__init__.py 7.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219
  1. '''
  2. searx is free software: you can redistribute it and/or modify
  3. it under the terms of the GNU Affero General Public License as published by
  4. the Free Software Foundation, either version 3 of the License, or
  5. (at your option) any later version.
  6. searx is distributed in the hope that it will be useful,
  7. but WITHOUT ANY WARRANTY; without even the implied warranty of
  8. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  9. GNU Affero General Public License for more details.
  10. You should have received a copy of the GNU Affero General Public License
  11. along with searx. If not, see < http://www.gnu.org/licenses/ >.
  12. (C) 2013- by Adam Tauber, <asciimoo@gmail.com>
  13. '''
  14. from os.path import realpath, dirname
  15. import sys
  16. from flask_babel import gettext
  17. from operator import itemgetter
  18. from json import loads
  19. from searx import settings
  20. from searx import logger
  21. from searx.utils import load_module
  22. logger = logger.getChild('engines')
  23. engine_dir = dirname(realpath(__file__))
  24. engines = {}
  25. categories = {'general': []}
  26. _initialized = False
  27. engine_shortcuts = {}
  28. engine_default_args = {'paging': False,
  29. 'categories': ['general'],
  30. 'language_support': True,
  31. 'supported_languages': [],
  32. 'safesearch': False,
  33. 'timeout': settings['outgoing']['request_timeout'],
  34. 'shortcut': '-',
  35. 'disabled': False,
  36. 'suspend_end_time': 0,
  37. 'continuous_errors': 0,
  38. 'time_range_support': False}
  39. def load_engine(engine_data):
  40. if '_' in engine_data['name']:
  41. logger.error('Engine name conains underscore: "{}"'.format(engine_data['name']))
  42. sys.exit(1)
  43. engine_module = engine_data['engine']
  44. try:
  45. engine = load_module(engine_module + '.py', engine_dir)
  46. except:
  47. logger.exception('Cannot load engine "{}"'.format(engine_module))
  48. return None
  49. for param_name in engine_data:
  50. if param_name == 'engine':
  51. continue
  52. if param_name == 'categories':
  53. if engine_data['categories'] == 'none':
  54. engine.categories = []
  55. else:
  56. engine.categories = map(
  57. str.strip, engine_data['categories'].split(','))
  58. continue
  59. setattr(engine, param_name, engine_data[param_name])
  60. for arg_name, arg_value in engine_default_args.iteritems():
  61. if not hasattr(engine, arg_name):
  62. setattr(engine, arg_name, arg_value)
  63. if engine_data['name'] in languages:
  64. setattr(engine, 'supported_languages', languages[engine_data['name']])
  65. # checking required variables
  66. for engine_attr in dir(engine):
  67. if engine_attr.startswith('_'):
  68. continue
  69. if getattr(engine, engine_attr) is None:
  70. logger.error('Missing engine config attribute: "{0}.{1}"'
  71. .format(engine.name, engine_attr))
  72. sys.exit(1)
  73. engine.stats = {
  74. 'result_count': 0,
  75. 'search_count': 0,
  76. 'page_load_time': 0,
  77. 'page_load_count': 0,
  78. 'engine_time': 0,
  79. 'engine_time_count': 0,
  80. 'score_count': 0,
  81. 'errors': 0
  82. }
  83. for category_name in engine.categories:
  84. categories.setdefault(category_name, []).append(engine)
  85. if engine.shortcut in engine_shortcuts:
  86. logger.error('Engine config error: ambigious shortcut: {0}'.format(engine.shortcut))
  87. sys.exit(1)
  88. engine_shortcuts[engine.shortcut] = engine.name
  89. return engine
  90. def to_percentage(stats, maxvalue):
  91. for engine_stat in stats:
  92. if maxvalue:
  93. engine_stat['percentage'] = int(engine_stat['avg'] / maxvalue * 100)
  94. else:
  95. engine_stat['percentage'] = 0
  96. return stats
  97. def get_engines_stats():
  98. # TODO refactor
  99. pageloads = []
  100. engine_times = []
  101. results = []
  102. scores = []
  103. errors = []
  104. scores_per_result = []
  105. max_pageload = max_engine_times = max_results = max_score = max_errors = max_score_per_result = 0 # noqa
  106. for engine in engines.values():
  107. if engine.stats['search_count'] == 0:
  108. continue
  109. results_num = \
  110. engine.stats['result_count'] / float(engine.stats['search_count'])
  111. if engine.stats['page_load_count'] != 0:
  112. load_times = engine.stats['page_load_time'] / float(engine.stats['page_load_count']) # noqa
  113. else:
  114. load_times = 0
  115. if engine.stats['engine_time_count'] != 0:
  116. this_engine_time = engine.stats['engine_time'] / float(engine.stats['engine_time_count']) # noqa
  117. else:
  118. this_engine_time = 0
  119. if results_num:
  120. score = engine.stats['score_count'] / float(engine.stats['search_count']) # noqa
  121. score_per_result = score / results_num
  122. else:
  123. score = score_per_result = 0.0
  124. max_pageload = max(load_times, max_pageload)
  125. max_engine_times = max(this_engine_time, max_engine_times)
  126. max_results = max(results_num, max_results)
  127. max_score = max(score, max_score)
  128. max_score_per_result = max(score_per_result, max_score_per_result)
  129. max_errors = max(max_errors, engine.stats['errors'])
  130. pageloads.append({'avg': load_times, 'name': engine.name})
  131. engine_times.append({'avg': this_engine_time, 'name': engine.name})
  132. results.append({'avg': results_num, 'name': engine.name})
  133. scores.append({'avg': score, 'name': engine.name})
  134. errors.append({'avg': engine.stats['errors'], 'name': engine.name})
  135. scores_per_result.append({
  136. 'avg': score_per_result,
  137. 'name': engine.name
  138. })
  139. pageloads = to_percentage(pageloads, max_pageload)
  140. engine_times = to_percentage(engine_times, max_engine_times)
  141. results = to_percentage(results, max_results)
  142. scores = to_percentage(scores, max_score)
  143. scores_per_result = to_percentage(scores_per_result, max_score_per_result)
  144. erros = to_percentage(errors, max_errors)
  145. return [
  146. (
  147. gettext('Engine time (sec)'),
  148. sorted(engine_times, key=itemgetter('avg'))
  149. ),
  150. (
  151. gettext('Page loads (sec)'),
  152. sorted(pageloads, key=itemgetter('avg'))
  153. ),
  154. (
  155. gettext('Number of results'),
  156. sorted(results, key=itemgetter('avg'), reverse=True)
  157. ),
  158. (
  159. gettext('Scores'),
  160. sorted(scores, key=itemgetter('avg'), reverse=True)
  161. ),
  162. (
  163. gettext('Scores per result'),
  164. sorted(scores_per_result, key=itemgetter('avg'), reverse=True)
  165. ),
  166. (
  167. gettext('Errors'),
  168. sorted(errors, key=itemgetter('avg'), reverse=True)
  169. ),
  170. ]
  171. if 'engines' not in settings or not settings['engines']:
  172. logger.error('No engines found. Edit your settings.yml')
  173. exit(2)
  174. languages = loads(open(engine_dir + '/../data/engines_languages.json').read())
  175. for engine_data in settings['engines']:
  176. engine = load_engine(engine_data)
  177. if engine is not None:
  178. engines[engine.name] = engine