__init__.py 6.9 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213
  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 searx import settings
  19. from searx import logger
  20. from searx.utils import load_module
  21. logger = logger.getChild('engines')
  22. engine_dir = dirname(realpath(__file__))
  23. engines = {}
  24. categories = {'general': []}
  25. _initialized = False
  26. engine_shortcuts = {}
  27. engine_default_args = {'paging': False,
  28. 'categories': ['general'],
  29. 'language_support': True,
  30. 'supported_languages': [],
  31. 'safesearch': False,
  32. 'timeout': settings['outgoing']['request_timeout'],
  33. 'shortcut': '-',
  34. 'disabled': False,
  35. 'suspend_end_time': 0,
  36. 'continuous_errors': 0,
  37. 'time_range_support': False}
  38. def load_engine(engine_data):
  39. if '_' in engine_data['name']:
  40. logger.error('Engine name conains underscore: "{}"'.format(engine_data['name']))
  41. sys.exit(1)
  42. engine_module = engine_data['engine']
  43. try:
  44. engine = load_module(engine_module + '.py', engine_dir)
  45. except:
  46. logger.exception('Cannot load engine "{}"'.format(engine_module))
  47. return None
  48. for param_name in engine_data:
  49. if param_name == 'engine':
  50. continue
  51. if param_name == 'categories':
  52. if engine_data['categories'] == 'none':
  53. engine.categories = []
  54. else:
  55. engine.categories = map(
  56. str.strip, engine_data['categories'].split(','))
  57. continue
  58. setattr(engine, param_name, engine_data[param_name])
  59. for arg_name, arg_value in engine_default_args.iteritems():
  60. if not hasattr(engine, arg_name):
  61. setattr(engine, arg_name, arg_value)
  62. # checking required variables
  63. for engine_attr in dir(engine):
  64. if engine_attr.startswith('_'):
  65. continue
  66. if getattr(engine, engine_attr) is None:
  67. logger.error('Missing engine config attribute: "{0}.{1}"'
  68. .format(engine.name, engine_attr))
  69. sys.exit(1)
  70. engine.stats = {
  71. 'result_count': 0,
  72. 'search_count': 0,
  73. 'page_load_time': 0,
  74. 'page_load_count': 0,
  75. 'engine_time': 0,
  76. 'engine_time_count': 0,
  77. 'score_count': 0,
  78. 'errors': 0
  79. }
  80. for category_name in engine.categories:
  81. categories.setdefault(category_name, []).append(engine)
  82. if engine.shortcut in engine_shortcuts:
  83. logger.error('Engine config error: ambigious shortcut: {0}'.format(engine.shortcut))
  84. sys.exit(1)
  85. engine_shortcuts[engine.shortcut] = engine.name
  86. return engine
  87. def to_percentage(stats, maxvalue):
  88. for engine_stat in stats:
  89. if maxvalue:
  90. engine_stat['percentage'] = int(engine_stat['avg'] / maxvalue * 100)
  91. else:
  92. engine_stat['percentage'] = 0
  93. return stats
  94. def get_engines_stats():
  95. # TODO refactor
  96. pageloads = []
  97. engine_times = []
  98. results = []
  99. scores = []
  100. errors = []
  101. scores_per_result = []
  102. max_pageload = max_engine_times = max_results = max_score = max_errors = max_score_per_result = 0 # noqa
  103. for engine in engines.values():
  104. if engine.stats['search_count'] == 0:
  105. continue
  106. results_num = \
  107. engine.stats['result_count'] / float(engine.stats['search_count'])
  108. if engine.stats['page_load_count'] != 0:
  109. load_times = engine.stats['page_load_time'] / float(engine.stats['page_load_count']) # noqa
  110. else:
  111. load_times = 0
  112. if engine.stats['engine_time_count'] != 0:
  113. this_engine_time = engine.stats['engine_time'] / float(engine.stats['engine_time_count']) # noqa
  114. else:
  115. this_engine_time = 0
  116. if results_num:
  117. score = engine.stats['score_count'] / float(engine.stats['search_count']) # noqa
  118. score_per_result = score / results_num
  119. else:
  120. score = score_per_result = 0.0
  121. max_pageload = max(load_times, max_pageload)
  122. max_engine_times = max(this_engine_time, max_engine_times)
  123. max_results = max(results_num, max_results)
  124. max_score = max(score, max_score)
  125. max_score_per_result = max(score_per_result, max_score_per_result)
  126. max_errors = max(max_errors, engine.stats['errors'])
  127. pageloads.append({'avg': load_times, 'name': engine.name})
  128. engine_times.append({'avg': this_engine_time, 'name': engine.name})
  129. results.append({'avg': results_num, 'name': engine.name})
  130. scores.append({'avg': score, 'name': engine.name})
  131. errors.append({'avg': engine.stats['errors'], 'name': engine.name})
  132. scores_per_result.append({
  133. 'avg': score_per_result,
  134. 'name': engine.name
  135. })
  136. pageloads = to_percentage(pageloads, max_pageload)
  137. engine_times = to_percentage(engine_times, max_engine_times)
  138. results = to_percentage(results, max_results)
  139. scores = to_percentage(scores, max_score)
  140. scores_per_result = to_percentage(scores_per_result, max_score_per_result)
  141. erros = to_percentage(errors, max_errors)
  142. return [
  143. (
  144. gettext('Engine time (sec)'),
  145. sorted(engine_times, key=itemgetter('avg'))
  146. ),
  147. (
  148. gettext('Page loads (sec)'),
  149. sorted(pageloads, key=itemgetter('avg'))
  150. ),
  151. (
  152. gettext('Number of results'),
  153. sorted(results, key=itemgetter('avg'), reverse=True)
  154. ),
  155. (
  156. gettext('Scores'),
  157. sorted(scores, key=itemgetter('avg'), reverse=True)
  158. ),
  159. (
  160. gettext('Scores per result'),
  161. sorted(scores_per_result, key=itemgetter('avg'), reverse=True)
  162. ),
  163. (
  164. gettext('Errors'),
  165. sorted(errors, key=itemgetter('avg'), reverse=True)
  166. ),
  167. ]
  168. if 'engines' not in settings or not settings['engines']:
  169. logger.error('No engines found. Edit your settings.yml')
  170. exit(2)
  171. for engine_data in settings['engines']:
  172. engine = load_engine(engine_data)
  173. if engine is not None:
  174. engines[engine.name] = engine