def predict(self, x_test):
desktops = Repository.get_data('desktops')
mobiles_sample = x_test
merge_key = 'anonymous_c2'
mobiles_sample = mobiles_sample[['device_id', merge_key]].set_index(
[merge_key, 'device_id'])
desktops_sample = desktops[['cookie_id', merge_key]].set_index([merge_key])
merge_data = mobiles_sample.join(desktops_sample).reset_index(1)
return merge_data
def regenerate_repodata(repository_path, marked_repository_path):
"""
Re-generates the repodata for the given repository.
@param repository_path The path to the repository.
Uses group.xml and patterns.xml from any path inside repository, if these
files don't exist they're unpacked from package-groups.rpm
"""
repository = Repository(repository_path)
repodata = repository.get_data()
if repodata.groups_data is None:
logging.warning("There is no groups data in "
"{0}".format(repository_path))
if repodata.patterns_data is None:
logging.warning("There is no patterns data in "
"{0}".format(repository_path))
repository.generate_derived_data()
marked_repository = Repository(marked_repository_path)
marked_repository.set_data(repodata)
marked_repository.generate_derived_data()
def create_submission(self):
submission = Repository.get_data('submission')
time_stamp = datetime.datetime.now().strftime("%Y_%M_%d_%H_%m")
file_name = '../output/{}-submission-{}.csv'.format(time_stamp, self.description)
submission.to_csv(file_name, header=True, index=False)
print "Submission file", file_name, "is created."
