| 
									
										
										
										
											2015-05-02 15:45:17 +02:00
										 |  |  | """
 | 
					
						
							|  |  |  |  Gigablast (Web) | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2015-12-22 20:25:57 +01:00
										 |  |  |  @website     https://gigablast.com | 
					
						
							|  |  |  |  @provide-api yes (https://gigablast.com/api.html) | 
					
						
							| 
									
										
										
										
											2015-05-02 15:45:17 +02:00
										 |  |  | 
 | 
					
						
							|  |  |  |  @using-api   yes | 
					
						
							|  |  |  |  @results     XML | 
					
						
							|  |  |  |  @stable      yes | 
					
						
							|  |  |  |  @parse       url, title, content | 
					
						
							|  |  |  | """
 | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  | from cgi import escape | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  | from json import loads | 
					
						
							| 
									
										
										
										
											2015-10-16 12:05:50 +02:00
										 |  |  | from random import randint | 
					
						
							|  |  |  | from time import time | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  | from urllib import urlencode | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  | # engine dependent config | 
					
						
							|  |  |  | categories = ['general'] | 
					
						
							|  |  |  | paging = True | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  | number_of_results = 10 | 
					
						
							|  |  |  | language_support = True | 
					
						
							|  |  |  | safesearch = True | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  | # search-url | 
					
						
							| 
									
										
										
										
											2015-12-22 20:25:57 +01:00
										 |  |  | base_url = 'https://gigablast.com/' | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  | search_string = 'search?{query}'\ | 
					
						
							|  |  |  |     '&n={number_of_results}'\ | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |     '&c=main'\ | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  |     '&s={offset}'\ | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |     '&format=json'\ | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  |     '&qh=0'\ | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |     '&rxiwd={rxiwd}'\ | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  |     '&qlang={lang}'\ | 
					
						
							|  |  |  |     '&ff={safesearch}' | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  | # specific xpath variables | 
					
						
							|  |  |  | results_xpath = '//response//result' | 
					
						
							|  |  |  | url_xpath = './/url' | 
					
						
							|  |  |  | title_xpath = './/title' | 
					
						
							|  |  |  | content_xpath = './/sum' | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | # do search-request | 
					
						
							|  |  |  | def request(query, params): | 
					
						
							|  |  |  |     offset = (params['pageno'] - 1) * number_of_results | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  |     if params['language'] == 'all': | 
					
						
							|  |  |  |         language = 'xx' | 
					
						
							|  |  |  |     else: | 
					
						
							|  |  |  |         language = params['language'][0:2] | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     if params['safesearch'] >= 1: | 
					
						
							|  |  |  |         safesearch = 1 | 
					
						
							|  |  |  |     else: | 
					
						
							|  |  |  |         safesearch = 0 | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     search_path = search_string.format(query=urlencode({'q': query}), | 
					
						
							|  |  |  |                                        offset=offset, | 
					
						
							|  |  |  |                                        number_of_results=number_of_results, | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |                                        rxiwd=1, | 
					
						
							|  |  |  |                                        #  rand=int(time()), | 
					
						
							| 
									
										
										
										
											2015-12-23 18:43:35 +01:00
										 |  |  |                                        lang=language, | 
					
						
							|  |  |  |                                        safesearch=safesearch) | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  |     params['url'] = base_url + search_path | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     return params | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | # get response from search-request | 
					
						
							|  |  |  | def response(resp): | 
					
						
							|  |  |  |     results = [] | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     # parse results | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |     response_json = loads(resp.text) | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |     for result in response_json['results']: | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  |         # append result | 
					
						
							| 
									
										
										
										
											2016-01-31 13:24:09 +01:00
										 |  |  |         results.append({'url': result['url'], | 
					
						
							|  |  |  |                         'title': escape(result['title']), | 
					
						
							|  |  |  |                         'content': escape(result['sum'])}) | 
					
						
							| 
									
										
										
										
											2015-02-08 14:12:14 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  |     # return results | 
					
						
							|  |  |  |     return results |