__init__.py 6.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202
  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.ext.babel import gettext
  18. from operator import itemgetter
  19. from searx import settings
  20. engine_dir = dirname(realpath(__file__))
  21. engines = {}
  22. categories = {'general': []}
  23. engine_shortcuts = {}
  24. def load_module(filename):
  25. modname = splitext(filename)[0]
  26. if modname in sys.modules:
  27. del sys.modules[modname]
  28. filepath = join(engine_dir, filename)
  29. module = load_source(modname, filepath)
  30. module.name = modname
  31. return module
  32. def load_engine(engine_data):
  33. engine_name = engine_data['engine']
  34. engine = load_module(engine_name + '.py')
  35. for param_name in engine_data:
  36. if param_name == 'engine':
  37. continue
  38. if param_name == 'categories':
  39. if engine_data['categories'] == 'none':
  40. engine.categories = []
  41. else:
  42. engine.categories = map(
  43. str.strip, engine_data['categories'].split(','))
  44. continue
  45. setattr(engine, param_name, engine_data[param_name])
  46. if not hasattr(engine, 'paging'):
  47. engine.paging = False
  48. if not hasattr(engine, 'categories'):
  49. engine.categories = ['general']
  50. if not hasattr(engine, 'language_support'):
  51. # engine.language_support = False
  52. engine.language_support = True
  53. if not hasattr(engine, 'timeout'):
  54. # engine.language_support = False
  55. engine.timeout = settings['server']['request_timeout']
  56. if not hasattr(engine, 'shortcut'):
  57. # engine.shortcut = '''
  58. engine.shortcut = ''
  59. # checking required variables
  60. for engine_attr in dir(engine):
  61. if engine_attr.startswith('_'):
  62. continue
  63. if getattr(engine, engine_attr) is None:
  64. print '[E] Engine config error: Missing attribute "{0}.{1}"'.format(engine.name, engine_attr) # noqa
  65. sys.exit(1)
  66. engine.stats = {
  67. 'result_count': 0,
  68. 'search_count': 0,
  69. 'page_load_time': 0,
  70. 'score_count': 0,
  71. 'errors': 0
  72. }
  73. if hasattr(engine, 'categories'):
  74. for category_name in engine.categories:
  75. categories.setdefault(category_name, []).append(engine)
  76. else:
  77. categories['general'].append(engine)
  78. if engine.shortcut:
  79. # TODO check duplications
  80. engine_shortcuts[engine.shortcut] = engine.name
  81. return engine
  82. def get_engines_stats():
  83. # TODO refactor
  84. pageloads = []
  85. results = []
  86. scores = []
  87. errors = []
  88. scores_per_result = []
  89. max_pageload = max_results = max_score = max_errors = max_score_per_result = 0 # noqa
  90. for engine in engines.values():
  91. if engine.stats['search_count'] == 0:
  92. continue
  93. results_num = \
  94. engine.stats['result_count'] / float(engine.stats['search_count'])
  95. load_times = engine.stats['page_load_time'] / float(engine.stats['search_count']) # noqa
  96. if results_num:
  97. score = engine.stats['score_count'] / float(engine.stats['search_count']) # noqa
  98. score_per_result = score / results_num
  99. else:
  100. score = score_per_result = 0.0
  101. max_results = max(results_num, max_results)
  102. max_pageload = max(load_times, max_pageload)
  103. max_score = max(score, max_score)
  104. max_score_per_result = max(score_per_result, max_score_per_result)
  105. max_errors = max(max_errors, engine.stats['errors'])
  106. pageloads.append({'avg': load_times, 'name': engine.name})
  107. results.append({'avg': results_num, 'name': engine.name})
  108. scores.append({'avg': score, 'name': engine.name})
  109. errors.append({'avg': engine.stats['errors'], 'name': engine.name})
  110. scores_per_result.append({
  111. 'avg': score_per_result,
  112. 'name': engine.name
  113. })
  114. for engine in pageloads:
  115. if max_pageload:
  116. engine['percentage'] = int(engine['avg'] / max_pageload * 100)
  117. else:
  118. engine['percentage'] = 0
  119. for engine in results:
  120. if max_results:
  121. engine['percentage'] = int(engine['avg'] / max_results * 100)
  122. else:
  123. engine['percentage'] = 0
  124. for engine in scores:
  125. if max_score:
  126. engine['percentage'] = int(engine['avg'] / max_score * 100)
  127. else:
  128. engine['percentage'] = 0
  129. for engine in scores_per_result:
  130. if max_score_per_result:
  131. engine['percentage'] = int(engine['avg']
  132. / max_score_per_result * 100)
  133. else:
  134. engine['percentage'] = 0
  135. for engine in errors:
  136. if max_errors:
  137. engine['percentage'] = int(float(engine['avg']) / max_errors * 100)
  138. else:
  139. engine['percentage'] = 0
  140. return [
  141. (
  142. gettext('Page loads (sec)'),
  143. sorted(pageloads, key=itemgetter('avg'))
  144. ),
  145. (
  146. gettext('Number of results'),
  147. sorted(results, key=itemgetter('avg'), reverse=True)
  148. ),
  149. (
  150. gettext('Scores'),
  151. sorted(scores, key=itemgetter('avg'), reverse=True)
  152. ),
  153. (
  154. gettext('Scores per result'),
  155. sorted(scores_per_result, key=itemgetter('avg'), reverse=True)
  156. ),
  157. (
  158. gettext('Errors'),
  159. sorted(errors, key=itemgetter('avg'), reverse=True)
  160. ),
  161. ]
  162. if 'engines' not in settings or not settings['engines']:
  163. print '[E] Error no engines found. Edit your settings.yml'
  164. exit(2)
  165. for engine_data in settings['engines']:
  166. engine = load_engine(engine_data)
  167. engines[engine.name] = engine