background.py 3.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122
  1. # SPDX-License-Identifier: AGPL-3.0-or-later
  2. import json
  3. import random
  4. import time
  5. import threading
  6. import os
  7. import signal
  8. from searx import logger, settings, searx_debug
  9. from searx.exceptions import SearxSettingsException
  10. from searx.search.processors import processors
  11. from searx.search.checker import Checker
  12. from searx.shared import schedule, storage
  13. CHECKER_RESULT = 'CHECKER_RESULT'
  14. running = threading.Lock()
  15. def _get_interval(every, error_msg):
  16. if isinstance(every, int):
  17. every = (every, every)
  18. if not isinstance(every, (tuple, list))\
  19. or len(every) != 2\
  20. or not isinstance(every[0], int)\
  21. or not isinstance(every[1], int):
  22. raise SearxSettingsException(error_msg, None)
  23. return every
  24. def _get_every():
  25. every = settings.get('checker', {}).get('scheduling', {}).get('every', (300, 1800))
  26. return _get_interval(every, 'checker.scheduling.every is not a int or list')
  27. def get_result():
  28. serialized_result = storage.get_str(CHECKER_RESULT)
  29. if serialized_result is not None:
  30. return json.loads(serialized_result)
  31. def _set_result(result):
  32. result['timestamp'] = int(time.time() / 3600) * 3600
  33. storage.set_str(CHECKER_RESULT, json.dumps(result))
  34. def run():
  35. if not running.acquire(blocking=False):
  36. return
  37. try:
  38. logger.info('Starting checker')
  39. result = {
  40. 'status': 'ok',
  41. 'engines': {}
  42. }
  43. for name, processor in processors.items():
  44. logger.debug('Checking %s engine', name)
  45. checker = Checker(processor)
  46. checker.run()
  47. if checker.test_results.succesfull:
  48. result['engines'][name] = {'success': True}
  49. else:
  50. result['engines'][name] = {'success': False, 'errors': checker.test_results.errors}
  51. _set_result(result)
  52. logger.info('Check done')
  53. except Exception:
  54. _set_result({'status': 'error'})
  55. logger.exception('Error while running the checker')
  56. finally:
  57. running.release()
  58. def _run_with_delay():
  59. every = _get_every()
  60. delay = random.randint(0, every[1] - every[0])
  61. logger.debug('Start checker in %i seconds', delay)
  62. time.sleep(delay)
  63. run()
  64. def _start_scheduling():
  65. every = _get_every()
  66. schedule(every[0], _run_with_delay)
  67. run()
  68. def _signal_handler(signum, frame):
  69. t = threading.Thread(target=run)
  70. t.daemon = True
  71. t.start()
  72. def initialize():
  73. logger.info('Send SIGUSR1 signal to pid %i to start the checker', os.getpid())
  74. signal.signal(signal.SIGUSR1, _signal_handler)
  75. # disabled by default
  76. _set_result({'status': 'disabled'})
  77. # special case when debug is activate
  78. if searx_debug and settings.get('checker', {}).get('off_when_debug', True):
  79. logger.info('debug mode: checker is disabled')
  80. return
  81. # check value of checker.scheduling.every now
  82. scheduling = settings.get('checker', {}).get('scheduling', None)
  83. if scheduling is None or not scheduling:
  84. logger.info('Checker scheduler is disabled')
  85. return
  86. #
  87. _set_result({'status': 'unknown'})
  88. start_after = scheduling.get('start_after', (300, 1800))
  89. start_after = _get_interval(start_after, 'checker.scheduling.start_after is not a int or list')
  90. delay = random.randint(start_after[0], start_after[1])
  91. logger.info('Start checker in %i seconds', delay)
  92. t = threading.Timer(delay, _start_scheduling)
  93. t.daemon = True
  94. t.start()