def _limit_corpus_size(corpus_url, size_limit):
"""Limit number of files in a corpus url."""
files_list = list(storage.list_blobs(corpus_url))
corpus_size = len(files_list)
if corpus_size <= size_limit:
# Corpus directory size is within limit, no more work to do.
return
logs.log(
'Limit corpus at {corpus_url} from {corpus_size} to {size_limit}.'.format(
corpus_url=corpus_url, corpus_size=corpus_size,
size_limit=size_limit))
files_to_delete = random.sample(files_list, corpus_size - size_limit)
bucket, _ = storage.get_bucket_name_and_path(corpus_url)
for file_to_delete in files_to_delete:
path_to_delete = storage.get_cloud_storage_file_path(bucket, file_to_delete)
storage.delete(path_to_delete)
def _limit_corpus_size(corpus_url):
"""Limit number of files and size of a corpus."""
corpus_count = 0
corpus_size = 0
deleted_corpus_count = 0
bucket, _ = storage.get_bucket_name_and_path(corpus_url)
for corpus_file in storage.get_blobs(corpus_url):
corpus_count += 1
corpus_size += corpus_file['size']
if (corpus_count > CORPUS_FILES_LIMIT_FOR_FAILURES
or corpus_size > CORPUS_SIZE_LIMIT_FOR_FAILURES):
path_to_delete = storage.get_cloud_storage_file_path(
bucket, corpus_file['name'])
storage.delete(path_to_delete)
deleted_corpus_count += 1
if deleted_corpus_count:
logs.log('Removed %d files from oversized corpus: %s.' %
(deleted_corpus_count, corpus_url))