
2016-11-30 18:43:03 +01:00
from searx.url_utils import quote, urlencode
from searx.utils import match_language
# search-url
2016-11-30 18:43:03 +01:00
base_url = u'https://{language}.wikipedia.org/'
search_url = base_url + u'w/api.php?'\
'action=query'\
'&format=json'\
'&{query}'\
'&prop=extracts|pageimages|pageprops'\
'&ppprop=disambiguation'\
'&exintro'\
'&explaintext'\
'&pithumbsize=300'\
'&redirects'
supported_languages_url = 'https://meta.wikimedia.org/wiki/List_of_Wikipedias'
# set language in base_url
def url_lang(lang):
lang_pre = lang.split('-')[0]
2019-01-07 21:28:58 +01:00
if lang_pre == 'all' or lang_pre not in supported_languages and lang_pre not in language_aliases:
return 'en'
return match_language(lang, supported_languages, language_aliases).split('-')[0]
# do search-request
def request(query, params):
if query.islower():
2016-11-30 18:43:03 +01:00
query = u'{0}|{1}'.format(query.decode('utf-8'), query.decode('utf-8').title()).encode('utf-8')
2016-11-30 18:43:03 +01:00
params['url'] = search_url.format(query=urlencode({'titles': query}),
language=url_lang(params['language']))
return params
# get response from search-request
def response(resp):
results = []
2016-11-30 18:43:03 +01:00
search_result = loads(resp.text)
# wikipedia article's unique id
# first valid id is assumed to be the requested article
if 'pages' not in search_result['query']:
return results
for article_id in search_result['query']['pages']:
page = search_result['query']['pages'][article_id]
if int(article_id) > 0:
break
if int(article_id) < 0 or 'disambiguation' in page.get('pageprops', {}):
return []
title = page.get('title')
image = page.get('thumbnail')
if image:
image = image.get('source')
2020-07-27 08:27:16 +02:00
summary = page.get('extract', '').split('\n')[0].replace('()', '')
# link to wikipedia article
2016-11-30 18:43:03 +01:00
wikipedia_link = base_url.format(language=url_lang(resp.search_params['language'])) \
+ 'wiki/' + quote(title.replace(' ', '_').encode('utf8'))
results.append({'url': wikipedia_link, 'title': title})
results.append({'infobox': title,
'id': wikipedia_link,
'content': summary,
'img_src': image,
'urls': [{'title': 'Wikipedia', 'url': wikipedia_link}]})
return results
# get supported languages from their site
def _fetch_supported_languages(resp):
supported_languages = {}
dom = fromstring(resp.text)
tables = dom.xpath('//table[contains(@class,"sortable")]')
for table in tables:
# exclude header row
trs = table.xpath('.//tr')[1:]
for tr in trs:
td = tr.xpath('./td')
code = td[3].xpath('./a')[0].text
name = td[2].xpath('./a')[0].text
english_name = td[1].xpath('./a')[0].text
articles = int(td[4].xpath('./a/b')[0].text.replace(',', ''))
# exclude languages with too few articles
if articles >= 100:
supported_languages[code] = {"name": name, "english_name": english_name, "articles": articles}
return supported_languages