searxng/searx/engines/__init__.py
2013-10-20 20:47:00 +02:00

120 lines
4.5 KiB
Python

'''
searx is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
searx is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with searx. If not, see < http://www.gnu.org/licenses/ >.
(C) 2013- by Adam Tauber, <asciimoo@gmail.com>
'''
from os.path import realpath, dirname, splitext, join
from os import listdir
from imp import load_source
import grequests
from itertools import izip_longest, chain
from operator import itemgetter
from urlparse import urlparse
from searx import settings
engine_dir = dirname(realpath(__file__))
engines = {}
categories = {'general': []}
for filename in listdir(engine_dir):
if filename.startswith('_') or not filename.endswith('.py'):
continue
modname = splitext(filename)[0]
if modname in settings.blacklist:
continue
filepath = join(engine_dir, filename)
engine = load_source(modname, filepath)
engine.name = modname
if not hasattr(engine, 'request') or not hasattr(engine, 'response'):
continue
engines[modname] = engine
if not hasattr(engine, 'categories'):
categories['general'].append(engine)
else:
for category_name in engine.categories:
categories.setdefault(category_name, []).append(engine)
def default_request_params():
return {'method': 'GET', 'headers': {}, 'data': {}, 'url': '', 'cookies': {}}
def make_callback(engine_name, results, callback):
def process_callback(response, **kwargs):
cb_res = []
for result in callback(response):
result['engine'] = engine_name
cb_res.append(result)
results[engine_name] = cb_res
return process_callback
def search(query, request, selected_engines):
global engines
requests = []
results = {}
user_agent = request.headers.get('User-Agent', '')
for ename, engine in engines.items():
if ename not in selected_engines:
continue
request_params = default_request_params()
request_params['headers']['User-Agent'] = user_agent
request_params = engine.request(query, request_params)
callback = make_callback(ename, results, engine.response)
if request_params['method'] == 'GET':
req = grequests.get(request_params['url']
,headers=request_params['headers']
,hooks=dict(response=callback)
,cookies = request_params['cookies']
)
else:
req = grequests.post(request_params['url']
,data=request_params['data']
,headers=request_params['headers']
,hooks=dict(response=callback)
,cookies = request_params['cookies']
)
requests.append(req)
grequests.map(requests)
flat_res = filter(None, chain.from_iterable(izip_longest(*results.values())))
flat_len = len(flat_res)
results = []
# deduplication + scoring
for i,res in enumerate(flat_res):
res['parsed_url'] = urlparse(res['url'])
score = (flat_len - i)*settings.weights.get(res['engine'], 1)
duplicated = False
for new_res in results:
if res['parsed_url'].netloc == new_res['parsed_url'].netloc and\
res['parsed_url'].path == new_res['parsed_url'].path and\
res['parsed_url'].query == new_res['parsed_url'].query:
duplicated = new_res
break
if duplicated:
if len(res.get('content', '')) > len(duplicated.get('content', '')):
duplicated['content'] = res['content']
duplicated['score'] += score
duplicated['engine'] += ', '+res['engine']
if duplicated['parsed_url'].scheme == 'https':
continue
elif res['parsed_url'].scheme == 'https':
duplicated['parsed_url'].scheme == 'https'
duplicated['url'] = duplicated['parsed_url'].geturl()
else:
res['score'] = score
results.append(res)
return sorted(results, key=itemgetter('score'), reverse=True)