using general mediawiki-engine

* writing general mediawiki-engine
* using this engine for wikipedia
* using this engine for uncyclopedia
This commit is contained in:
Thomas Pointhuber 2014-09-03 11:40:29 +02:00
parent 629a05e149
commit 52ad49ccba
3 changed files with 60 additions and 81 deletions

View File

@ -1,22 +1,67 @@
## Wikipedia (Web)
#
# @website http://www.wikipedia.org
# @provide-api yes (http://www.mediawiki.org/wiki/API:Search)
#
# @using-api yes
# @results JSON
# @stable yes
# @parse url, title
#
# @todo content
from json import loads from json import loads
from urllib import urlencode, quote from urllib import urlencode, quote
url = 'https://en.wikipedia.org/' # engine dependent config
categories = ['general']
search_url = url + 'w/api.php?action=query&list=search&{query}&srprop=timestamp&format=json&sroffset={offset}' # noqa language_support = True
paging = True
number_of_results = 10 number_of_results = 1
# search-url
base_url = 'https://{language}.wikipedia.org/'
search_url = base_url + 'w/api.php?action=query&list=search&{query}&srprop=timestamp&format=json&sroffset={offset}&srlimit={limit}' # noqa
# do search-request
def request(query, params): def request(query, params):
offset = (params['pageno'] - 1) * 10 offset = (params['pageno'] - 1) * number_of_results
if params['language'] == 'all':
language = 'en'
else:
language = params['language'].split('_')[0]
# write search-language back to params, required in response
params['language'] = language
params['url'] = search_url.format(query=urlencode({'srsearch': query}), params['url'] = search_url.format(query=urlencode({'srsearch': query}),
offset=offset) offset=offset,
limit=number_of_results,
language=language)
return params return params
# get response from search-request
def response(resp): def response(resp):
results = []
search_results = loads(resp.text) search_results = loads(resp.text)
res = search_results.get('query', {}).get('search', [])
return [{'url': url + 'wiki/' + quote(result['title'].replace(' ', '_').encode('utf-8')), # noqa # return empty array if there are no results
'title': result['title']} for result in res[:int(number_of_results)]] if not search_results.get('query', {}).get('search'):
return []
# parse results
for result in search_results['query']['search']:
url = base_url.format(language=resp.search_params['language']) + 'wiki/' + quote(result['title'].replace(' ', '_').encode('utf-8'))
# append result
results.append({'url': url,
'title': result['title'],
'content': ''})
# return results
return results

View File

@ -1,67 +0,0 @@
## Wikipedia (Web)
#
# @website http://www.wikipedia.org
# @provide-api yes (http://www.mediawiki.org/wiki/API:Search)
#
# @using-api yes
# @results JSON
# @stable yes
# @parse url, title
#
# @todo content
from json import loads
from urllib import urlencode, quote
# engine dependent config
categories = ['general']
language_support = True
paging = True
number_of_results = 1
# search-url
url = 'https://{language}.wikipedia.org/'
search_url = url + 'w/api.php?action=query&list=search&{query}&srprop=timestamp&format=json&sroffset={offset}&srlimit={limit}' # noqa
# do search-request
def request(query, params):
offset = (params['pageno'] - 1) * number_of_results
if params['language'] == 'all':
language = 'en'
else:
language = params['language'].split('_')[0]
# write search-language back to params, required in response
params['language'] = language
params['url'] = search_url.format(query=urlencode({'srsearch': query}),
offset=offset,
limit=number_of_results,
language=language)
return params
# get response from search-request
def response(resp):
results = []
search_results = loads(resp.text)
# return empty array if there are no results
if not search_results.get('query', {}).get('search'):
return []
# parse results
for result in search_results['query']['search']:
res_url = url.format(language=resp.search_params['language']) + 'wiki/' + quote(result['title'].replace(' ', '_').encode('utf-8'))
# append result
results.append({'url': res_url,
'title': result['title'],
'content': ''})
# return results
return results

View File

@ -10,9 +10,10 @@ server:
engines: engines:
- name : wikipedia - name : wikipedia
engine : wikipedia engine : mediawiki
shortcut : wp shortcut : wp
# number_of_results : 1 # default is 1 base_url : 'https://{language}.wikipedia.org/'
number_of_results : 1
- name : bing - name : bing
engine : bing engine : bing
@ -108,9 +109,9 @@ engines:
# maybe in a fun category # maybe in a fun category
# - name : uncyclopedia # - name : uncyclopedia
# engine : mediawiki # engine : mediawiki
# categories : general
# shortcut : unc # shortcut : unc
# url : https://uncyclopedia.wikia.com/ # base_url : https://uncyclopedia.wikia.com/
# number_of_results : 5
# tmp suspended - too slow, too many errors # tmp suspended - too slow, too many errors
# - name : urbandictionary # - name : urbandictionary