A semantic search engine for source code https://bitshift.benkurtovic.com/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

99 lines
2.8 KiB

  1. """
  2. :synopsis: Parent crawler module, which supervises all crawlers.
  3. Contains functions for initializing all subsidiary, threaded crawlers.
  4. """
  5. import logging
  6. import logging.handlers
  7. import os
  8. import Queue
  9. import sys
  10. import time
  11. from threading import Event
  12. from .crawler import GitHubCrawler, BitbucketCrawler
  13. from .indexer import GitIndexer, GitRepository
  14. from ..parser import start_parse_servers
  15. __all__ = ["crawl"]
  16. MAX_URL_QUEUE_SIZE = 5e3
  17. def crawl():
  18. """
  19. Initialize all crawlers (and indexers).
  20. Start the:
  21. 1. GitHub crawler, :class:`crawler.GitHubCrawler`.
  22. 2. Bitbucket crawler, :class:`crawler.BitbucketCrawler`.
  23. 3. Git indexer, :class:`bitshift.crawler.indexer.GitIndexer`.
  24. """
  25. _configure_logging()
  26. parse_servers = start_parse_servers()
  27. time.sleep(5)
  28. repo_clone_queue = Queue.Queue(maxsize=MAX_URL_QUEUE_SIZE)
  29. run_event = Event()
  30. run_event.set()
  31. threads = [GitIndexer(repo_clone_queue, run_event)]
  32. if sys.argv[1:]:
  33. names = sys.argv[1:]
  34. ranks = GitHubCrawler.get_ranks(names)
  35. for name in names:
  36. repo = GitRepository("https://github.com/" + name, name, "GitHub",
  37. ranks[name])
  38. repo_clone_queue.put(repo)
  39. else:
  40. threads += [GitHubCrawler(repo_clone_queue, run_event),
  41. BitbucketCrawler(repo_clone_queue, run_event)]
  42. for thread in threads:
  43. thread.start()
  44. try:
  45. while 1:
  46. time.sleep(0.1)
  47. except KeyboardInterrupt:
  48. run_event.clear()
  49. with repo_clone_queue.mutex:
  50. repo_clone_queue.queue.clear()
  51. for thread in threads:
  52. thread.join()
  53. for server in parse_servers:
  54. server.kill()
  55. def _configure_logging():
  56. # This isn't ideal, since it means the bitshift python package must be kept
  57. # inside the app, but it works for now:
  58. root = os.path.abspath(os.path.join(os.path.dirname(__file__), "..", ".."))
  59. log_dir = os.path.join(root, "logs")
  60. if not os.path.exists(log_dir):
  61. os.mkdir(log_dir)
  62. logging.getLogger("requests").setLevel(logging.WARNING)
  63. logging.getLogger("urllib3").setLevel(logging.WARNING)
  64. formatter = logging.Formatter(
  65. fmt=("%(asctime)s %(levelname)s %(name)s:%(funcName)s"
  66. " %(message)s"), datefmt="%y-%m-%d %H:%M:%S")
  67. file_handler = logging.handlers.TimedRotatingFileHandler(
  68. "%s/%s" % (log_dir, "app.log"), when="H", interval=1,
  69. backupCount=20)
  70. stream_handler = logging.StreamHandler()
  71. file_handler.setFormatter(formatter)
  72. stream_handler.setFormatter(formatter)
  73. root_logger = logging.getLogger()
  74. root_logger.handlers = []
  75. root_logger.addHandler(file_handler)
  76. root_logger.addHandler(stream_handler)
  77. root_logger.setLevel(logging.NOTSET)
  78. if __name__ == "__main__":
  79. crawl()