A framework for quick web archiving
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.

89 lines
3.1 KiB

  1. import argparse
  2. import asyncio
  3. import importlib.util
  4. import logging
  5. import os.path
  6. import qwarc
  7. import qwarc.utils
  8. import qwarc.version
  9. import sys
  10. def setup_logging(logFilename):
  11. rootLogger = logging.getLogger()
  12. rootLogger.handlers = []
  13. rootLogger.setLevel(logging.INFO)
  14. formatter = qwarc.utils.LogFormatter()
  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(f'Error: "{specFilename}" does not exist or is not a regular file', file = sys.stderr)
  25. success = False
  26. if os.path.exists(logFilename):
  27. print(f'Error: "{logFilename}" already exists', 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('--version', action = 'version', version = f'qwarc {qwarc.version.__version__}')
  36. parser.add_argument('--log', metavar = 'LOGFILE', default = './qwarc.log')
  37. parser.add_argument('--database', metavar = 'DBFILE', default = './qwarc.db')
  38. parser.add_argument('--warc', metavar = 'PREFIX', help = 'prefix for the WARC filenames', default = './qwarc')
  39. parser.add_argument('--concurrency', type = int, default = 1)
  40. parser.add_argument('--memorylimit', metavar = 'LIMIT', type = int, help = 'pause when less than LIMIT bytes memory is free; disable if 0', default = 0)
  41. parser.add_argument('--disklimit', metavar = 'LIMIT', type = int, help = 'pause when less than LIMIT bytes disk space is free; disable if 0', default = 0)
  42. parser.add_argument('--warcsplit', metavar = 'SIZE', type = int, help = 'split WARCs into files of SIZE bytes; disable if 0', default = 0)
  43. parser.add_argument('--warcdedupe', action = 'store_true', help = 'enable deduplication of WARC records')
  44. parser.add_argument('specfile')
  45. args = parser.parse_args()
  46. if not check_files(args.specfile, args.log):
  47. sys.exit(1)
  48. setup_logging(args.log)
  49. spec = importlib.util.spec_from_file_location('spec', args.specfile)
  50. specMod = importlib.util.module_from_spec(spec)
  51. spec.loader.exec_module(specMod)
  52. specDependencies = specMod.__dict__.get('specDependencies', qwarc.utils.SpecDependencies())
  53. a = qwarc.QWARC(
  54. itemClasses = set(qwarc.Item.get_subclasses()),
  55. warcBasePath = args.warc,
  56. dbPath = args.database,
  57. command = sys.argv,
  58. specFile = args.specfile,
  59. specDependencies = specDependencies,
  60. logFilename = args.log,
  61. concurrency = args.concurrency,
  62. memoryLimit = args.memorylimit,
  63. minFreeDisk = args.disklimit,
  64. warcSizeLimit = args.warcsplit,
  65. warcDedupe = args.warcdedupe,
  66. )
  67. loop = asyncio.get_event_loop()
  68. try:
  69. loop.run_until_complete(a.run(loop))
  70. except (Exception, KeyboardInterrupt) as e:
  71. logging.exception('Unhandled error')
  72. finally:
  73. loop.run_until_complete(a.close())
  74. loop.close()