__init__.py 6.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214
  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, splitext, join
  15. import sys
  16. from imp import load_source
  17. from flask_babel import gettext
  18. from operator import itemgetter
  19. from searx import settings
  20. from searx import logger
  21. logger = logger.getChild('engines')
  22. engine_dir = dirname(realpath(__file__))
  23. engines = {}
  24. categories = {'general': []}
  25. engine_shortcuts = {}
  26. engine_default_args = {'paging': False,
  27. 'categories': ['general'],
  28. 'language_support': True,
  29. 'safesearch': False,
  30. 'timeout': settings['outgoing']['request_timeout'],
  31. 'shortcut': '-',
  32. 'disabled': False,
  33. 'suspend_end_time': 0,
  34. 'continuous_errors': 0,
  35. 'time_range_support': False}
  36. def load_module(filename):
  37. modname = splitext(filename)[0]
  38. if modname in sys.modules:
  39. del sys.modules[modname]
  40. filepath = join(engine_dir, filename)
  41. module = load_source(modname, filepath)
  42. module.name = modname
  43. return module
  44. def load_engine(engine_data):
  45. if '_' in engine_data['name']:
  46. logger.error('Engine name conains underscore: "{}"'.format(engine_data['name']))
  47. sys.exit(1)
  48. engine_module = engine_data['engine']
  49. try:
  50. engine = load_module(engine_module + '.py')
  51. except:
  52. logger.exception('Cannot load engine "{}"'.format(engine_module))
  53. return None
  54. for param_name in engine_data:
  55. if param_name == 'engine':
  56. continue
  57. if param_name == 'categories':
  58. if engine_data['categories'] == 'none':
  59. engine.categories = []
  60. else:
  61. engine.categories = map(
  62. str.strip, engine_data['categories'].split(','))
  63. continue
  64. setattr(engine, param_name, engine_data[param_name])
  65. for arg_name, arg_value in engine_default_args.iteritems():
  66. if not hasattr(engine, arg_name):
  67. setattr(engine, arg_name, arg_value)
  68. # checking required variables
  69. for engine_attr in dir(engine):
  70. if engine_attr.startswith('_'):
  71. continue
  72. if getattr(engine, engine_attr) is None:
  73. logger.error('Missing engine config attribute: "{0}.{1}"'
  74. .format(engine.name, engine_attr))
  75. sys.exit(1)
  76. engine.stats = {
  77. 'result_count': 0,
  78. 'search_count': 0,
  79. 'page_load_time': 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 get_engines_stats():
  91. # TODO refactor
  92. pageloads = []
  93. results = []
  94. scores = []
  95. errors = []
  96. scores_per_result = []
  97. max_pageload = max_results = max_score = max_errors = max_score_per_result = 0 # noqa
  98. for engine in engines.values():
  99. if engine.stats['search_count'] == 0:
  100. continue
  101. results_num = \
  102. engine.stats['result_count'] / float(engine.stats['search_count'])
  103. load_times = engine.stats['page_load_time'] / float(engine.stats['search_count']) # noqa
  104. if results_num:
  105. score = engine.stats['score_count'] / float(engine.stats['search_count']) # noqa
  106. score_per_result = score / results_num
  107. else:
  108. score = score_per_result = 0.0
  109. max_results = max(results_num, max_results)
  110. max_pageload = max(load_times, max_pageload)
  111. max_score = max(score, max_score)
  112. max_score_per_result = max(score_per_result, max_score_per_result)
  113. max_errors = max(max_errors, engine.stats['errors'])
  114. pageloads.append({'avg': load_times, 'name': engine.name})
  115. results.append({'avg': results_num, 'name': engine.name})
  116. scores.append({'avg': score, 'name': engine.name})
  117. errors.append({'avg': engine.stats['errors'], 'name': engine.name})
  118. scores_per_result.append({
  119. 'avg': score_per_result,
  120. 'name': engine.name
  121. })
  122. for engine in pageloads:
  123. if max_pageload:
  124. engine['percentage'] = int(engine['avg'] / max_pageload * 100)
  125. else:
  126. engine['percentage'] = 0
  127. for engine in results:
  128. if max_results:
  129. engine['percentage'] = int(engine['avg'] / max_results * 100)
  130. else:
  131. engine['percentage'] = 0
  132. for engine in scores:
  133. if max_score:
  134. engine['percentage'] = int(engine['avg'] / max_score * 100)
  135. else:
  136. engine['percentage'] = 0
  137. for engine in scores_per_result:
  138. if max_score_per_result:
  139. engine['percentage'] = int(engine['avg']
  140. / max_score_per_result * 100)
  141. else:
  142. engine['percentage'] = 0
  143. for engine in errors:
  144. if max_errors:
  145. engine['percentage'] = int(float(engine['avg']) / max_errors * 100)
  146. else:
  147. engine['percentage'] = 0
  148. return [
  149. (
  150. gettext('Page loads (sec)'),
  151. sorted(pageloads, key=itemgetter('avg'))
  152. ),
  153. (
  154. gettext('Number of results'),
  155. sorted(results, key=itemgetter('avg'), reverse=True)
  156. ),
  157. (
  158. gettext('Scores'),
  159. sorted(scores, key=itemgetter('avg'), reverse=True)
  160. ),
  161. (
  162. gettext('Scores per result'),
  163. sorted(scores_per_result, key=itemgetter('avg'), reverse=True)
  164. ),
  165. (
  166. gettext('Errors'),
  167. sorted(errors, key=itemgetter('avg'), reverse=True)
  168. ),
  169. ]
  170. if 'engines' not in settings or not settings['engines']:
  171. logger.error('No engines found. Edit your settings.yml')
  172. exit(2)
  173. for engine_data in settings['engines']:
  174. engine = load_engine(engine_data)
  175. if engine is not None:
  176. engines[engine.name] = engine