[Dask] Dask Example Tests (#16346)

* add examples

* update dask docs

* add build file

* formatting

* fix ci command

* fix

* Update python/ray/util/dask/BUILD

* newline

* fix pytest fixtures

* fixes

* formatting

* fix shuffle example

.buildkite/pipeline.yml

@@ -150,6 +150,8 @@
       --test_tag_filters=-kubernetes,-jenkins_only,client_tests,-flaky
       --test_env=RAY_CLIENT_MODE=1 --test_env=RAY_PROFILING=1
       python/ray/tests/...
+    # Dask tests and examples.
+    - bazel test --config=ci $(./scripts/bazel_export_options) --build_tests_only --test_tag_filters=-client python/ray/util/dask/...
 - label: ":python: (Medium A-J)"
   conditions: ["RAY_CI_PYTHON_AFFECTED"]
   commands:
@@ -296,6 +298,7 @@
     - cleanup() { if [ "${BUILDKITE_PULL_REQUEST}" = "false" ]; then ./ci/travis/upload_build_info.sh; fi }; trap cleanup EXIT
     - TUNE_TESTING=1 PYTHON=3.7 INSTALL_HOROVOD=1 ./ci/travis/install-dependencies.sh
     - rm -rf ./python/ray/thirdparty_files; ./ci/travis/ci.sh build
+    - bazel test --config=ci $(./scripts/bazel_export_options) --build_tests_only --test_tag_filters=client --test_env=RAY_CLIENT_MODE=1 python/ray/util/dask/...
     - bazel test --config=ci $(./scripts/bazel_export_options) --build_tests_only --test_tag_filters=client,-flaky python/ray/tune/...
     - bazel test --config=ci $(./scripts/bazel_export_options) --build_tests_only --test_tag_filters=client,-client_unit_tests python/ray/util/sgd/...
     - bazel test --config=ci $(./scripts/bazel_export_options) --build_tests_only --test_tag_filters=client,-flaky python/ray/util/xgboost/...

doc/source/dask-on-ray.rst

@@ -31,33 +31,8 @@ any Dask `.compute() <https://docs.dask.org/en/latest/api.html#dask.compute>`__
 call.
 Here's an example:
 
-.. code-block:: python
+.. literalinclude:: ../../python/ray/util/dask/examples/dask_ray_scheduler_example.py
-
+    :language: python
-   import ray
-   from ray.util.dask import ray_dask_get
-   import dask
-   import dask.array as da
-   import dask.dataframe as dd
-   import numpy as np
-   import pandas as pd
-
-   # Start Ray.
-   # Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
-   ray.init()
-
-   d_arr = da.from_array(np.random.randint(0, 1000, size=(256, 256)))
-
-   # The Dask scheduler submits the underlying task graph to Ray.
-   d_arr.mean().compute(scheduler=ray_dask_get)
-
-   # Set the scheduler to ray_dask_get in your config so you don't have to specify it on
-   # each compute call.
-   dask.config.set(scheduler=ray_dask_get)
-
-   df = dd.from_pandas(pd.DataFrame(
-       np.random.randint(0, 100, size=(1024, 2)),
-       columns=["age", "grade"]), npartitions=2)
-   df.groupby(["age"]).mean().compute()
 
 .. note::
   For execution on a Ray cluster, you should *not* use the
@@ -121,42 +96,8 @@ aggregations): those downstream computations will be faster since that base coll
 computation was kicked off early and referenced by all downstream computations, often
 via shared memory.
 
-.. code-block:: python
+.. literalinclude:: ../../python/ray/util/dask/examples/dask_ray_persist_example.py
-
+    :language: python
-   import ray
-   from ray.util.dask import ray_dask_get
-   import dask
-   import dask.array as da
-   import numpy as np
-   # Start Ray.
-   # Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
-   ray.init()
-
-   # Set the scheduler to ray_dask_get in your config so you don't have to specify it on
-   # each compute call.
-   dask.config.set(scheduler=ray_dask_get)
-
-   d_arr = da.ones(100)
-   print(dask.base.collections_to_dsk([d_arr]))
-   # {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
-   #   0): (functools.partial(<function _broadcast_trick_inner at 0x7f27f1a71f80>,
-   #   dtype=dtype('float64')), (5,))}
-
-   # This submits all underlying Ray tasks to the cluster and returns a Dask array with
-   # the Ray futures inlined.
-   d_arr_p = d_arr.persist()
-
-   # Notice that the Ray ObjectRef is inlined. The dask.ones() task has been submitted
-   # to and is running on the Ray cluster.
-   dask.base.collections_to_dsk([d_arr_p])
-   # {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
-   #   0): ObjectRef(8b4e50dc1ddac855ffffffffffffffffffffffff0100000001000000)}
-
-   # Future computations on this persisted Dask Array will be fast since we already
-   # started computing d_arr_p in the background.
-   d_arr_p.sum().compute()
-   d_arr_p.min().compute()
-   d_arr_p.max().compute()
 
 
 Custom optimization for Dask DataFrame shuffling
@@ -168,31 +109,8 @@ Dask on Ray provides a Dask DataFrame optimizer that leverages Ray's ability to
 execute multiple-return tasks in order to speed up shuffling by as much as 4x on Ray.
 Simply set the `dataframe_optimize` configuration option to our optimizer function, similar to how you specify the Dask-on-Ray scheduler:
 
-.. code-block:: python
+.. literalinclude:: ../../python/ray/util/dask/examples/dask_ray_shuffle_optimization.py
-
+    :language: python
-   import ray
-   from ray.util.dask import ray_dask_get, dataframe_optimize
-   import dask
-   import dask.dataframe as dd
-   import numpy as np
-   import pandas as pd
-
-   # Start Ray.
-   # Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
-   ray.init()
-
-   # Set the scheduler to ray_dask_get, and set the Dask DataFrame optimizer to our
-   # custom optimization function, this time using the config setter as a context manager.
-   with dask.config.set(scheduler=ray_dask_get, dataframe_optimize=dataframe_optimize):
-       npartitions = 100
-       df = dd.from_pandas(pd.DataFrame(
-           np.random.randint(0, 100, size=(10000, 2)),
-           columns=["age", "grade"]), npartitions=npartitions)
-       # We set max_branch to infinity in order to ensure that the task-based shuffle
-       # happens in a single stage, which is required in order for our optimization to
-       # work.
-       df.set_index(
-           ["age"], shuffle="tasks", max_branch=float("inf")).head(10, npartitions=-1)
 
 Callbacks
 ---------

python/ray/tests/BUILD

@@ -125,8 +125,6 @@ py_test_module_list(
     "test_command_runner.py",
     "test_component_failures.py",
     "test_coordinator_server.py",
-    "test_dask_callback.py",
-    "test_dask_scheduler.py",
     "test_debug_tools.py",
     "test_job.py",
     "test_k8s_operator_unit_tests.py",
@@ -239,9 +237,6 @@ py_test_module_list(
     "test_basic_2.py",
     "test_basic_3.py",
     "test_asyncio.py",
-    "test_dask_callback.py",
-    "test_dask_scheduler.py",
-    "test_dask_optimization.py",
   ],
   size = "medium",
   extra_srcs = SRCS,

python/ray/util/dask/BUILD

@@ -0,0 +1,118 @@
+# --------------------------------------------------------------------
+# Tests from the python/ray/util/dask/tests directory.
+# Please keep these sorted alphabetically.
+# --------------------------------------------------------------------
+
+py_test(
+    name = "test_dask_callback",
+    size = "small",
+    srcs = ["tests/test_dask_callback.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"]
+)
+
+py_test(
+    name = "test_dask_callback_client_mode",
+    size = "small",
+    main = "test_dask_callback.py",
+    srcs = ["tests/test_dask_callback.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"]
+)
+
+py_test(
+    name = "test_dask_optimization",
+    size = "small",
+    srcs = ["tests/test_dask_optimization.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"]
+)
+
+py_test(
+    name = "test_dask_optimization_client_mode",
+    size = "small",
+    main = "test_dask_optimization.py",
+    srcs = ["tests/test_dask_optimization.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"]
+)
+
+py_test(
+    name = "test_dask_scheduler",
+    size = "small",
+    srcs = ["tests/test_dask_scheduler.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"]
+)
+
+py_test(
+    name = "test_dask_scheduler_client_mode",
+    size = "small",
+    main = "test_dask_scheduler.py",
+    srcs = ["tests/test_dask_scheduler.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"]
+)
+
+# --------------------------------------------------------------------
+# Tests from the python/ray/util/dask/examples directory.
+# Please keep these sorted alphabetically.
+# --------------------------------------------------------------------
+
+py_test(
+    name = "dask_ray_persist_example",
+    size = "medium",
+    srcs = ["examples/dask_ray_persist_example.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"],
+)
+
+py_test(
+    name = "dask_ray_persist_example_client_mode",
+    size = "medium",
+    main = "dask_ray_persist_example.py",
+    srcs = ["examples/dask_ray_persist_example.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"],
+)
+
+py_test(
+    name = "dask_ray_scheduler_example",
+    size = "medium",
+    srcs = ["examples/dask_ray_scheduler_example.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"],
+)
+
+py_test(
+    name = "dask_ray_scheduler_example_client_mode",
+    size = "medium",
+    main = "dask_ray_scheduler_example.py",
+    srcs = ["examples/dask_ray_scheduler_example.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"],
+)
+
+py_test(
+    name = "dask_ray_shuffle_optimization",
+    size = "medium",
+    srcs = ["examples/dask_ray_shuffle_optimization.py"],
+    tags = ["exclusive"],
+    deps = [":dask_lib"],
+)
+
+py_test(
+    name = "dask_ray_shuffle_optimization_client_mode",
+    size = "medium",
+    main = "dask_ray_shuffle_optimization.py",
+    srcs = ["examples/dask_ray_shuffle_optimization.py"],
+    tags = ["exclusive", "client"],
+    deps = [":dask_lib"],
+)
+
+# This is a dummy test dependency that causes the above tests to be
+# re-run if any of these files changes.
+py_library(
+    name = "dask_lib",
+    srcs = glob(["**/*.py"], exclude=["tests/*.py"]),
+)

python/ray/util/dask/examples/dask_ray_persist_example.py

@@ -0,0 +1,36 @@
+import ray
+from ray.util.dask import ray_dask_get
+import dask
+import dask.array as da
+
+# Start Ray.
+# Tip: If connecting to an existing cluster, use ray.init(address="auto").
+ray.init()
+
+# Set the scheduler to ray_dask_get in your config so you don't
+# have to specify it on each compute call.
+dask.config.set(scheduler=ray_dask_get)
+
+d_arr = da.ones(100)
+print(dask.base.collections_to_dsk([d_arr]))
+# {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
+#   0): (functools.partial(<function _broadcast_trick_inner at 0x7f27f1a71f80>,
+#   dtype=dtype('float64')), (5,))}
+
+# This submits all underlying Ray tasks to the cluster and returns
+# a Dask array with the Ray futures inlined.
+d_arr_p = d_arr.persist()
+
+# Notice that the Ray ObjectRef is inlined. The dask.ones() task has
+# been submitted to and is running on the Ray cluster.
+dask.base.collections_to_dsk([d_arr_p])
+# {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
+#   0): ObjectRef(8b4e50dc1ddac855ffffffffffffffffffffffff0100000001000000)}
+
+# Future computations on this persisted Dask Array will be fast since we
+# already started computing d_arr_p in the background.
+d_arr_p.sum().compute()
+d_arr_p.min().compute()
+d_arr_p.max().compute()
+
+ray.shutdown()

python/ray/util/dask/examples/dask_ray_scheduler_example.py

@@ -0,0 +1,28 @@
+import ray
+from ray.util.dask import ray_dask_get
+import dask
+import dask.array as da
+import dask.dataframe as dd
+import numpy as np
+import pandas as pd
+
+# Start Ray.
+# Tip: If connecting to an existing cluster, use ray.init(address="auto").
+ray.init()
+
+d_arr = da.from_array(np.random.randint(0, 1000, size=(256, 256)))
+
+# The Dask scheduler submits the underlying task graph to Ray.
+d_arr.mean().compute(scheduler=ray_dask_get)
+
+# Set the scheduler to ray_dask_get in your config so you don't have to
+# specify it on each compute call.
+dask.config.set(scheduler=ray_dask_get)
+
+df = dd.from_pandas(
+    pd.DataFrame(
+        np.random.randint(0, 100, size=(1024, 2)), columns=["age", "grade"]),
+    npartitions=2)
+df.groupby(["age"]).mean().compute()
+
+ray.shutdown()

python/ray/util/dask/examples/dask_ray_shuffle_optimization.py

@@ -0,0 +1,29 @@
+import ray
+from ray.util.dask import dataframe_optimize
+import dask
+import dask.dataframe as dd
+import numpy as np
+import pandas as pd
+
+# Start Ray.
+# Tip: If connecting to an existing cluster, use ray.init(address="auto").
+ray.init()
+
+# Set the Dask DataFrame optimizer to
+# our custom optimization function, this time using the config setter as a
+# context manager.
+with dask.config.set(dataframe_optimize=dataframe_optimize):
+    npartitions = 100
+    df = dd.from_pandas(
+        pd.DataFrame(
+            np.random.randint(0, 100, size=(10000, 2)),
+            columns=["age", "grade"]),
+        npartitions=npartitions)
+    # We set max_branch to infinity in order to ensure that the task-based
+    # shuffle happens in a single stage, which is required in order for our
+    # optimization to work.
+    df.set_index(
+        ["age"], shuffle="tasks", max_branch=float("inf")).head(
+            10, npartitions=-1)
+
+ray.shutdown()

python/ray/util/dask/tests/test_dask_callback.py

@@ -5,6 +5,14 @@ import ray
 from ray.util.dask import ray_dask_get, RayDaskCallback
 
 
+@pytest.fixture
+def ray_start_1_cpu():
+    address_info = ray.init(num_cpus=2)
+    yield address_info
+    # The code after the yield will run as teardown code.
+    ray.shutdown()
+
+
 @dask.delayed
 def add(x, y):
     return x + y
@@ -20,7 +28,7 @@ def test_callback_active():
     assert not RayDaskCallback.ray_active
 
 
-def test_presubmit_shortcircuit(ray_start_regular_shared):
+def test_presubmit_shortcircuit(ray_start_1_cpu):
     """
     Test that presubmit return short-circuits task submission, and that task's
     result is set to the presubmit return value.
@@ -41,7 +49,7 @@ def test_presubmit_shortcircuit(ray_start_regular_shared):
     assert result == 0
 
 
-def test_pretask_posttask_shared_state(ray_start_regular_shared):
+def test_pretask_posttask_shared_state(ray_start_1_cpu):
     """
     Test that pretask return value is passed to corresponding posttask
     callback.
@@ -61,7 +69,7 @@ def test_pretask_posttask_shared_state(ray_start_regular_shared):
     assert result == 5
 
 
-def test_postsubmit(ray_start_regular_shared):
+def test_postsubmit(ray_start_1_cpu):
     """
     Test that postsubmit is called after each task.
     """
@@ -94,7 +102,7 @@ def test_postsubmit(ray_start_regular_shared):
     assert result == 5
 
 
-def test_postsubmit_all(ray_start_regular_shared):
+def test_postsubmit_all(ray_start_1_cpu):
     """
     Test that postsubmit_all is called once.
     """
@@ -126,7 +134,7 @@ def test_postsubmit_all(ray_start_regular_shared):
     assert result == 5
 
 
-def test_finish(ray_start_regular_shared):
+def test_finish(ray_start_1_cpu):
     """
     Test that finish callback is called once.
     """
@@ -158,7 +166,7 @@ def test_finish(ray_start_regular_shared):
     assert result == 5
 
 
-def test_multiple_callbacks(ray_start_regular_shared):
+def test_multiple_callbacks(ray_start_1_cpu):
     """
     Test that multiple callbacks are supported.
     """
@@ -194,7 +202,7 @@ def test_multiple_callbacks(ray_start_regular_shared):
     assert result == 5
 
 
-def test_pretask_posttask_shared_state_multi(ray_start_regular_shared):
+def test_pretask_posttask_shared_state_multi(ray_start_1_cpu):
     """
     Test that pretask return values are passed to the correct corresponding
     posttask callbacks when multiple callbacks are given.

python/ray/util/dask/tests/test_dask_scheduler.py

@@ -7,8 +7,16 @@ import ray
 from ray.util.client.common import ClientObjectRef
 
 
+@pytest.fixture
+def ray_start_1_cpu():
+    address_info = ray.init(num_cpus=2)
+    yield address_info
+    # The code after the yield will run as teardown code.
+    ray.shutdown()
+
+
 @unittest.skipIf(sys.platform == "win32", "Failing on Windows.")
-def test_ray_dask_basic(ray_start_regular_shared):
+def test_ray_dask_basic(ray_start_1_cpu):
     from ray.util.dask import ray_dask_get
 
     @ray.remote
@@ -36,7 +44,7 @@ def test_ray_dask_basic(ray_start_regular_shared):
 
 
 @unittest.skipIf(sys.platform == "win32", "Failing on Windows.")
-def test_ray_dask_persist(ray_start_regular_shared):
+def test_ray_dask_persist(ray_start_1_cpu):
     from ray.util.dask import ray_dask_get
     arr = da.ones(5) + 2
     result = arr.persist(scheduler=ray_dask_get)