| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-17 00:32:32 +02:00
										 |  |  | '''
 | 
					
						
							|  |  |  | 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> | 
					
						
							|  |  |  | '''
 | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  | from os.path import realpath, dirname, splitext, join | 
					
						
							| 
									
										
										
										
											2014-02-05 20:24:31 +01:00
										 |  |  | import sys | 
					
						
							| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  | from imp import load_source | 
					
						
							| 
									
										
										
										
											2016-07-04 22:46:43 +02:00
										 |  |  | from flask_babel import gettext | 
					
						
							| 
									
										
										
										
											2014-07-07 13:59:27 +02:00
										 |  |  | from operator import itemgetter | 
					
						
							| 
									
										
										
										
											2014-02-05 20:24:31 +01:00
										 |  |  | from searx import settings | 
					
						
							| 
									
										
										
										
											2015-01-09 04:13:05 +01:00
										 |  |  | from searx import logger | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | logger = logger.getChild('engines') | 
					
						
							| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  | 
 | 
					
						
							|  |  |  | engine_dir = dirname(realpath(__file__)) | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-15 19:11:43 +02:00
										 |  |  | engines = {} | 
					
						
							| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-17 21:06:28 +02:00
										 |  |  | categories = {'general': []} | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | engine_shortcuts = {} | 
					
						
							| 
									
										
										
										
											2016-02-19 15:13:01 +01:00
										 |  |  | engine_default_args = {'paging': False, | 
					
						
							|  |  |  |                        'categories': ['general'], | 
					
						
							|  |  |  |                        'language_support': True, | 
					
						
							|  |  |  |                        'safesearch': False, | 
					
						
							|  |  |  |                        'timeout': settings['outgoing']['request_timeout'], | 
					
						
							|  |  |  |                        'shortcut': '-', | 
					
						
							| 
									
										
										
										
											2016-02-20 00:21:56 +01:00
										 |  |  |                        'disabled': False, | 
					
						
							|  |  |  |                        'suspend_end_time': 0, | 
					
						
							| 
									
										
										
										
											2016-07-17 18:42:30 +02:00
										 |  |  |                        'continuous_errors': 0, | 
					
						
							|  |  |  |                        'time_range_support': False} | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  | def load_module(filename): | 
					
						
							| 
									
										
										
										
											2013-10-19 19:01:06 +02:00
										 |  |  |     modname = splitext(filename)[0] | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  |     if modname in sys.modules: | 
					
						
							|  |  |  |         del sys.modules[modname] | 
					
						
							| 
									
										
										
										
											2013-10-14 23:09:13 +02:00
										 |  |  |     filepath = join(engine_dir, filename) | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  |     module = load_source(modname, filepath) | 
					
						
							|  |  |  |     module.name = modname | 
					
						
							|  |  |  |     return module | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-26 18:44:58 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-12-13 19:26:40 +01:00
										 |  |  | def load_engine(engine_data): | 
					
						
							| 
									
										
										
										
											2016-09-28 22:30:05 +02:00
										 |  |  | 
 | 
					
						
							|  |  |  |     if '_' in engine_data['name']: | 
					
						
							|  |  |  |         logger.error('Engine name conains underscore: "{}"'.format(engine_data['name'])) | 
					
						
							|  |  |  |         sys.exit(1) | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     engine_module = engine_data['engine'] | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2016-07-25 23:36:52 +02:00
										 |  |  |     try: | 
					
						
							| 
									
										
										
										
											2016-09-28 22:30:05 +02:00
										 |  |  |         engine = load_module(engine_module + '.py') | 
					
						
							| 
									
										
										
										
											2016-07-25 23:36:52 +02:00
										 |  |  |     except: | 
					
						
							| 
									
										
										
										
											2016-09-28 22:30:05 +02:00
										 |  |  |         logger.exception('Cannot load engine "{}"'.format(engine_module)) | 
					
						
							| 
									
										
										
										
											2016-07-25 23:36:52 +02:00
										 |  |  |         return None | 
					
						
							| 
									
										
										
										
											2014-01-31 04:35:23 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  |     for param_name in engine_data: | 
					
						
							|  |  |  |         if param_name == 'engine': | 
					
						
							|  |  |  |             continue | 
					
						
							|  |  |  |         if param_name == 'categories': | 
					
						
							| 
									
										
										
										
											2014-01-19 00:17:02 +01:00
										 |  |  |             if engine_data['categories'] == 'none': | 
					
						
							| 
									
										
										
										
											2013-11-04 21:46:23 +01:00
										 |  |  |                 engine.categories = [] | 
					
						
							|  |  |  |             else: | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |                 engine.categories = map( | 
					
						
							|  |  |  |                     str.strip, engine_data['categories'].split(',')) | 
					
						
							| 
									
										
										
										
											2013-10-23 23:54:46 +02:00
										 |  |  |             continue | 
					
						
							| 
									
										
										
										
											2014-01-19 00:17:02 +01:00
										 |  |  |         setattr(engine, param_name, engine_data[param_name]) | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2016-02-19 15:13:01 +01:00
										 |  |  |     for arg_name, arg_value in engine_default_args.iteritems(): | 
					
						
							|  |  |  |         if not hasattr(engine, arg_name): | 
					
						
							|  |  |  |             setattr(engine, arg_name, arg_value) | 
					
						
							| 
									
										
										
										
											2015-01-31 23:11:45 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  |     # checking required variables | 
					
						
							| 
									
										
										
										
											2013-10-25 23:41:14 +02:00
										 |  |  |     for engine_attr in dir(engine): | 
					
						
							|  |  |  |         if engine_attr.startswith('_'): | 
					
						
							|  |  |  |             continue | 
					
						
							| 
									
										
										
										
											2014-01-20 02:31:20 +01:00
										 |  |  |         if getattr(engine, engine_attr) is None: | 
					
						
							| 
									
										
										
										
											2015-01-09 04:13:05 +01:00
										 |  |  |             logger.error('Missing engine config attribute: "{0}.{1}"' | 
					
						
							| 
									
										
										
										
											2015-05-02 15:45:17 +02:00
										 |  |  |                          .format(engine.name, engine_attr)) | 
					
						
							| 
									
										
										
										
											2013-10-25 23:41:14 +02:00
										 |  |  |             sys.exit(1) | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |     engine.stats = { | 
					
						
							|  |  |  |         'result_count': 0, | 
					
						
							|  |  |  |         'search_count': 0, | 
					
						
							|  |  |  |         'page_load_time': 0, | 
					
						
							|  |  |  |         'score_count': 0, | 
					
						
							|  |  |  |         'errors': 0 | 
					
						
							|  |  |  |     } | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2016-02-19 15:13:01 +01:00
										 |  |  |     for category_name in engine.categories: | 
					
						
							|  |  |  |         categories.setdefault(category_name, []).append(engine) | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     if engine.shortcut in engine_shortcuts: | 
					
						
							|  |  |  |         logger.error('Engine config error: ambigious shortcut: {0}'.format(engine.shortcut)) | 
					
						
							|  |  |  |         sys.exit(1) | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  |     engine_shortcuts[engine.shortcut] = engine.name | 
					
						
							| 
									
										
										
										
											2013-10-15 18:19:06 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-12-13 19:26:40 +01:00
										 |  |  |     return engine | 
					
						
							| 
									
										
										
										
											2014-01-31 15:45:18 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-27 01:02:28 +02:00
										 |  |  | def get_engines_stats(): | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |     # TODO refactor | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  |     pageloads = [] | 
					
						
							|  |  |  |     results = [] | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |     scores = [] | 
					
						
							| 
									
										
										
										
											2013-10-27 20:45:21 +01:00
										 |  |  |     errors = [] | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |     scores_per_result = [] | 
					
						
							| 
									
										
										
										
											2013-10-27 01:02:28 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |     max_pageload = max_results = max_score = max_errors = max_score_per_result = 0  # noqa | 
					
						
							| 
									
										
										
										
											2013-10-27 01:02:28 +02:00
										 |  |  |     for engine in engines.values(): | 
					
						
							|  |  |  |         if engine.stats['search_count'] == 0: | 
					
						
							|  |  |  |             continue | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |         results_num = \ | 
					
						
							|  |  |  |             engine.stats['result_count'] / float(engine.stats['search_count']) | 
					
						
							|  |  |  |         load_times = engine.stats['page_load_time'] / float(engine.stats['search_count'])  # noqa | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |         if results_num: | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |             score = engine.stats['score_count'] / float(engine.stats['search_count'])  # noqa | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |             score_per_result = score / results_num | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |         else: | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |             score = score_per_result = 0.0 | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  |         max_results = max(results_num, max_results) | 
					
						
							|  |  |  |         max_pageload = max(load_times, max_pageload) | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |         max_score = max(score, max_score) | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |         max_score_per_result = max(score_per_result, max_score_per_result) | 
					
						
							| 
									
										
										
										
											2013-10-27 20:45:21 +01:00
										 |  |  |         max_errors = max(max_errors, engine.stats['errors']) | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  |         pageloads.append({'avg': load_times, 'name': engine.name}) | 
					
						
							|  |  |  |         results.append({'avg': results_num, 'name': engine.name}) | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |         scores.append({'avg': score, 'name': engine.name}) | 
					
						
							| 
									
										
										
										
											2013-10-27 20:45:21 +01:00
										 |  |  |         errors.append({'avg': engine.stats['errors'], 'name': engine.name}) | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |         scores_per_result.append({ | 
					
						
							|  |  |  |             'avg': score_per_result, | 
					
						
							|  |  |  |             'name': engine.name | 
					
						
							|  |  |  |         }) | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  | 
 | 
					
						
							|  |  |  |     for engine in pageloads: | 
					
						
							| 
									
										
										
										
											2014-10-01 19:16:34 +02:00
										 |  |  |         if max_pageload: | 
					
						
							|  |  |  |             engine['percentage'] = int(engine['avg'] / max_pageload * 100) | 
					
						
							|  |  |  |         else: | 
					
						
							|  |  |  |             engine['percentage'] = 0 | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  | 
 | 
					
						
							|  |  |  |     for engine in results: | 
					
						
							| 
									
										
										
										
											2014-10-01 19:16:34 +02:00
										 |  |  |         if max_results: | 
					
						
							|  |  |  |             engine['percentage'] = int(engine['avg'] / max_results * 100) | 
					
						
							|  |  |  |         else: | 
					
						
							|  |  |  |             engine['percentage'] = 0 | 
					
						
							| 
									
										
										
										
											2013-10-27 02:50:24 +02:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  |     for engine in scores: | 
					
						
							| 
									
										
										
										
											2014-10-01 19:16:34 +02:00
										 |  |  |         if max_score: | 
					
						
							|  |  |  |             engine['percentage'] = int(engine['avg'] / max_score * 100) | 
					
						
							|  |  |  |         else: | 
					
						
							|  |  |  |             engine['percentage'] = 0 | 
					
						
							| 
									
										
										
										
											2013-10-27 14:21:23 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  |     for engine in scores_per_result: | 
					
						
							| 
									
										
										
										
											2014-10-01 19:16:34 +02:00
										 |  |  |         if max_score_per_result: | 
					
						
							| 
									
										
										
										
											2014-10-19 12:41:04 +02:00
										 |  |  |             engine['percentage'] = int(engine['avg'] | 
					
						
							|  |  |  |                                        / max_score_per_result * 100) | 
					
						
							| 
									
										
										
										
											2014-10-01 19:16:34 +02:00
										 |  |  |         else: | 
					
						
							|  |  |  |             engine['percentage'] = 0 | 
					
						
							| 
									
										
										
										
											2014-01-05 23:13:53 +01:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2013-10-27 20:45:21 +01:00
										 |  |  |     for engine in errors: | 
					
						
							|  |  |  |         if max_errors: | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |             engine['percentage'] = int(float(engine['avg']) / max_errors * 100) | 
					
						
							| 
									
										
										
										
											2013-10-27 20:45:21 +01:00
										 |  |  |         else: | 
					
						
							|  |  |  |             engine['percentage'] = 0 | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |     return [ | 
					
						
							|  |  |  |         ( | 
					
						
							| 
									
										
										
										
											2014-01-22 00:17:49 +01:00
										 |  |  |             gettext('Page loads (sec)'), | 
					
						
							|  |  |  |             sorted(pageloads, key=itemgetter('avg')) | 
					
						
							|  |  |  |         ), | 
					
						
							|  |  |  |         ( | 
					
						
							|  |  |  |             gettext('Number of results'), | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |             sorted(results, key=itemgetter('avg'), reverse=True) | 
					
						
							|  |  |  |         ), | 
					
						
							|  |  |  |         ( | 
					
						
							| 
									
										
										
										
											2014-01-22 00:17:49 +01:00
										 |  |  |             gettext('Scores'), | 
					
						
							|  |  |  |             sorted(scores, key=itemgetter('avg'), reverse=True) | 
					
						
							|  |  |  |         ), | 
					
						
							|  |  |  |         ( | 
					
						
							|  |  |  |             gettext('Scores per result'), | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |             sorted(scores_per_result, key=itemgetter('avg'), reverse=True) | 
					
						
							|  |  |  |         ), | 
					
						
							| 
									
										
										
										
											2014-01-22 00:17:49 +01:00
										 |  |  |         ( | 
					
						
							|  |  |  |             gettext('Errors'), | 
					
						
							|  |  |  |             sorted(errors, key=itemgetter('avg'), reverse=True) | 
					
						
							|  |  |  |         ), | 
					
						
							| 
									
										
										
										
											2014-01-19 22:59:01 +01:00
										 |  |  |     ] | 
					
						
							| 
									
										
										
										
											2014-12-13 19:26:40 +01:00
										 |  |  | 
 | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | if 'engines' not in settings or not settings['engines']: | 
					
						
							| 
									
										
										
										
											2015-01-09 04:13:05 +01:00
										 |  |  |     logger.error('No engines found. Edit your settings.yml') | 
					
						
							| 
									
										
										
										
											2014-12-13 19:26:40 +01:00
										 |  |  |     exit(2) | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | for engine_data in settings['engines']: | 
					
						
							|  |  |  |     engine = load_engine(engine_data) | 
					
						
							| 
									
										
										
										
											2016-07-25 23:36:52 +02:00
										 |  |  |     if engine is not None: | 
					
						
							|  |  |  |         engines[engine.name] = engine |