def test_read_7z():
# file was created with:
# 7z a tests/data.7z tests/readme_example.py
# (brew install p7zip)
rdd = Context().textFile('tests/data.7z')
print(rdd.collect())
assert 'from pysparkling import Context' in rdd.collect()
def test_read_7z():
# file was created with:
# 7z a tests/data.7z tests/readme_example.py
# (brew install p7zip)
rdd = Context().textFile('tests/data.7z')
print(rdd.collect())
assert 'from pysparkling import Context' in rdd.collect()
def test_saveAsTextFile_zip():
tempFile = tempfile.NamedTemporaryFile(delete=True)
tempFile.close()
Context().parallelize(range(10)).saveAsTextFile(tempFile.name+'.zip')
read_rdd = Context().textFile(tempFile.name+'.zip')
print(read_rdd.collect())
assert '5' in read_rdd.collect()
def test_saveAsTextFile_zip():
tempFile = tempfile.NamedTemporaryFile(delete=True)
tempFile.close()
Context().parallelize(range(10)).saveAsTextFile(tempFile.name + '.zip')
read_rdd = Context().textFile(tempFile.name + '.zip')
print(read_rdd.collect())
assert '5' in read_rdd.collect()
def test_lazy_execution():
r = Context().textFile('tests/test_multiprocessing.py')
r = r.map(indent_line)
exec_before_collect = INDENT_WAS_EXECUTED
# at this point, no map() or foreach() should have been executed
r.collect()
exec_after_collect = INDENT_WAS_EXECUTED
assert not exec_before_collect and exec_after_collect
def test_cache():
# this crashes in version 0.2.28
lines = Context().textFile('tests/*textFil*.py')
lines = lines.map(lambda l: '-' + l).cache()
print(len(lines.collect()))
lines = lines.map(lambda l: '+' + l)
lines = lines.map(lambda l: '-' + l).cache()
lines = lines.collect()
print(lines)
assert '-+-from pysparkling import Context' in lines
def test_cache():
# this crashes in version 0.2.28
lines = Context().textFile('tests/*textFil*.py')
lines = lines.map(lambda l: '-'+l).cache()
print(len(lines.collect()))
lines = lines.map(lambda l: '+'+l)
lines = lines.map(lambda l: '-'+l).cache()
lines = lines.collect()
print(lines)
assert '-+-from pysparkling import Context' in lines
def test_cache():
my_rdd = Context().parallelize([1, 2, 3, 4], 2)
my_rdd = my_rdd.map(lambda x: x*x).cache()
print('no exec until here')
print(my_rdd.first())
print('executed map on first partition only')
print(my_rdd.collect())
print('now map() was executed on all partitions and should '
'not be executed again')
print(my_rdd.collect())
assert len(my_rdd.collect()) == 4 and 16 in my_rdd.collect()
def test_s3_textFile_loop():
random.seed()
fn = f'{S3_TEST_PATH}/pysparkling_test_{random.random() * 999999.0:d}.txt'
rdd = Context().parallelize(f'Line {n}' for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (rdd.count() == rdd_check.count()
and all(e1 == e2
for e1, e2 in zip(rdd.collect(), rdd_check.collect())))
def test_lazy_execution_threadpool():
def indent_line(l):
return '--- ' + l
with futures.ThreadPoolExecutor(4) as p:
r = Context(pool=p).textFile('tests/test_multiprocessing.py')
r = r.map(indent_line).cache()
r.collect()
r = r.map(indent_line)
r = r.collect()
# ThreadPool is not lazy although it returns generators.
print(r)
assert '--- --- from pysparkling import Context' in r
def test_lazy_execution_threadpool():
def indent_line(l):
return '--- '+l
with futures.ThreadPoolExecutor(4) as p:
r = Context(pool=p).textFile('tests/test_multiprocessing.py')
r = r.map(indent_line).cache()
r.collect()
r = r.map(indent_line)
r = r.collect()
# ThreadPool is not lazy although it returns generators.
print(r)
assert '--- --- from pysparkling import Context' in r
def test_gs_textFile_loop():
random.seed()
fn = '{}/pysparkling_test_{:d}.txt'.format(GS_TEST_PATH,
random.random() * 999999.0)
rdd = Context().parallelize('Line {0}'.format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (rdd.count() == rdd_check.count()
and all(e1 == e2
for e1, e2 in zip(rdd.collect(), rdd_check.collect())))
def test_gs_textFile_loop():
random.seed()
fn = '{}/pysparkling_test_{:d}.txt'.format(
GS_TEST_PATH, random.random() * 999999.0)
rdd = Context().parallelize('Line {0}'.format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (
rdd.count() == rdd_check.count() and
all(e1 == e2 for e1, e2 in zip(rdd.collect(), rdd_check.collect()))
)
def test_hdfs_textFile_loop():
random.seed()
fn = f'{HDFS_TEST_PATH}/pysparkling_test_{random.random() * 999999.0:d}.txt'
print(f'HDFS test file: {fn}')
rdd = Context().parallelize(f'Hello World {x}' for x in range(10))
rdd.saveAsTextFile(fn)
read_rdd = Context().textFile(fn)
print(rdd.collect())
print(read_rdd.collect())
assert (rdd.count() == read_rdd.count()
and all(r1 == r2
for r1, r2 in zip(rdd.collect(), read_rdd.collect())))
def test_lazy_execution_processpool():
with futures.ProcessPoolExecutor(4) as p:
r = Context(
pool=p,
serializer=dill.dumps,
deserializer=dill.loads,
).textFile('tests/test_multiprocessing.py')
r = r.map(indent_line).cache()
r.collect()
r = r.map(indent_line)
r = r.collect()
# ProcessPool is not lazy although it returns generators.
print(r)
assert '--- --- from pysparkling import Context' in r
def test_processpool_distributed_cache():
with futures.ProcessPoolExecutor(4) as p:
r = Context(
pool=p,
serializer=cloudpickle.dumps,
deserializer=pickle.loads,
).parallelize(range(3), 3)
r = r.map(lambda _: time.sleep(0.1)).cache()
r.collect()
time_start = time.time()
print(r.collect())
time_end = time.time()
assert time_end - time_start < 0.3
def test_processpool_distributed_cache():
with futures.ProcessPoolExecutor(4) as p:
r = Context(
pool=p,
serializer=cloudpickle.dumps,
deserializer=pickle.loads,
).parallelize(range(3), 3)
r = r.map(lambda _: time.sleep(0.1)).cache()
r.collect()
time_start = time.time()
print(r.collect())
time_end = time.time()
assert time_end - time_start < 0.3
def test_hdfs_textFile_loop():
random.seed()
fn = '{}/pysparkling_test_{:d}.txt'.format(HDFS_TEST_PATH,
random.random() * 999999.0)
print('HDFS test file: {0}'.format(fn))
rdd = Context().parallelize('Hello World {0}'.format(x) for x in range(10))
rdd.saveAsTextFile(fn)
read_rdd = Context().textFile(fn)
print(rdd.collect())
print(read_rdd.collect())
assert (rdd.count() == read_rdd.count()
and all(r1 == r2
for r1, r2 in zip(rdd.collect(), read_rdd.collect())))
def test_gs_textFile_loop():
if not OAUTH2_CLIENT_ID or not GS_TEST_PATH:
raise SkipTest
random.seed()
fn = GS_TEST_PATH + '/pysparkling_test_{0}.txt'.format(
int(random.random() * 999999.0))
rdd = Context().parallelize("Line {0}".format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (rdd.count() == rdd_check.count()
and all(e1 == e2
for e1, e2 in zip(rdd.collect(), rdd_check.collect())))
def test_s3_textFile_loop():
if not AWS_ACCESS_KEY_ID or not S3_TEST_PATH:
raise SkipTest
random.seed()
fn = '{}/pysparkling_test_{0}.txt'.format(S3_TEST_PATH,
int(random.random() * 999999.0))
rdd = Context().parallelize("Line {0}".format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (rdd.count() == rdd_check.count()
and all(e1 == e2
for e1, e2 in zip(rdd.collect(), rdd_check.collect())))
def test_hdfs_textFile_loop():
random.seed()
fn = '{}/pysparkling_test_{:d}.txt'.format(
HDFS_TEST_PATH, random.random() * 999999.0)
print('HDFS test file: {0}'.format(fn))
rdd = Context().parallelize('Hello World {0}'.format(x) for x in range(10))
rdd.saveAsTextFile(fn)
read_rdd = Context().textFile(fn)
print(rdd.collect())
print(read_rdd.collect())
assert (
rdd.count() == read_rdd.count() and
all(r1 == r2 for r1, r2 in zip(rdd.collect(), read_rdd.collect()))
)
def test_filter():
my_rdd = Context().parallelize(
[1, 2, 2, 4, 1, 3, 5, 9],
3,
).filter(lambda x: x % 2 == 0)
print(my_rdd.collect())
print(my_rdd.count())
assert my_rdd.count() == 3
def test_hdfs_textFile_loop():
if not HDFS_TEST_PATH:
raise SkipTest
random.seed()
fn = HDFS_TEST_PATH+'/pysparkling_test_{0}.txt'.format(
int(random.random()*999999.0)
)
rdd = Context().parallelize('Hello World {0}'.format(x) for x in range(10))
rdd.saveAsTextFile(fn)
read_rdd = Context().textFile(fn)
assert (
rdd.count() == read_rdd.count() and
all(r1 == r2 for r1, r2 in zip(rdd.collect(), read_rdd.collect()))
)
def test_s3_textFile():
myrdd = Context().textFile(
's3n://aws-publicdatasets/common-crawl/crawl-data/'
'CC-MAIN-2015-11/warc.paths.*')
assert (
'common-crawl/crawl-data/CC-MAIN-2015-11/segments/1424937481488.49/'
'warc/CC-MAIN-20150226075801-00329-ip-10-28-5-156.ec2.'
'internal.warc.gz' in myrdd.collect())
def test_gs_textFile_loop():
if not OAUTH2_CLIENT_ID or not GS_TEST_PATH:
raise SkipTest
random.seed()
fn = '{}/pysparkling_test_{0}.txt'.format(
GS_TEST_PATH, int(random.random() * 999999.0)
)
rdd = Context().parallelize("Line {0}".format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (
rdd.count() == rdd_check.count() and
all(e1 == e2 for e1, e2 in zip(rdd.collect(), rdd_check.collect()))
)
def test_s3_textFile_loop():
if not AWS_ACCESS_KEY_ID or not S3_TEST_PATH:
raise SkipTest
random.seed()
fn = S3_TEST_PATH+'/pysparkling_test_{0}.txt'.format(
int(random.random()*999999.0)
)
rdd = Context().parallelize("Line {0}".format(n) for n in range(200))
rdd.saveAsTextFile(fn)
rdd_check = Context().textFile(fn)
assert (
rdd.count() == rdd_check.count() and
all(e1 == e2 for e1, e2 in zip(rdd.collect(), rdd_check.collect()))
)
def test_s3_textFile():
myrdd = Context().textFile(
's3n://aws-publicdatasets/common-crawl/crawl-data/'
'CC-MAIN-2015-11/warc.paths.*'
)
assert (
'common-crawl/crawl-data/CC-MAIN-2015-11/segments/1424937481488.49/'
'warc/CC-MAIN-20150226075801-00329-ip-10-28-5-156.ec2.'
'internal.warc.gz' in myrdd.collect()
)
def test_lazy_execution_processpool():
def indent_line(l):
return '--- '+l
with futures.ProcessPoolExecutor(4) as p:
r = Context(
pool=p,
serializer=cloudpickle.dumps,
deserializer=pickle.loads,
).textFile('tests/test_multiprocessing.py') # .take(10)
print(r.collect())
r = r.map(indent_line)
print(r.collect())
r = r.cache()
print(r.collect())
r = r.map(indent_line)
r = r.collect()
# ProcessPool is not lazy although it returns generators.
print(r)
assert '--- --- from pysparkling import Context' in r
def test_lazy_execution_processpool():
def indent_line(l):
return '--- ' + l
with futures.ProcessPoolExecutor(4) as p:
r = Context(
pool=p,
serializer=cloudpickle.dumps,
deserializer=pickle.loads,
).textFile('tests/test_multiprocessing.py') # .take(10)
print(r.collect())
r = r.map(indent_line)
print(r.collect())
r = r.cache()
print(r.collect())
r = r.map(indent_line)
r = r.collect()
# ProcessPool is not lazy although it returns generators.
print(r)
assert '--- --- from pysparkling import Context' in r
def test_lazy_execution():
class I(object):
def __init__(self):
self.executed = False
def indent_line(self, l):
# global indent_was_executed
self.executed = True
return '--- ' + l
r = Context().textFile('tests/test_multiprocessing.py')
i = I()
r = r.map(i.indent_line)
exec_before_collect = i.executed
# at this point, no map() or foreach() should have been executed
r = r.map(i.indent_line).cache()
print(r.collect())
r = r.map(i.indent_line)
r.collect()
exec_after_collect = i.executed
print((exec_before_collect, exec_after_collect))
assert not exec_before_collect and exec_after_collect
def test_lazy_execution():
class I(object):
def __init__(self):
self.executed = False
def indent_line(self, l):
# global indent_was_executed
self.executed = True
return '--- '+l
r = Context().textFile('tests/test_multiprocessing.py')
i = I()
r = r.map(i.indent_line)
exec_before_collect = i.executed
# at this point, no map() or foreach() should have been executed
r = r.map(i.indent_line).cache()
print(r.collect())
r = r.map(i.indent_line)
r.collect()
exec_after_collect = i.executed
print((exec_before_collect, exec_after_collect))
assert not exec_before_collect and exec_after_collect
def test_http_textFile():
myrdd = Context().textFile(
'https://s3-us-west-2.amazonaws.com/human-microbiome-project/DEMO/'
'HM16STR/46333/by_subject/1139.fsa'
)
assert u'TGCTGCGGTGAATGCGTTCCCGGGTCT' in myrdd.collect()
def test_local_textFile_1():
lines = Context().textFile('{}/*textFil*.py'.format(LOCAL_TEST_PATH))
lines = lines.collect()
print(lines)
assert 'from pysparkling import Context' in lines
def test_http_textFile():
myrdd = Context().textFile(
'https://s3-us-west-2.amazonaws.com/human-microbiome-project/DEMO/'
'HM16STR/46333/by_subject/1139.fsa')
assert u'TGCTGCGGTGAATGCGTTCCCGGGTCT' in myrdd.collect()
from pysparkling import Context
counts = Context().textFile('README.rst').flatMap(lambda line: line.split(
' ')).map(lambda word: (word, 1)).reduceByKey(lambda a, b: a + b)
print(counts.collect())
def test_parallelize_matched_elements():
my_rdd = Context().parallelize([1, 2, 3, 4, 5], 5)
assert my_rdd.collect()[2] == 3 and len(my_rdd.collect()) == 5
from pysparkling import Context
counts = Context().textFile(
'README.rst'
).flatMap(
lambda line: line.split(' ')
).map(
lambda word: (word, 1)
).reduceByKey(
lambda a, b: a + b
)
print(counts.collect())
def test_local_textFile_1():
lines = Context().textFile('{}/*textFil*.py'.format(LOCAL_TEST_PATH))
lines = lines.collect()
print(lines)
assert 'from pysparkling import Context' in lines
def test_read_tar_gz():
# file was created with:
# tar -cvzf data.tar.gz hello.txt
rdd = Context().textFile('{}/data.tar.gz'.format(LOCAL_TEST_PATH))
print(rdd.collect())
assert 'Hello pysparkling!' in rdd.collect()
def test_map():
my_rdd = Context().parallelize([1, 2, 3]).map(lambda x: x+1)
assert my_rdd.collect()[0] == 2
def test_read_tar_gz():
# file was created with:
# tar -cvzf data.tar.gz hello.txt
rdd = Context().textFile('tests/data.tar.gz')
print(rdd.collect())
assert 'Hello pysparkling!' in rdd.collect()
def test_parallelize_single_element():
my_rdd = Context().parallelize([7], 100)
assert my_rdd.collect()[0] == 7
def test_collect():
my_rdd = Context().parallelize([1, 2, 3])
assert my_rdd.collect()[0] == 1
from pysparkling import Context
# read all the paths of warc and wat files of the latest Common Crawl
paths_rdd = Context().textFile(
's3n://aws-publicdatasets/common-crawl/crawl-data/CC-MAIN-2015-11/warc.paths.*,'
's3n://aws-publicdatasets/common-crawl/crawl-data/CC-MAIN-2015-11/wat.paths.gz'
)
print(paths_rdd.collect())
def test_read_tar_gz():
# file was created with:
# tar -cvzf data.tar.gz hello.txt
rdd = Context().textFile('{}/data.tar.gz'.format(LOCAL_TEST_PATH))
print(rdd.collect())
assert 'Hello pysparkling!' in rdd.collect()
def test_keys():
rdd = Context().parallelize([(0, 1), (1, 1)]).keys()
assert rdd.collect()[0] == 0
def test_parallelize_matched_elements():
my_rdd = Context().parallelize([1, 2, 3, 4, 5], 5)
assert my_rdd.collect()[2] == 3 and len(my_rdd.collect()) == 5
def test_read_tar_gz():
# file was created with:
# tar -cvzf data.tar.gz hello.txt
rdd = Context().textFile('tests/data.tar.gz')
print(rdd.collect())
assert 'Hello pysparkling!' in rdd.collect()
def test_keyBy():
rdd = Context().parallelize([0, 4, 7, 4, 10])
rdd = rdd.keyBy(lambda x: x % 2)
assert rdd.collect()[2][0] == 1 # the third element (7) is odd
def test_parallelize_single_element():
my_rdd = Context().parallelize([7], 100)
assert my_rdd.collect()[0] == 7
def test_count_partitions():
my_rdd = Context().parallelize([1, 2, 3], 2)
print(my_rdd.collect())
my_rdd.foreach(print)
assert my_rdd.count() == 3