__init__.py 6.8 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212
  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. 'safesearch': False,
  31. 'timeout': settings['outgoing']['request_timeout'],
  32. 'shortcut': '-',
  33. 'disabled': False,
  34. 'suspend_end_time': 0,
  35. 'continuous_errors': 0,
  36. 'time_range_support': False}
  37. def load_engine(engine_data):
  38. if '_' in engine_data['name']:
  39. logger.error('Engine name conains underscore: "{}"'.format(engine_data['name']))
  40. sys.exit(1)
  41. engine_module = engine_data['engine']
  42. try:
  43. engine = load_module(engine_module + '.py', engine_dir)
  44. except:
  45. logger.exception('Cannot load engine "{}"'.format(engine_module))
  46. return None
  47. for param_name in engine_data:
  48. if param_name == 'engine':
  49. continue
  50. if param_name == 'categories':
  51. if engine_data['categories'] == 'none':
  52. engine.categories = []
  53. else:
  54. engine.categories = map(
  55. str.strip, engine_data['categories'].split(','))
  56. continue
  57. setattr(engine, param_name, engine_data[param_name])
  58. for arg_name, arg_value in engine_default_args.iteritems():
  59. if not hasattr(engine, arg_name):
  60. setattr(engine, arg_name, arg_value)
  61. # checking required variables
  62. for engine_attr in dir(engine):
  63. if engine_attr.startswith('_'):
  64. continue
  65. if getattr(engine, engine_attr) is None:
  66. logger.error('Missing engine config attribute: "{0}.{1}"'
  67. .format(engine.name, engine_attr))
  68. sys.exit(1)
  69. engine.stats = {
  70. 'result_count': 0,
  71. 'search_count': 0,
  72. 'page_load_time': 0,
  73. 'page_load_count': 0,
  74. 'engine_time': 0,
  75. 'engine_time_count': 0,
  76. 'score_count': 0,
  77. 'errors': 0
  78. }
  79. for category_name in engine.categories:
  80. categories.setdefault(category_name, []).append(engine)
  81. if engine.shortcut in engine_shortcuts:
  82. logger.error('Engine config error: ambigious shortcut: {0}'.format(engine.shortcut))
  83. sys.exit(1)
  84. engine_shortcuts[engine.shortcut] = engine.name
  85. return engine
  86. def to_percentage(stats, maxvalue):
  87. for engine_stat in stats:
  88. if maxvalue:
  89. engine_stat['percentage'] = int(engine_stat['avg'] / maxvalue * 100)
  90. else:
  91. engine_stat['percentage'] = 0
  92. return stats
  93. def get_engines_stats():
  94. # TODO refactor
  95. pageloads = []
  96. engine_times = []
  97. results = []
  98. scores = []
  99. errors = []
  100. scores_per_result = []
  101. max_pageload = max_engine_times = max_results = max_score = max_errors = max_score_per_result = 0 # noqa
  102. for engine in engines.values():
  103. if engine.stats['search_count'] == 0:
  104. continue
  105. results_num = \
  106. engine.stats['result_count'] / float(engine.stats['search_count'])
  107. if engine.stats['page_load_count'] != 0:
  108. load_times = engine.stats['page_load_time'] / float(engine.stats['page_load_count']) # noqa
  109. else:
  110. load_times = 0
  111. if engine.stats['engine_time_count'] != 0:
  112. this_engine_time = engine.stats['engine_time'] / float(engine.stats['engine_time_count']) # noqa
  113. else:
  114. this_engine_time = 0
  115. if results_num:
  116. score = engine.stats['score_count'] / float(engine.stats['search_count']) # noqa
  117. score_per_result = score / results_num
  118. else:
  119. score = score_per_result = 0.0
  120. max_pageload = max(load_times, max_pageload)
  121. max_engine_times = max(this_engine_time, max_engine_times)
  122. max_results = max(results_num, max_results)
  123. max_score = max(score, max_score)
  124. max_score_per_result = max(score_per_result, max_score_per_result)
  125. max_errors = max(max_errors, engine.stats['errors'])
  126. pageloads.append({'avg': load_times, 'name': engine.name})
  127. engine_times.append({'avg': this_engine_time, 'name': engine.name})
  128. results.append({'avg': results_num, 'name': engine.name})
  129. scores.append({'avg': score, 'name': engine.name})
  130. errors.append({'avg': engine.stats['errors'], 'name': engine.name})
  131. scores_per_result.append({
  132. 'avg': score_per_result,
  133. 'name': engine.name
  134. })
  135. pageloads = to_percentage(pageloads, max_pageload)
  136. engine_times = to_percentage(engine_times, max_engine_times)
  137. results = to_percentage(results, max_results)
  138. scores = to_percentage(scores, max_score)
  139. scores_per_result = to_percentage(scores_per_result, max_score_per_result)
  140. erros = to_percentage(errors, max_errors)
  141. return [
  142. (
  143. gettext('Engine time (sec)'),
  144. sorted(engine_times, key=itemgetter('avg'))
  145. ),
  146. (
  147. gettext('Page loads (sec)'),
  148. sorted(pageloads, key=itemgetter('avg'))
  149. ),
  150. (
  151. gettext('Number of results'),
  152. sorted(results, key=itemgetter('avg'), reverse=True)
  153. ),
  154. (
  155. gettext('Scores'),
  156. sorted(scores, key=itemgetter('avg'), reverse=True)
  157. ),
  158. (
  159. gettext('Scores per result'),
  160. sorted(scores_per_result, key=itemgetter('avg'), reverse=True)
  161. ),
  162. (
  163. gettext('Errors'),
  164. sorted(errors, key=itemgetter('avg'), reverse=True)
  165. ),
  166. ]
  167. if 'engines' not in settings or not settings['engines']:
  168. logger.error('No engines found. Edit your settings.yml')
  169. exit(2)
  170. for engine_data in settings['engines']:
  171. engine = load_engine(engine_data)
  172. if engine is not None:
  173. engines[engine.name] = engine