add duckduckgo images engine

This commit is contained in:
marc 2017-05-20 22:33:08 -05:00 committed by Adam Tauber
parent e60e98156f
commit c65a409f0d
4 changed files with 205 additions and 31 deletions

View File

@ -41,6 +41,39 @@ title_xpath = './/a[@class="result__a"]'
content_xpath = './/a[@class="result__snippet"]' content_xpath = './/a[@class="result__snippet"]'
# match query's language to a region code that duckduckgo will accept
def get_region_code(lang):
# custom fixes for languages
if lang == 'all':
region_code = None
elif lang[:2] == 'ja':
region_code = 'jp-jp'
elif lang[:2] == 'sl':
region_code = 'sl-sl'
elif lang == 'zh-TW':
region_code = 'tw-tzh'
elif lang == 'zh-HK':
region_code = 'hk-tzh'
elif lang[-2:] == 'SA':
region_code = 'xa-' + lang.split('-')[0]
elif lang[-2:] == 'GB':
region_code = 'uk-' + lang.split('-')[0]
else:
region_code = lang.split('-')
if len(region_code) == 2:
# country code goes first
region_code = region_code[1].lower() + '-' + region_code[0].lower()
else:
# tries to get a country code from language
region_code = region_code[0].lower()
for lc in supported_languages:
lc = lc.split('-')
if region_code == lc[0]:
region_code = lc[1].lower() + '-' + lc[0].lower()
break
return region_code
# do search-request # do search-request
def request(query, params): def request(query, params):
if params['time_range'] and params['time_range'] not in time_range_dict: if params['time_range'] and params['time_range'] not in time_range_dict:
@ -49,38 +82,10 @@ def request(query, params):
offset = 30 + (params['pageno'] - 1) * 50 offset = 30 + (params['pageno'] - 1) * 50
dc_param = offset + 1 dc_param = offset + 1
# custom fixes for languages region_code = get_region_code(params['language'])
if params['language'] == 'all': if region_code:
locale = None
elif params['language'][:2] == 'ja':
locale = 'jp-jp'
elif params['language'][:2] == 'sl':
locale = 'sl-sl'
elif params['language'] == 'zh-TW':
locale = 'tw-tzh'
elif params['language'] == 'zh-HK':
locale = 'hk-tzh'
elif params['language'][-2:] == 'SA':
locale = 'xa-' + params['language'].split('-')[0]
elif params['language'][-2:] == 'GB':
locale = 'uk-' + params['language'].split('-')[0]
else:
locale = params['language'].split('-')
if len(locale) == 2:
# country code goes first
locale = locale[1].lower() + '-' + locale[0].lower()
else:
# tries to get a country code from language
locale = locale[0].lower()
for lc in supported_languages:
lc = lc.split('-')
if locale == lc[0]:
locale = lc[1].lower() + '-' + lc[0].lower()
break
if locale:
params['url'] = url.format( params['url'] = url.format(
query=urlencode({'q': query, 'kl': locale}), offset=offset, dc_param=dc_param) query=urlencode({'q': query, 'kl': region_code}), offset=offset, dc_param=dc_param)
else: else:
params['url'] = url.format( params['url'] = url.format(
query=urlencode({'q': query}), offset=offset, dc_param=dc_param) query=urlencode({'q': query}), offset=offset, dc_param=dc_param)

View File

@ -0,0 +1,91 @@
"""
DuckDuckGo (Images)
@website https://duckduckgo.com/
@provide-api yes (https://duckduckgo.com/api),
but images are not supported
@using-api no
@results JSON (site requires js to get images)
@stable no (JSON can change)
@parse url, title, img_src
@todo avoid extra request
"""
from requests import get
from json import loads
from searx.engines.xpath import extract_text
from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url, get_region_code
from searx.url_utils import urlencode
# engine dependent config
categories = ['images']
paging = True
language_support = True
safesearch = True
# search-url
images_url = 'https://duckduckgo.com/i.js?{query}&s={offset}&p={safesearch}&o=json&vqd={vqd}'
site_url = 'https://duckduckgo.com/?{query}&iar=images&iax=1&ia=images'
# run query in site to get vqd number needed for requesting images
# TODO: find a way to get this number without an extra request (is it a hash of the query?)
def get_vqd(query):
res = get(site_url.format(query=urlencode({'q': query})))
content = res.text
vqd = content[content.find('vqd=\'') + 5:]
vqd = vqd[:vqd.find('\'')]
return vqd
# do search-request
def request(query, params):
# to avoid running actual external requests when testing
if 'is_test' not in params:
vqd = get_vqd(query)
else:
vqd = '12345'
offset = (params['pageno'] - 1) * 50
safesearch = params['safesearch'] - 1
region_code = get_region_code(params['language'])
if region_code:
params['url'] = images_url.format(
query=urlencode({'q': query, 'l': region_code}), offset=offset, safesearch=safesearch, vqd=vqd)
else:
params['url'] = images_url.format(
query=urlencode({'q': query}), offset=offset, safesearch=safesearch, vqd=vqd)
return params
# get response from search-request
def response(resp):
results = []
content = resp.text
try:
res_json = loads(content)
except:
return []
# parse results
for result in res_json['results']:
title = result['title']
url = result['url']
thumbnail = result['thumbnail']
image = result['image']
# append result
results.append({'template': 'images.html',
'title': title,
'content': '',
'thumbnail_src': thumbnail,
'img_src': image,
'url': url})
return results

View File

@ -167,6 +167,12 @@ engines:
shortcut : ddg shortcut : ddg
disabled : True disabled : True
- name : duckduckgo images
engine : duckduckgo_images
shortcut : ddi
timeout: 3.0
disabled : True
- name : etymonline - name : etymonline
engine : xpath engine : xpath
paging : True paging : True

View File

@ -0,0 +1,72 @@
# -*- coding: utf-8 -*-
from collections import defaultdict
import mock
from searx.engines import duckduckgo_images
from searx.testing import SearxTestCase
class TestDuckduckgoImagesEngine(SearxTestCase):
def test_request(self):
query = 'test_query'
dicto = defaultdict(dict)
dicto['is_test'] = True
dicto['pageno'] = 1
dicto['safesearch'] = 0
dicto['language'] = 'all'
params = duckduckgo_images.request(query, dicto)
self.assertIn('url', params)
self.assertIn(query, params['url'])
self.assertIn('duckduckgo.com', params['url'])
self.assertIn('s=0', params['url'])
self.assertIn('p=-1', params['url'])
self.assertIn('vqd=12345', params['url'])
# test paging and safe search
dicto['pageno'] = 2
dicto['safesearch'] = 2
params = duckduckgo_images.request(query, dicto)
self.assertIn('url', params)
self.assertIn(query, params['url'])
self.assertIn('s=50', params['url'])
self.assertIn('p=1', params['url'])
def test_response(self):
self.assertRaises(AttributeError, duckduckgo_images.response, None)
self.assertRaises(AttributeError, duckduckgo_images.response, [])
self.assertRaises(AttributeError, duckduckgo_images.response, '')
self.assertRaises(AttributeError, duckduckgo_images.response, '[]')
response = mock.Mock(text='If this error persists, please let us know: ops@duckduckgo.com')
self.assertEqual(duckduckgo_images.response(response), [])
json = u"""
{
"query": "test_query",
"results": [
{
"title": "Result 1",
"url": "https://site1.url",
"thumbnail": "https://thumb1.nail",
"image": "https://image1"
},
{
"title": "Result 2",
"url": "https://site2.url",
"thumbnail": "https://thumb2.nail",
"image": "https://image2"
}
]
}
"""
response = mock.Mock(text=json)
results = duckduckgo_images.response(response)
self.assertEqual(len(results), 2)
self.assertEqual(results[0]['title'], 'Result 1')
self.assertEqual(results[0]['url'], 'https://site1.url')
self.assertEqual(results[0]['thumbnail_src'], 'https://thumb1.nail')
self.assertEqual(results[0]['img_src'], 'https://image1')
self.assertEqual(results[1]['title'], 'Result 2')
self.assertEqual(results[1]['url'], 'https://site2.url')
self.assertEqual(results[1]['thumbnail_src'], 'https://thumb2.nail')
self.assertEqual(results[1]['img_src'], 'https://image2')