def test_timestamped_with_combiners(self):
p = Pipeline('DirectPipelineRunner')
result = (p
# Create some initial test values.
| Create('start', [(k, k) for k in range(10)])
# The purpose of the WindowInto transform is to establish a
# FixedWindows windowing function for the PCollection.
# It does not bucket elements into windows since the timestamps
# from Create are not spaced 5 ms apart and very likely they all
# fall into the same window.
| WindowInto('w', FixedWindows(5))
# Generate timestamped values using the values as timestamps.
# Now there are values 5 ms apart and since Map propagates the
# windowing function from input to output the output PCollection
# will have elements falling into different 5ms windows.
| Map(lambda (x, t): TimestampedValue(x, t))
# We add a 'key' to each value representing the index of the
# window. This is important since there is no guarantee of
# order for the elements of a PCollection.
| Map(lambda v: (v / 5, v)))
# Sum all elements associated with a key and window. Although it
# is called CombinePerKey it is really CombinePerKeyAndWindow the
# same way GroupByKey is really GroupByKeyAndWindow.
sum_per_window = result | CombinePerKey(sum)
# Compute mean per key and window.
mean_per_window = result | combiners.Mean.PerKey()
assert_that(sum_per_window, equal_to([(0, 10), (1, 35)]),
label='assert:sum')
assert_that(mean_per_window, equal_to([(0, 2.0), (1, 7.0)]),
label='assert:mean')
p.run()
def test_timestamped_with_combiners(self):
p = Pipeline('DirectPipelineRunner')
result = (p
# Create some initial test values.
| Create('start', [(k, k) for k in range(10)])
# The purpose of the WindowInto transform is to establish a
# FixedWindows windowing function for the PCollection.
# It does not bucket elements into windows since the timestamps
# from Create are not spaced 5 ms apart and very likely they all
# fall into the same window.
| WindowInto('w', FixedWindows(5))
# Generate timestamped values using the values as timestamps.
# Now there are values 5 ms apart and since Map propagates the
# windowing function from input to output the output PCollection
# will have elements falling into different 5ms windows.
| Map(lambda (x, t): TimestampedValue(x, t))
# We add a 'key' to each value representing the index of the
# window. This is important since there is no guarantee of
# order for the elements of a PCollection.
| Map(lambda v: (v / 5, v)))
# Sum all elements associated with a key and window. Although it
# is called CombinePerKey it is really CombinePerKeyAndWindow the
# same way GroupByKey is really GroupByKeyAndWindow.
sum_per_window = result | CombinePerKey(sum)
# Compute mean per key and window.
mean_per_window = result | combiners.Mean.PerKey()
assert_that(sum_per_window, equal_to([(0, 10), (1, 35)]),
label='assert:sum')
assert_that(mean_per_window, equal_to([(0, 2.0), (1, 7.0)]),
label='assert:mean')
p.run()
def test_reuse_cloned_custom_transform_instance(self):
pipeline = Pipeline(DirectPipelineRunner())
pcoll1 = pipeline | Create('pcoll1', [1, 2, 3])
pcoll2 = pipeline | Create('pcoll2', [4, 5, 6])
transform = PipelineTest.CustomTransform()
result1 = pcoll1 | transform
result2 = pcoll2 | transform.clone('new label')
assert_that(result1, equal_to([2, 3, 4]), label='r1')
assert_that(result2, equal_to([5, 6, 7]), label='r2')
pipeline.run()
def test_create(self):
pipeline = Pipeline('DirectPipelineRunner')
pcoll = pipeline | Create('label1', [1, 2, 3])
assert_that(pcoll, equal_to([1, 2, 3]))
# Test if initial value is an iterator object.
pcoll2 = pipeline | Create('label2', iter((4, 5, 6)))
pcoll3 = pcoll2 | FlatMap('do', lambda x: [x + 10])
assert_that(pcoll3, equal_to([14, 15, 16]), label='pcoll3')
pipeline.run()
def test_create(self):
pipeline = Pipeline('DirectPipelineRunner')
pcoll = pipeline | Create('label1', [1, 2, 3])
assert_that(pcoll, equal_to([1, 2, 3]))
# Test if initial value is an iterator object.
pcoll2 = pipeline | Create('label2', iter((4, 5, 6)))
pcoll3 = pcoll2 | FlatMap('do', lambda x: [x + 10])
assert_that(pcoll3, equal_to([14, 15, 16]), label='pcoll3')
pipeline.run()
def test_reuse_cloned_custom_transform_instance(self):
pipeline = Pipeline(DirectPipelineRunner())
pcoll1 = pipeline | Create('pcoll1', [1, 2, 3])
pcoll2 = pipeline | Create('pcoll2', [4, 5, 6])
transform = PipelineTest.CustomTransform()
result1 = pcoll1 | transform
result2 = pcoll2 | transform.clone('new label')
assert_that(result1, equal_to([2, 3, 4]), label='r1')
assert_that(result2, equal_to([5, 6, 7]), label='r2')
pipeline.run()
def test_empty_side_outputs(self):
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 3, 5])
results = nums | FlatMap(
'ClassifyNumbers',
lambda x: [x, SideOutputValue('even' if x % 2 == 0 else 'odd', x)]
).with_outputs('odd', 'even', main='main')
assert_that(results.main, equal_to([1, 3, 5]))
assert_that(results.odd, equal_to([1, 3, 5]), label='assert:odd')
assert_that(results.even, equal_to([]), label='assert:even')
pipeline.run()
def test_empty_side_outputs(self):
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 3, 5])
results = nums | FlatMap(
'ClassifyNumbers', lambda x:
[x, SideOutputValue('even'
if x % 2 == 0 else 'odd', x)]).with_outputs()
assert_that(results[None], equal_to([1, 3, 5]))
assert_that(results.odd, equal_to([1, 3, 5]), label='assert:odd')
assert_that(results.even, equal_to([]), label='assert:even')
pipeline.run()
def test_combine_globally_with_default_side_input(self):
class CombineWithSideInput(PTransform):
def apply(self, pcoll):
side = pcoll | CombineGlobally(sum).as_singleton_view()
main = pcoll.pipeline | Create([None])
return main | Map(lambda _, s: s, side)
p = Pipeline('DirectPipelineRunner')
result1 = p | Create('label1', []) | CombineWithSideInput('L1')
result2 = p | Create('label2', [1, 2, 3, 4]) | CombineWithSideInput('L2')
assert_that(result1, equal_to([0]), label='r1')
assert_that(result2, equal_to([10]), label='r2')
p.run()
def test_undeclared_side_outputs(self):
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | FlatMap(
'ClassifyNumbers',
lambda x: [x, SideOutputValue('even' if x % 2 == 0 else 'odd', x)]
).with_outputs('odd', 'even', main='main')
# TODO(silviuc): Revisit this test to check for undeclared side outputs.
# This should work with .with_outputs() without any tags declared and
# the results[None] should work also.
assert_that(results.main, equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_undeclared_side_outputs(self):
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | FlatMap(
'ClassifyNumbers', lambda x:
[x, SideOutputValue('even'
if x % 2 == 0 else 'odd', x)]).with_outputs()
# TODO(silviuc): Revisit this test to check for undeclared side outputs.
# This should work with .with_outputs() without any tags declared and
# the results[None] should work also.
assert_that(results[None], equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_combine_globally_with_default_side_input(self):
class CombineWithSideInput(PTransform):
def apply(self, pcoll):
side = pcoll | CombineGlobally(sum).as_singleton_view()
main = pcoll.pipeline | Create([None])
return main | Map(lambda _, s: s, side)
p = Pipeline('DirectPipelineRunner')
result1 = p | Create('label1', []) | CombineWithSideInput('L1')
result2 = p | Create('label2',
[1, 2, 3, 4]) | CombineWithSideInput('L2')
assert_that(result1, equal_to([0]), label='r1')
assert_that(result2, equal_to([10]), label='r2')
p.run()
def test_par_do_with_multiple_outputs_and_using_return(self):
def some_fn(v):
if v % 2 == 0:
return [v, SideOutputValue('even', v)]
else:
return [v, SideOutputValue('odd', v)]
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | FlatMap('ClassifyNumbers', some_fn).with_outputs(
'odd', 'even', main='main')
assert_that(results.main, equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_par_do_with_multiple_outputs_and_using_return(self):
def some_fn(v):
if v % 2 == 0:
return [v, SideOutputValue('even', v)]
else:
return [v, SideOutputValue('odd', v)]
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | FlatMap(
'ClassifyNumbers', some_fn).with_outputs('odd', 'even', main='main')
assert_that(results.main, equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_top_shorthands(self):
pipeline = Pipeline('DirectPipelineRunner')
pcoll = pipeline | Create('start', [6, 3, 1, 1, 9, 1, 5, 2, 0, 6])
result_top = pcoll | df.CombineGlobally('top', combiners.Largest(5))
result_bot = pcoll | df.CombineGlobally('bot', combiners.Smallest(4))
assert_that(result_top, equal_to([[9, 6, 6, 5, 3]]), label='assert:top')
assert_that(result_bot, equal_to([[0, 1, 1, 1]]), label='assert:bot')
pcoll = pipeline | Create(
'start-perkey', [('a', x) for x in [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]])
result_ktop = pcoll | df.CombinePerKey('top-perkey', combiners.Largest(5))
result_kbot = pcoll | df.CombinePerKey('bot-perkey', combiners.Smallest(4))
assert_that(result_ktop, equal_to([('a', [9, 6, 6, 5, 3])]), label='k:top')
assert_that(result_kbot, equal_to([('a', [0, 1, 1, 1])]), label='k:bot')
pipeline.run()
def test_run_direct(self):
file_name = self._create_temp_file('aaaa\nbbbb\ncccc\ndddd')
pipeline = df.Pipeline('DirectPipelineRunner')
pcoll = pipeline | df.Read(LineSource(file_name))
assert_that(pcoll, equal_to(['aaaa', 'bbbb', 'cccc', 'dddd']))
pipeline.run()
def test_map(self):
pipeline = Pipeline('DirectPipelineRunner')
lines = pipeline | Create('input', ['a', 'b', 'c'])
result = (lines
| Map('upper', str.upper)
| Map('prefix', lambda x, prefix: prefix + x, 'foo-'))
assert_that(result, equal_to(['foo-A', 'foo-B', 'foo-C']))
pipeline.run()
def test_map(self):
pipeline = Pipeline('DirectPipelineRunner')
lines = pipeline | Create('input', ['a', 'b', 'c'])
result = (lines
| Map('upper', str.upper)
| Map('prefix', lambda x, prefix: prefix + x, 'foo-'))
assert_that(result, equal_to(['foo-A', 'foo-B', 'foo-C']))
pipeline.run()
def test_word_count_using_get(self):
pipeline = Pipeline('DirectPipelineRunner')
lines = pipeline | Create('SomeWords', [DataflowTest.SAMPLE_DATA])
result = (
(lines | FlatMap('GetWords', lambda x: re.findall(r'\w+', x)))
.apply('CountWords', DataflowTest.Count))
assert_that(result, equal_to(DataflowTest.SAMPLE_RESULT))
pipeline.run()
def test_iterable_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', [3, 4]) # 2 values in side input.
result = pcol | FlatMap('compute', lambda x, s: [x * y for y in s],
AllOf(side))
assert_that(result, equal_to([3, 4, 6, 8]))
pipeline.run()
def test_iterable_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', [3, 4]) # 2 values in side input.
result = pcol | FlatMap('compute',
lambda x, s: [x * y for y in s], AllOf(side))
assert_that(result, equal_to([3, 4, 6, 8]))
pipeline.run()
def test_default_value_singleton_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', []) # 0 values in side input.
result = (pcol | FlatMap('compute', lambda x, s: [x * s],
AsSingleton(side, 10)))
assert_that(result, equal_to([10, 20]))
pipeline.run()
def test_word_count_using_get(self):
pipeline = Pipeline('DirectPipelineRunner')
lines = pipeline | Create('SomeWords', [DataflowTest.SAMPLE_DATA])
result = ((lines | FlatMap('GetWords',
lambda x: re.findall(r'\w+', x))).apply(
'CountWords', DataflowTest.Count))
assert_that(result, equal_to(DataflowTest.SAMPLE_RESULT))
pipeline.run()
def test_default_value_singleton_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', []) # 0 values in side input.
result = (
pcol | FlatMap('compute', lambda x, s: [x * s], AsSingleton(side, 10)))
assert_that(result, equal_to([10, 20]))
pipeline.run()
def test_par_do_with_multiple_outputs_and_using_yield(self):
class SomeDoFn(DoFn):
"""A custom DoFn using yield."""
def process(self, context):
yield context.element
if context.element % 2 == 0:
yield SideOutputValue('even', context.element)
else:
yield SideOutputValue('odd', context.element)
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | ParDo('ClassifyNumbers', SomeDoFn()).with_outputs(
'odd', 'even', main='main')
assert_that(results.main, equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_tuple_combine_fn(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| Create([('a', 100, 0.0), ('b', 10, -1), ('c', 1, 100)])
| df.CombineGlobally(
combine.TupleCombineFn(max, combine.MeanCombineFn(),
sum)).without_defaults())
assert_that(result, equal_to([('c', 111.0 / 3, 99.0)]))
p.run()
def test_tuple_combine_fn_without_defaults(self):
p = Pipeline('DirectPipelineRunner')
result = (
p
| Create([1, 1, 2, 3])
| df.CombineGlobally(
combine.TupleCombineFn(min, combine.MeanCombineFn(), max)
.with_common_input()).without_defaults())
assert_that(result, equal_to([(1, 7.0 / 4, 3)]))
p.run()
def test_par_do_with_multiple_outputs_and_using_yield(self):
class SomeDoFn(DoFn):
"""A custom DoFn using yield."""
def process(self, context):
yield context.element
if context.element % 2 == 0:
yield SideOutputValue('even', context.element)
else:
yield SideOutputValue('odd', context.element)
pipeline = Pipeline('DirectPipelineRunner')
nums = pipeline | Create('Some Numbers', [1, 2, 3, 4])
results = nums | ParDo(
'ClassifyNumbers', SomeDoFn()).with_outputs('odd', 'even', main='main')
assert_that(results.main, equal_to([1, 2, 3, 4]))
assert_that(results.odd, equal_to([1, 3]), label='assert:odd')
assert_that(results.even, equal_to([2, 4]), label='assert:even')
pipeline.run()
def test_tuple_combine_fn(self):
p = Pipeline('DirectPipelineRunner')
result = (
p
| Create([('a', 100, 0.0), ('b', 10, -1), ('c', 1, 100)])
| df.CombineGlobally(combine.TupleCombineFn(max,
combine.MeanCombineFn(),
sum)).without_defaults())
assert_that(result, equal_to([('c', 111.0 / 3, 99.0)]))
p.run()
def test_tuple_combine_fn_without_defaults(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| Create([1, 1, 2, 3])
| df.CombineGlobally(
combine.TupleCombineFn(
min, combine.MeanCombineFn(),
max).with_common_input()).without_defaults())
assert_that(result, equal_to([(1, 7.0 / 4, 3)]))
p.run()
def test_cached_pvalues_are_refcounted(self):
"""Test that cached PValues are refcounted and deleted.
The intermediary PValues computed by the workflow below contain
one million elements so if the refcounting does not work the number of
objects tracked by the garbage collector will increase by a few millions
by the time we execute the final Map checking the objects tracked.
Anything that is much larger than what we started with will fail the test.
"""
def check_memory(value, count_threshold):
gc.collect()
objects_count = len(gc.get_objects())
if objects_count > count_threshold:
raise RuntimeError(
'PValues are not refcounted: %s, %s' % (
objects_count, count_threshold))
return value
def create_dupes(o, _):
yield o
yield SideOutputValue('side', o)
pipeline = Pipeline('DirectPipelineRunner')
gc.collect()
count_threshold = len(gc.get_objects()) + 10000
biglist = pipeline | Create('oom:create', ['x'] * 1000000)
dupes = (
biglist
| Map('oom:addone', lambda x: (x, 1))
| FlatMap('oom:dupes', create_dupes,
AsIter(biglist)).with_outputs('side', main='main'))
result = (
(dupes.side, dupes.main, dupes.side)
| Flatten('oom:flatten')
| CombinePerKey('oom:combine', sum)
| Map('oom:check', check_memory, count_threshold))
assert_that(result, equal_to([('x', 3000000)]))
pipeline.run()
self.assertEqual(
pipeline.runner.debug_counters['element_counts'],
{
'oom:flatten': 3000000,
('oom:combine/GroupByKey/reify_windows', None): 3000000,
('oom:dupes/oom:dupes', 'side'): 1000000,
('oom:dupes/oom:dupes', None): 1000000,
'oom:create': 1000000,
('oom:addone', None): 1000000,
'oom:combine/GroupByKey/group_by_key': 1,
('oom:check', None): 1,
'assert_that/singleton': 1,
('assert_that/Map(match)', None): 1,
('oom:combine/GroupByKey/group_by_window', None): 1,
('oom:combine/Combine/ParDo(CombineValuesDoFn)', None): 1})
def test_cached_pvalues_are_refcounted(self):
"""Test that cached PValues are refcounted and deleted.
The intermediary PValues computed by the workflow below contain
one million elements so if the refcounting does not work the number of
objects tracked by the garbage collector will increase by a few millions
by the time we execute the final Map checking the objects tracked.
Anything that is much larger than what we started with will fail the test.
"""
def check_memory(value, count_threshold):
gc.collect()
objects_count = len(gc.get_objects())
if objects_count > count_threshold:
raise RuntimeError(
'PValues are not refcounted: %s, %s' % (
objects_count, count_threshold))
return value
def create_dupes(o, _):
yield o
yield SideOutputValue('side', o)
pipeline = Pipeline('DirectPipelineRunner')
gc.collect()
count_threshold = len(gc.get_objects()) + 10000
biglist = pipeline | Create('oom:create', ['x'] * 1000000)
dupes = (
biglist
| Map('oom:addone', lambda x: (x, 1))
| FlatMap('oom:dupes', create_dupes,
AsIter(biglist)).with_outputs('side', main='main'))
result = (
(dupes.side, dupes.main, dupes.side)
| Flatten('oom:flatten')
| CombinePerKey('oom:combine', sum)
| Map('oom:check', check_memory, count_threshold))
assert_that(result, equal_to([('x', 3000000)]))
pipeline.run()
self.assertEqual(
pipeline.runner.debug_counters['element_counts'],
{
'oom:flatten': 3000000,
('oom:combine/GroupByKey/reify_windows', None): 3000000,
('oom:dupes/oom:dupes', 'side'): 1000000,
('oom:dupes/oom:dupes', None): 1000000,
'oom:create': 1000000,
('oom:addone', None): 1000000,
'oom:combine/GroupByKey/group_by_key': 1,
('oom:check', None): 1,
'assert_that/singleton': 1,
('assert_that/Map(match)', None): 1,
('oom:combine/GroupByKey/group_by_window', None): 1,
('oom:combine/Combine/ParDo(CombineValuesDoFn)', None): 1})
def test_par_do_with_side_input_as_arg(self):
pipeline = Pipeline('DirectPipelineRunner')
words_list = ['aa', 'bb', 'cc']
words = pipeline | Create('SomeWords', words_list)
prefix = pipeline | Create('SomeString', ['xyz']) # side in
suffix = 'zyx'
result = words | FlatMap(
'DecorateWords', lambda x, pfx, sfx: ['%s-%s-%s' % (pfx, x, sfx)],
AsSingleton(prefix), suffix)
assert_that(result, equal_to(['xyz-%s-zyx' % x for x in words_list]))
pipeline.run()
def test_timestamped_value(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| Create('start', [(k, k) for k in range(10)])
| Map(lambda (x, t): TimestampedValue(x, t))
| WindowInto('w', FixedWindows(5))
| Map(lambda v: ('key', v))
| GroupByKey())
assert_that(result, equal_to([('key', [0, 1, 2, 3, 4]),
('key', [5, 6, 7, 8, 9])]))
p.run()
def test_timestamped_value(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| Create('start', [(k, k) for k in range(10)])
| Map(lambda (x, t): TimestampedValue(x, t))
| WindowInto('w', FixedWindows(5))
| Map(lambda v: ('key', v))
| GroupByKey())
assert_that(result, equal_to([('key', [0, 1, 2, 3, 4]),
('key', [5, 6, 7, 8, 9])]))
p.run()
def test_empty_singleton_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', []) # Empty side input.
def my_fn(k, s):
v = ('empty' if isinstance(s, EmptySideInput) else 'full')
return [(k, v)]
result = pcol | FlatMap('compute', my_fn, AsSingleton(side))
assert_that(result, equal_to([(1, 'empty'), (2, 'empty')]))
pipeline.run()
def test_sliding_windows(self):
p = Pipeline('DirectPipelineRunner')
pcoll = self.timestamped_key_values(p, 'key', 1, 2, 3)
result = (pcoll
| WindowInto('w', SlidingWindows(period=2, size=4))
| GroupByKey()
| reify_windows)
expected = [('key @ [-2.0, 2.0)', [1]),
('key @ [0.0, 4.0)', [1, 2, 3]),
('key @ [2.0, 6.0)', [2, 3])]
assert_that(result, equal_to(expected))
p.run()
def test_empty_singleton_side_input(self):
pipeline = Pipeline('DirectPipelineRunner')
pcol = pipeline | Create('start', [1, 2])
side = pipeline | Create('side', []) # Empty side input.
def my_fn(k, s):
v = ('empty' if isinstance(s, EmptySideInput) else 'full')
return [(k, v)]
result = pcol | FlatMap('compute', my_fn, AsSingleton(side))
assert_that(result, equal_to([(1, 'empty'), (2, 'empty')]))
pipeline.run()
def test_par_do_with_side_input_as_keyword_arg(self):
pipeline = Pipeline('DirectPipelineRunner')
words_list = ['aa', 'bb', 'cc']
words = pipeline | Create('SomeWords', words_list)
prefix = 'zyx'
suffix = pipeline | Create('SomeString', ['xyz']) # side in
result = words | FlatMap(
'DecorateWords',
lambda x, pfx, sfx: ['%s-%s-%s' % (pfx, x, sfx)],
prefix, sfx=AsSingleton(suffix))
assert_that(result, equal_to(['zyx-%s-xyz' % x for x in words_list]))
pipeline.run()
def test_builtin_combines(self):
pipeline = Pipeline('DirectPipelineRunner')
vals = [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]
mean = sum(vals) / float(len(vals))
size = len(vals)
# First for global combines.
pcoll = pipeline | Create('start', vals)
result_mean = pcoll | combine.Mean.Globally('mean')
result_count = pcoll | combine.Count.Globally('count')
assert_that(result_mean, equal_to([mean]), label='assert:mean')
assert_that(result_count, equal_to([size]), label='assert:size')
# Again for per-key combines.
pcoll = pipeline | Create('start-perkey', [('a', x) for x in vals])
result_key_mean = pcoll | combine.Mean.PerKey('mean-perkey')
result_key_count = pcoll | combine.Count.PerKey('count-perkey')
assert_that(result_key_mean, equal_to([('a', mean)]), label='key:mean')
assert_that(result_key_count, equal_to([('a', size)]), label='key:size')
pipeline.run()
def test_sliding_windows(self):
p = Pipeline('DirectPipelineRunner')
pcoll = self.timestamped_key_values(p, 'key', 1, 2, 3)
result = (pcoll
| WindowInto('w', SlidingWindows(period=2, size=4))
| GroupByKey()
| reify_windows)
expected = [('key @ [-2, 2)', [1]),
('key @ [0, 4)', [1, 2, 3]),
('key @ [2, 6)', [2, 3])]
assert_that(result, equal_to(expected))
p.run()
def test_sessions(self):
p = Pipeline('DirectPipelineRunner')
pcoll = self.timestamped_key_values(p, 'key', 1, 2, 3, 20, 35, 27)
result = (pcoll
| WindowInto('w', Sessions(10))
| GroupByKey()
| sort_values
| reify_windows)
expected = [('key @ [1, 13)', [1, 2, 3]),
('key @ [20, 45)', [20, 27, 35])]
assert_that(result, equal_to(expected))
p.run()
def test_sessions(self):
p = Pipeline('DirectPipelineRunner')
pcoll = self.timestamped_key_values(p, 'key', 1, 2, 3, 20, 35, 27)
result = (pcoll
| WindowInto('w', Sessions(10))
| GroupByKey()
| sort_values
| reify_windows)
expected = [('key @ [1.0, 13.0)', [1, 2, 3]),
('key @ [20.0, 45.0)', [20, 27, 35])]
assert_that(result, equal_to(expected))
p.run()
def test_builtin_combines(self):
pipeline = Pipeline('DirectPipelineRunner')
vals = [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]
mean = sum(vals) / float(len(vals))
size = len(vals)
# First for global combines.
pcoll = pipeline | Create('start', vals)
result_mean = pcoll | combine.Mean.Globally('mean')
result_count = pcoll | combine.Count.Globally('count')
assert_that(result_mean, equal_to([mean]), label='assert:mean')
assert_that(result_count, equal_to([size]), label='assert:size')
# Again for per-key combines.
pcoll = pipeline | Create('start-perkey', [('a', x) for x in vals])
result_key_mean = pcoll | combine.Mean.PerKey('mean-perkey')
result_key_count = pcoll | combine.Count.PerKey('count-perkey')
assert_that(result_key_mean, equal_to([('a', mean)]), label='key:mean')
assert_that(result_key_count,
equal_to([('a', size)]),
label='key:size')
pipeline.run()
def match(actual): # There is always exactly one result. equal_to([1])([len(actual)]) # There are always exactly three samples in the result. equal_to([3])([len(actual[0])]) # Sampling is without replacement. num_ones = sum(1 for x in actual[0] if x == 1) num_twos = sum(1 for x in actual[0] if x == 2) equal_to([1, 2])([num_ones, num_twos])
def match(actual):
# There is always exactly one result.
equal_to([1])([len(actual)])
# There are always exactly three samples in the result.
equal_to([3])([len(actual[0])])
# Sampling is without replacement.
num_ones = sum(1 for x in actual[0] if x == 1)
num_twos = sum(1 for x in actual[0] if x == 2)
equal_to([1, 2])([num_ones, num_twos])
def test_top_shorthands(self):
pipeline = Pipeline('DirectPipelineRunner')
pcoll = pipeline | Create('start', [6, 3, 1, 1, 9, 1, 5, 2, 0, 6])
result_top = pcoll | df.CombineGlobally('top', combiners.Largest(5))
result_bot = pcoll | df.CombineGlobally('bot', combiners.Smallest(4))
assert_that(result_top,
equal_to([[9, 6, 6, 5, 3]]),
label='assert:top')
assert_that(result_bot, equal_to([[0, 1, 1, 1]]), label='assert:bot')
pcoll = pipeline | Create(
'start-perkey', [('a', x) for x in [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]])
result_ktop = pcoll | df.CombinePerKey('top-perkey',
combiners.Largest(5))
result_kbot = pcoll | df.CombinePerKey('bot-perkey',
combiners.Smallest(4))
assert_that(result_ktop,
equal_to([('a', [9, 6, 6, 5, 3])]),
label='k:top')
assert_that(result_kbot,
equal_to([('a', [0, 1, 1, 1])]),
label='k:bot')
pipeline.run()
def test_deferred_side_input_iterable(self):
@typehints.with_input_types(str, typehints.Iterable[str])
def concat(glue, items):
return glue.join(sorted(items))
p = df.Pipeline(options=PipelineOptions([]))
main_input = p | df.Create(['a', 'bb', 'c'])
side_input = p | df.Create('side', ['x', 'y', 'z'])
result = main_input | df.Map(concat, pvalue.AsIter(side_input))
assert_that(result, equal_to(['xayaz', 'xbbybbz', 'xcycz']))
p.run()
bad_side_input = p | df.Create('bad_side', [1, 2, 3])
with self.assertRaises(typehints.TypeCheckError):
main_input | df.Map('fail', concat, pvalue.AsIter(bad_side_input))
def test_par_do_with_do_fn_object(self):
class SomeDoFn(DoFn):
"""A custom DoFn for a FlatMap transform."""
def process(self, context, prefix, suffix):
return ['%s-%s-%s' % (prefix, context.element, suffix)]
pipeline = Pipeline('DirectPipelineRunner')
words_list = ['aa', 'bb', 'cc']
words = pipeline | Create('SomeWords', words_list)
prefix = 'zyx'
suffix = pipeline | Create('SomeString', ['xyz']) # side in
result = words | ParDo('DecorateWordsDoFn', SomeDoFn(), prefix,
suffix=AsSingleton(suffix))
assert_that(result, equal_to(['zyx-%s-xyz' % x for x in words_list]))
pipeline.run()
def test_deferred_side_inputs(self):
@typehints.with_input_types(str, int)
def repeat(s, times):
return s * times
p = df.Pipeline(options=PipelineOptions([]))
main_input = p | df.Create(['a', 'bb', 'c'])
side_input = p | df.Create('side', [3])
result = main_input | df.Map(repeat, pvalue.AsSingleton(side_input))
assert_that(result, equal_to(['aaa', 'bbbbbb', 'ccc']))
p.run()
bad_side_input = p | df.Create('bad_side', ['z'])
with self.assertRaises(typehints.TypeCheckError):
main_input | df.Map('again', repeat,
pvalue.AsSingleton(bad_side_input))
def test_after_count(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| df.Create([1, 2, 3, 4, 5, 10, 11])
| df.FlatMap(lambda t: [('A', t), ('B', t + 5)])
| df.Map(lambda (k, t): TimestampedValue((k, t), t))
| df.WindowInto(FixedWindows(10), trigger=AfterCount(3),
accumulation_mode=AccumulationMode.DISCARDING)
| df.GroupByKey()
| df.Map(lambda (k, v): ('%s-%s' % (k, len(v)), set(v))))
assert_that(result, equal_to(
{
'A-5': {1, 2, 3, 4, 5},
# A-10, A-11 never emitted due to AfterCount(3) never firing.
'B-4': {6, 7, 8, 9},
'B-3': {10, 15, 16},
}.iteritems()))
def test_par_do_with_do_fn_object(self):
class SomeDoFn(DoFn):
"""A custom DoFn for a FlatMap transform."""
def process(self, context, prefix, suffix):
return ['%s-%s-%s' % (prefix, context.element, suffix)]
pipeline = Pipeline('DirectPipelineRunner')
words_list = ['aa', 'bb', 'cc']
words = pipeline | Create('SomeWords', words_list)
prefix = 'zyx'
suffix = pipeline | Create('SomeString', ['xyz']) # side in
result = words | ParDo('DecorateWordsDoFn',
SomeDoFn(),
prefix,
suffix=AsSingleton(suffix))
assert_that(result, equal_to(['zyx-%s-xyz' % x for x in words_list]))
pipeline.run()
def test_after_count(self):
p = Pipeline('DirectPipelineRunner')
result = (p
| df.Create([1, 2, 3, 4, 5, 10, 11])
| df.FlatMap(lambda t: [('A', t), ('B', t + 5)])
| df.Map(lambda (k, t): TimestampedValue((k, t), t))
| df.WindowInto(
FixedWindows(10),
trigger=AfterCount(3),
accumulation_mode=AccumulationMode.DISCARDING)
| df.GroupByKey()
| df.Map(lambda (k, v): ('%s-%s' % (k, len(v)), set(v))))
assert_that(
result,
equal_to({
'A-5': {1, 2, 3, 4, 5},
# A-10, A-11 never emitted due to AfterCount(3) never firing.
'B-4': {6, 7, 8, 9},
'B-3': {10, 15, 16},
}.iteritems()))
def test_window_transform(self):
class TestWindowFn(WindowFn):
"""Windowing function adding two disjoint windows to each element."""
def assign(self, assign_context):
_ = assign_context
return [IntervalWindow(10, 20), IntervalWindow(20, 30)]
def merge(self, existing_windows):
return existing_windows
pipeline = Pipeline('DirectPipelineRunner')
numbers = pipeline | Create('KVs', [(1, 10), (2, 20), (3, 30)])
result = (numbers
| WindowInto('W', windowfn=TestWindowFn())
| GroupByKey('G'))
assert_that(
result, equal_to([(1, [10]), (1, [10]), (2, [20]),
(2, [20]), (3, [30]), (3, [30])]))
pipeline.run()
def test_window_transform(self):
class TestWindowFn(WindowFn):
"""Windowing function adding two disjoint windows to each element."""
def assign(self, assign_context):
_ = assign_context
return [IntervalWindow(10, 20), IntervalWindow(20, 30)]
def merge(self, existing_windows):
return existing_windows
pipeline = Pipeline('DirectPipelineRunner')
numbers = pipeline | Create('KVs', [(1, 10), (2, 20), (3, 30)])
result = (numbers
| WindowInto('W', windowfn=TestWindowFn())
| GroupByKey('G'))
assert_that(
result,
equal_to([(1, [10]), (1, [10]), (2, [20]), (2, [20]), (3, [30]),
(3, [30])]))
pipeline.run()
def test_top(self):
pipeline = Pipeline('DirectPipelineRunner')
# A parameter we'll be sharing with a custom comparator.
names = {
0: 'zo',
1: 'one',
2: 'twoo',
3: 'three',
5: 'fiiive',
6: 'sssssix',
9: 'nniiinne'
}
# First for global combines.
pcoll = pipeline | Create('start', [6, 3, 1, 1, 9, 1, 5, 2, 0, 6])
result_top = pcoll | combine.Top.Largest('top', 5)
result_bot = pcoll | combine.Top.Smallest('bot', 4)
result_cmp = pcoll | combine.Top.Of(
'cmp', 6, lambda a, b, names: len(names[a]) < len(names[b]),
names) # Note parameter passed to comparator.
assert_that(result_top,
equal_to([[9, 6, 6, 5, 3]]),
label='assert:top')
assert_that(result_bot, equal_to([[0, 1, 1, 1]]), label='assert:bot')
assert_that(result_cmp,
equal_to([[9, 6, 6, 5, 3, 2]]),
label='assert:cmp')
# Again for per-key combines.
pcoll = pipeline | Create(
'start-perkey', [('a', x) for x in [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]])
result_key_top = pcoll | combine.Top.LargestPerKey('top-perkey', 5)
result_key_bot = pcoll | combine.Top.SmallestPerKey('bot-perkey', 4)
result_key_cmp = pcoll | combine.Top.PerKey(
'cmp-perkey', 6, lambda a, b, names: len(names[a]) < len(names[b]),
names) # Note parameter passed to comparator.
assert_that(result_key_top,
equal_to([('a', [9, 6, 6, 5, 3])]),
label='key:top')
assert_that(result_key_bot,
equal_to([('a', [0, 1, 1, 1])]),
label='key:bot')
assert_that(result_key_cmp,
equal_to([('a', [9, 6, 6, 5, 3, 2])]),
label='key:cmp')
pipeline.run()
def test_top(self):
pipeline = Pipeline('DirectPipelineRunner')
# A parameter we'll be sharing with a custom comparator.
names = {0: 'zo',
1: 'one',
2: 'twoo',
3: 'three',
5: 'fiiive',
6: 'sssssix',
9: 'nniiinne'}
# First for global combines.
pcoll = pipeline | Create('start', [6, 3, 1, 1, 9, 1, 5, 2, 0, 6])
result_top = pcoll | combine.Top.Largest('top', 5)
result_bot = pcoll | combine.Top.Smallest('bot', 4)
result_cmp = pcoll | combine.Top.Of(
'cmp',
6,
lambda a, b, names: len(names[a]) < len(names[b]),
names) # Note parameter passed to comparator.
assert_that(result_top, equal_to([[9, 6, 6, 5, 3]]), label='assert:top')
assert_that(result_bot, equal_to([[0, 1, 1, 1]]), label='assert:bot')
assert_that(result_cmp, equal_to([[9, 6, 6, 5, 3, 2]]), label='assert:cmp')
# Again for per-key combines.
pcoll = pipeline | Create(
'start-perkey', [('a', x) for x in [6, 3, 1, 1, 9, 1, 5, 2, 0, 6]])
result_key_top = pcoll | combine.Top.LargestPerKey('top-perkey', 5)
result_key_bot = pcoll | combine.Top.SmallestPerKey('bot-perkey', 4)
result_key_cmp = pcoll | combine.Top.PerKey(
'cmp-perkey',
6,
lambda a, b, names: len(names[a]) < len(names[b]),
names) # Note parameter passed to comparator.
assert_that(result_key_top, equal_to([('a', [9, 6, 6, 5, 3])]),
label='key:top')
assert_that(result_key_bot, equal_to([('a', [0, 1, 1, 1])]),
label='key:bot')
assert_that(result_key_cmp, equal_to([('a', [9, 6, 6, 5, 3, 2])]),
label='key:cmp')
pipeline.run()
def test_create_singleton_pcollection(self):
pipeline = Pipeline(DirectPipelineRunner())
pcoll = pipeline | Create('label', [[1, 2, 3]])
assert_that(pcoll, equal_to([[1, 2, 3]]))
pipeline.run()
def test_apply_custom_callable(self):
pipeline = Pipeline('DirectPipelineRunner')
pcoll = pipeline | Create('pcoll', [1, 2, 3])
result = pipeline.apply(PipelineTest.custom_callable, pcoll)
assert_that(result, equal_to([2, 3, 4]))
pipeline.run()