A framework for quick web archiving
Você não pode selecionar mais de 25 tópicos Os tópicos devem começar com uma letra ou um número, podem incluir traços ('-') e podem ter até 35 caracteres.

82 linhas
2.7 KiB

  1. import argparse
  2. import asyncio
  3. import importlib.util
  4. import logging
  5. import os.path
  6. import qwarc
  7. import sys
  8. import time
  9. def setup_logging(logFilename):
  10. rootLogger = logging.getLogger()
  11. rootLogger.handlers = []
  12. rootLogger.setLevel(logging.INFO)
  13. formatter = logging.Formatter('%(asctime)s.%(msecs)03dZ %(levelname)s %(message)s', datefmt = '%Y-%m-%d %H:%M:%S')
  14. formatter.converter = time.gmtime
  15. fileHandler = logging.FileHandler(logFilename)
  16. fileHandler.setFormatter(formatter)
  17. rootLogger.addHandler(fileHandler)
  18. stderrHandler = logging.StreamHandler()
  19. stderrHandler.setFormatter(formatter)
  20. rootLogger.addHandler(stderrHandler)
  21. def check_files(specFilename, logFilename):
  22. success = True
  23. if not os.path.isfile(specFilename):
  24. print('Error: "{}" does not exist or is not a regular file', file = sys.stderr)
  25. success = False
  26. if os.path.exists(logFilename):
  27. print('Error: "{}" already exists'.format(logFilename), file = sys.stderr)
  28. success = False
  29. if os.path.exists('STOP'):
  30. print('Error: "STOP" exists', file = sys.stderr)
  31. success = False
  32. return success
  33. def main():
  34. parser = argparse.ArgumentParser(formatter_class = argparse.ArgumentDefaultsHelpFormatter)
  35. parser.add_argument('--log', metavar = 'LOGFILE', default = './qwarc.log')
  36. parser.add_argument('--database', metavar = 'DBFILE', default = './qwarc.db')
  37. parser.add_argument('--warc', metavar = 'PREFIX', help = 'prefix for the WARC filenames', default = './qwarc')
  38. parser.add_argument('--concurrency', type = int, default = 1)
  39. parser.add_argument('--memorylimit', metavar = 'LIMIT', help = 'pause when less than LIMIT bytes memory is free; disable if 0', default = 0)
  40. parser.add_argument('--disklimit', metavar = 'LIMIT', help = 'pause when less than LIMIT bytes disk space is free; disable if 0', default = 0)
  41. parser.add_argument('--warcsplit', metavar = 'SIZE', help = 'split WARCs into files of SIZE bytes; disable if 0', default = 0)
  42. parser.add_argument('specfile')
  43. args = parser.parse_args()
  44. if not check_files(args.specfile, args.log):
  45. sys.exit(1)
  46. setup_logging(args.log)
  47. spec = importlib.util.spec_from_file_location('spec', args.specfile)
  48. specMod = importlib.util.module_from_spec(spec)
  49. spec.loader.exec_module(specMod)
  50. a = qwarc.QWARC(
  51. itemClasses = qwarc.Item.__subclasses__(),
  52. warcBasePath = args.warc,
  53. dbPath = args.database,
  54. concurrency = args.concurrency,
  55. memoryLimit = args.memorylimit,
  56. minFreeDisk = args.disklimit,
  57. warcSizeLimit = args.warcsplit,
  58. )
  59. if not os.path.exists(args.database):
  60. a.create_db()
  61. loop = asyncio.get_event_loop()
  62. try:
  63. loop.run_until_complete(a.run(loop))
  64. except (Exception, KeyboardInterrupt) as e:
  65. logging.exception('Unhandled error')
  66. loop.close()