def load_test_archive(test_file, offset, length):
path = utils.test_data_dir() + 'warcs/' + test_file
archive = testloader.load(path, offset, length)
pprint.pprint((archive.type, archive.rec_headers, archive.status_headers))
self[header] = field
self.cdxline = cdxline
def __setitem__(self, key, value):
OrderedDict.__setitem__(self, key, value)
# force regen on next __str__ call
self.cdxline = None
def __str__(self):
if self.cdxline:
return self.cdxline
li = itertools.imap(lambda (n, val): val, self.items())
return ' '.join(li)
# Testing
import utils
if __name__ == "__main__" or utils.enable_doctests():
from pprint import pprint
test_dir = utils.test_data_dir() + 'cdx/'
import doctest
doctest.testmod()
def create_test_cdx(test_file):
path = utils.test_data_dir() + 'cdx/' + test_file
return FileReader(path)