def test_file_path():
"""
Test S3IterableDataset for existing and nonexistent path
"""
# existing path
s3_path = 's3://pt-s3plugin-test-data-west2/images/test'
s3_dataset = S3IterableDataset(s3_path)
assert s3_dataset
# non-existent path
s3_path_none = 's3://pt-s3plugin-test-data-west2/non_existent_path/test'
with pytest.raises(AssertionError) as excinfo:
s3_dataset = S3IterableDataset(s3_path_none)
assert 'does not contain any objects' in str(excinfo.value)
def test_urls_list():
"""
Test whether urls_list input for S3IterableDataset works properly
"""
os.environ['AWS_REGION'] = 'us-west-2'
# provide url prefix (path within bucket)
prefix_to_directory = 'images/test'
prefix_to_file = 'test_1.JPEG'
prefix_list = [prefix_to_directory, prefix_to_file]
# set up boto3
s3 = boto3.resource('s3')
bucket_name = 'pt-s3plugin-test-data-west2'
test_bucket = s3.Bucket(bucket_name)
# try individual valid urls and collect url_list and all_boto3_files to test url list input
urls_list = list()
all_boto3_files = list()
for prefix in prefix_list:
# collect list of all file names using S3IterableDataset
url = os.path.join('s3://', bucket_name, prefix)
urls_list.append(url)
s3_dataset = S3IterableDataset(url)
s3_files = [item[0] for item in s3_dataset]
# collect list of all file names using boto3
boto3_files = [os.path.join('s3://', url.bucket_name, url.key) \
for url in test_bucket.objects.filter(Prefix=prefix)]
all_boto3_files.extend(boto3_files)
assert s3_files == boto3_files
# test list of two valid urls as input
s3_dataset = S3IterableDataset(urls_list)
s3_files = [item[0] for item in s3_dataset]
assert s3_files == all_boto3_files
# add an non-existent url to list of urls
url_to_non_existent = 's3://pt-s3plugin-test-data-west2/non_existent_directory'
urls_list.append(url_to_non_existent)
with pytest.raises(AssertionError) as excinfo:
s3_dataset = S3IterableDataset(urls_list)
assert 'does not contain any objects' in str(excinfo.value)
del os.environ['AWS_REGION']
def test_zip_file_s3iterabledataset():
s3_dataset_path = 's3://pt-s3plugin-test-data-west2/tiny-imagenet-200.zip'
dataset = S3IterableDataset(s3_dataset_path)
list_of_files = []
for files in dataset:
list_of_files.append(files[0][0])
result1 = len(list_of_files)
result2 = get_zip(s3_dataset_path)
assert result1 == len(result2)
def test_csv_file_s3iterabledataset():
os.environ['AWS_REGION'] = 'us-east-1'
s3_dataset_path = 's3://pt-s3plugin-test-data-east1/genome-scores.csv'
dataset = S3IterableDataset(s3_dataset_path)
import pandas as pd
for files in dataset:
result1 = pd.read_csv(io.BytesIO(files[1]))
s3 = boto3.client('s3')
obj = s3.get_object(Bucket=s3_dataset_path.split('/')[2],
Key=s3_dataset_path.split('/')[3])
result2 = pd.read_csv(io.BytesIO(obj['Body'].read()))
assert result1.equals(result2)
del os.environ['AWS_REGION']
def test_shuffle_true():
"""
Tests shuffle_urls parameter, len and set_epoch functions
"""
os.environ['AWS_REGION'] = 'us-west-2'
# create two datasets, one shuffled with self.epoch
s3_dataset_path = 's3://pt-s3plugin-test-data-west2/images/test'
s3_dataset0 = S3IterableDataset(s3_dataset_path)
s3_dataset1 = S3IterableDataset(s3_dataset_path, shuffle_urls=True)
s3_dataset1.set_epoch(5)
# len is defined as the length of the urls_list created by the path
assert len(s3_dataset0) == len(s3_dataset1)
# check to make sure shuffling works
filenames0 = [item[0] for item in s3_dataset0]
filenames1 = [item[0] for item in s3_dataset1]
assert len(filenames0) == len(filenames1)
assert filenames0 != filenames1
del os.environ['AWS_REGION']
def test_disable_multi_download():
s3_dataset_path = 's3://pt-s3plugin-test-data-east1/genome-scores.csv'
os.environ['S3_DISABLE_MULTI_PART_DOWNLOAD'] = "ON"
os.environ['AWS_REGION'] = 'us-east-1'
dataset = S3IterableDataset(s3_dataset_path)
import pandas as pd
for files in dataset:
result1 = pd.read_csv(io.BytesIO(files[1]))
s3 = boto3.client('s3')
obj = s3.get_object(Bucket=s3_dataset_path.split('/')[2],
Key=s3_dataset_path.split('/')[3])
result2 = pd.read_csv(io.BytesIO(obj['Body'].read()))
assert result1.equals(result2)
del os.environ['S3_DISABLE_MULTI_PART_DOWNLOAD'], os.environ['AWS_REGION']
def test_ShuffleDataset():
"""
Args:
bucket: name of the bucket
tarfiles_list: list of all tarfiles with the prefix
buffer_size: number of files the ShuffleDataset object caches
Logic:
Loop over the ShuffleDataset Dataloader twice
For the runs, the corresponding batches returned should not be the same
- ensures that shuffling is happening within tarfile constituents
After both the runs, the overall dataloaded should be the same
If either of these conditions fails, then test fails
"""
bucket = "pt-s3plugin-test-data-west2"
tarfiles_list = [
"integration_tests/imagenet-train-000000.tar",
"integration_tests/imagenet-train-000001.tar"
]
url_list = ["s3://" + bucket + "/" + tarfile for tarfile in tarfiles_list]
batch_size = 32
buffer_size = 300
for num_workers in [0, 16]:
for buffer_size in [30, 300, 3000]:
dataset = ShuffleDataset(S3IterableDataset(url_list),
buffer_size=buffer_size)
dataloader = DataLoader(dataset,
batch_size=batch_size,
num_workers=num_workers)
batch_list1 = get_batches(dataloader)
batch_list2 = get_batches(dataloader)
assert batches_shuffled(
batch_list1,
batch_list2), "ShuffleDataset Test fails: batches not shuffled"
assert batches_congruent(
batch_list1,
batch_list2), "ShuffleDataset Test fails: data mismatch"
print(
"ShuffleDataset test passes for {} buffer_size & {} workers ".
format(buffer_size, num_workers))
def test_shuffleurls():
"""
Args:
bucket : name of the bucket
files_prefix : prefix of the location where files stored
Logic:
Loop over dataloader twice, once with shuffle_urls as True and once as False
After both runs,
the dataloaded should be the same, the loading order should be different
Maintains a dictionary each of sets and lists.
The keys of the dictionary is the state of shuffle_urls(True/False)
Values are the set/list of the samples
Test passes if the set of samples loaded in both cases is same and the list of
samples is diffrent(loading order different - data being shuffled)
"""
bucket = "pt-s3plugin-test-data-west2"
files_prefix = "integration_tests/files"
assert files_prefix[
-1] != "/", "Enter Prefix without trailing \"/\" else error"
prefix_list = get_file_list(bucket, files_prefix)
url_list = ["s3://" + bucket + "/" + prefix for prefix in prefix_list]
batch_size = 32
shuffled_sets = defaultdict(set)
shuffled_lists = defaultdict(list)
print("\nINITIATING SHUFFLE TEST")
for shuffle_urls in [True, False]:
dataset = S3IterableDataset(url_list, shuffle_urls=shuffle_urls)
dataloader = DataLoader(dataset, batch_size=batch_size)
for fname, fobj in dataloader:
fname = [x.split("/")[-1] for x in fname]
batch_set = set(map(tuple, zip(fname, fobj)))
batch_list = list(map(tuple, zip(fname, fobj)))
shuffled_sets[str(shuffle_urls)].update(batch_set)
shuffled_lists[str(shuffle_urls)].append(batch_list)
assert shuffled_sets['True'] == shuffled_sets['False'] and shuffled_lists['True'] != shuffled_lists['False'], \
"Shuffling not working correctly"
print("Shuffle test passed for S3IterableDataset")
def __init__(self, url_list, shuffle_urls=False, transform=None):
self.s3_iter_dataset = S3IterableDataset(url_list,
shuffle_urls)
self.transform = transform
def __init__(self, s3_directory):
self.s3_directory = s3_directory
self.dataset = S3IterableDataset(self.s3_directory, shuffle_urls=True)
