Merge with ray master

BUILD.bazel

@@ -495,6 +495,7 @@ flatbuffer_py_library(
         "ConfigTableData.py",
         "CustomSerializerData.py",
         "DriverTableData.py",
+        "EntryType.py",
         "ErrorTableData.py",
         "ErrorType.py",
         "FunctionTableData.py",

README.rst

@@ -6,12 +6,12 @@
 .. image:: https://readthedocs.org/projects/ray/badge/?version=latest
     :target: http://ray.readthedocs.io/en/latest/?badge=latest
 
-.. image:: https://img.shields.io/badge/pypi-0.6.6-blue.svg
+.. image:: https://img.shields.io/badge/pypi-0.7.0-blue.svg
     :target: https://pypi.org/project/ray/
 
 |
 
-**Ray is a flexible, high-performance distributed execution framework.**
+**Ray is a fast and simple framework for building and running distributed applications.**
 
 
 Ray is easy to install: ``pip install ray``

bazel/ray.bzl

@@ -25,8 +25,10 @@ def flatbuffer_java_library(name, srcs, outs, out_prefix, includes = [], include
     )
 
 def define_java_module(name, additional_srcs = [], additional_resources = [], define_test_lib = False, test_deps = [], **kwargs):
+    lib_name = "org_ray_ray_" + name
+    pom_file_targets = [lib_name]
     native.java_library(
-        name = "org_ray_ray_" + name,
+        name = lib_name,
         srcs = additional_srcs + native.glob([name + "/src/main/java/**/*.java"]),
         resources = native.glob([name + "/src/main/resources/**"]) + additional_resources,
         **kwargs
@@ -40,8 +42,10 @@ def define_java_module(name, additional_srcs = [], additional_resources = [], de
         tags = ["checkstyle"],
     )
     if define_test_lib:
+        test_lib_name = "org_ray_ray_" + name + "_test"
+        pom_file_targets.append(test_lib_name)
         native.java_library(
-            name = "org_ray_ray_" + name + "_test",
+            name = test_lib_name,
             srcs = native.glob([name + "/src/test/java/**/*.java"]),
             deps = test_deps,
         )
@@ -53,12 +57,11 @@ def define_java_module(name, additional_srcs = [], additional_resources = [], de
             size = "small",
             tags = ["checkstyle"],
         )
-
-def gen_java_pom_file(name):
     pom_file(
         name = "org_ray_ray_" + name + "_pom",
-        targets = [
-            ":org_ray_ray_" + name,
-        ],
+        targets = pom_file_targets,
         template_file = name + "/pom_template.xml",
+        substitutions = {
+            "{auto_gen_header}": "<!-- This file is auto-generated by Bazel from pom_template.xml, do not modify it. -->",
+        },
     )

bazel/ray_deps_setup.bzl

@@ -80,18 +80,18 @@ def ray_deps_setup():
 
     http_archive(
         name = "io_opencensus_cpp",
-        strip_prefix = "opencensus-cpp-0.3.0",
-        urls = ["https://github.com/census-instrumentation/opencensus-cpp/archive/v0.3.0.zip"],
+        strip_prefix = "opencensus-cpp-3aa11f20dd610cb8d2f7c62e58d1e69196aadf11",
+        urls = ["https://github.com/census-instrumentation/opencensus-cpp/archive/3aa11f20dd610cb8d2f7c62e58d1e69196aadf11.zip"],
     )
 
     # OpenCensus depends on Abseil so we have to explicitly pull it in.
     # This is how diamond dependencies are prevented.
     git_repository(
         name = "com_google_absl",
-        commit = "88a152ae747c3c42dc9167d46c590929b048d436",
+        commit = "5b65c4af5107176555b23a638e5947686410ac1f",
         remote = "https://github.com/abseil/abseil-cpp.git",
     )
-   
+
     # OpenCensus depends on jupp0r/prometheus-cpp
     http_archive(
         name = "com_github_jupp0r_prometheus_cpp",

ci/jenkins_tests/perf_integration_tests/run_perf_integration.sh

@@ -9,7 +9,7 @@ pushd "$ROOT_DIR"
 
 python -m pip install pytest-benchmark
 
-pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev3-cp27-cp27mu-manylinux1_x86_64.whl
+pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp27-cp27mu-manylinux1_x86_64.whl
 python -m pytest --benchmark-autosave --benchmark-min-rounds=10 --benchmark-columns="min, max, mean" $ROOT_DIR/../../../python/ray/tests/perf_integration_tests/test_perf_integration.py
 
 pushd $ROOT_DIR/../../../python

ci/jenkins_tests/run_rllib_tests.sh

@@ -289,6 +289,9 @@ docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
 docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
     /ray/ci/suppress_output python /ray/python/ray/rllib/tests/test_local.py
 
+docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
+    /ray/ci/suppress_output python /ray/python/ray/rllib/tests/test_dependency.py
+
 docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
     /ray/ci/suppress_output python /ray/python/ray/rllib/tests/test_legacy.py
 
@@ -365,6 +368,9 @@ docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
 docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
     /ray/ci/suppress_output python /ray/python/ray/rllib/examples/multiagent_cartpole.py --num-iters=2
 
+docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
+    /ray/ci/suppress_output python /ray/python/ray/rllib/examples/multiagent_cartpole.py --num-iters=2 --simple
+
 docker run --rm --shm-size=${SHM_SIZE} --memory=${MEMORY_SIZE} $DOCKER_SHA \
     /ray/ci/suppress_output python /ray/python/ray/rllib/examples/multiagent_two_trainers.py --num-iters=2
 

ci/long_running_tests/config.yaml

@@ -49,6 +49,7 @@ setup_commands:
     # - sudo apt-get update
     # - sudo apt-get install -y build-essential curl unzip
     # - git clone https://github.com/ray-project/ray || true
+    # - ray/ci/travis/install-bazel.sh
     # - cd ray/python; git checkout master; git pull; pip install -e . --verbose
     # Install nightly Ray wheels.
     - pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/<<RAY_RELEASE_HASH>>/ray-<<RAY_RELEASE_VERSION>>-cp36-cp36m-manylinux1_x86_64.whl

ci/stress_tests/application_cluster_template.yaml

@@ -90,8 +90,8 @@ file_mounts: {
 # List of shell commands to run to set up nodes.
 setup_commands:
     - echo 'export PATH="$HOME/anaconda3/envs/tensorflow_<<<PYTHON_VERSION>>>/bin:$PATH"' >> ~/.bashrc
-    - ray || wget https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-<<<WHEEL_STR>>>-manylinux1_x86_64.whl
-    - rllib || pip install -U ray-0.7.0.dev2-<<<WHEEL_STR>>>-manylinux1_x86_64.whl[rllib]
+    - ray || wget https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-<<<WHEEL_STR>>>-manylinux1_x86_64.whl
+    - rllib || pip install -U ray-0.8.0.dev0-<<<WHEEL_STR>>>-manylinux1_x86_64.whl[rllib]
     - pip install tensorflow-gpu==1.12.0
     - echo "sudo halt" | at now + 60 minutes
     # Consider uncommenting these if you also want to run apt-get commands during setup

ci/stress_tests/stress_testing_config.yaml

@@ -98,9 +98,10 @@ setup_commands:
     - echo 'export PATH="$HOME/anaconda3/bin:$PATH"' >> ~/.bashrc
     # # Build Ray.
     # - git clone https://github.com/ray-project/ray || true
+    # - ray/ci/travis/install-bazel.sh
     - pip install boto3==1.4.8 cython==0.29.0
     # - cd ray/python; git checkout master; git pull; pip install -e . --verbose
-    - pip install https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp36-cp36m-manylinux1_x86_64.whl
+    - pip install https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp36-cp36m-manylinux1_x86_64.whl
     - echo "sudo halt" | at now + 60 minutes
 
 # Custom commands that will be run on the head node after common setup.

ci/stress_tests/test_many_tasks_and_transfers.py

@@ -24,10 +24,10 @@
 
 # Wait until the expected number of nodes have joined the cluster.
 while True:
-    if len(ray.global_state.client_table()) >= num_remote_nodes + 1:
+    if len(ray.nodes()) >= num_remote_nodes + 1:
         break
 logger.info("Nodes have all joined. There are %s resources.",
-            ray.global_state.cluster_resources())
+            ray.cluster_resources())
 
 
 # Require 1 GPU to force the tasks to be on remote machines.

dev/RELEASE_PROCESS.rst

@@ -41,12 +41,12 @@ This document describes the process for creating new releases.
 
 6. **Download all the wheels:** Now the release is ready to begin final
    testing. The wheels are automatically uploaded to S3, even on the release
-   branch. The wheels can ``pip install``ed from the following URLs:
+   branch. To test, ``pip install`` from the following URLs:
 
    .. code-block:: bash
 
        export RAY_HASH=...  # e.g., 618147f57fb40368448da3b2fb4fd213828fa12b
-       export RAY_VERSION=...  # e.g., 0.6.6
+       export RAY_VERSION=...  # e.g., 0.7.0
        pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/$RAY_HASH/ray-$RAY_VERSION-cp27-cp27mu-manylinux1_x86_64.whl
        pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/$RAY_HASH/ray-$RAY_VERSION-cp35-cp35m-manylinux1_x86_64.whl
        pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/$RAY_HASH/ray-$RAY_VERSION-cp36-cp36m-manylinux1_x86_64.whl
@@ -120,10 +120,28 @@ This document describes the process for creating new releases.
       git pull origin master --tags
       git log $(git describe --tags --abbrev=0)..HEAD --pretty=format:"%s" | sort
 
+11. **Bump version on Ray master branch:** Create a pull request to increment the
+    version of the master branch. The format of the new version is as follows:
+
+    New minor release (e.g., 0.7.0): Increment the minor version and append ``.dev0`` to
+    the version. For example, if the version of the new release is 0.7.0, the master
+    branch needs to be updated to 0.8.0.dev0. `Example PR for minor release`
+
+    New micro release (e.g., 0.7.1): Increment the ``dev`` number, such that the number
+    after ``dev`` equals the micro version. For example, if the version of the new
+    release is 0.7.1, the master branch needs to be updated to 0.8.0.dev1.
+
+12. **Update version numbers throughout codebase:** Suppose we just released 0.7.1. The
+    previous release version number (in this case 0.7.0) and the previous dev version number
+    (in this case 0.8.0.dev0) appear in many places throughout the code base including
+    the installation documentation, the example autoscaler config files, and the testing
+    scripts. Search for all of the occurrences of these version numbers and update them to
+    use the new release and dev version numbers.
+
 .. _documentation: https://ray.readthedocs.io/en/latest/installation.html#trying-snapshots-from-master
 .. _`documentation for building wheels`: https://github.com/ray-project/ray/blob/master/python/README-building-wheels.md
 .. _`ci/stress_tests/run_stress_tests.sh`: https://github.com/ray-project/ray/blob/master/ci/stress_tests/run_stress_tests.sh
 .. _`ci/stress_tests/run_application_stress_tests.sh`: https://github.com/ray-project/ray/blob/master/ci/stress_tests/run_application_stress_tests.sh
 .. _`this example`: https://github.com/ray-project/ray/pull/4226
-.. _`these wheels here`: https://ray.readthedocs.io/en/latest/installation.html
 .. _`GitHub website`: https://github.com/ray-project/ray/releases
+.. _`Example PR for minor release`: https://github.com/ray-project/ray/pull/4845

doc/source/api.rst

@@ -27,6 +27,26 @@ The Ray API
 
 .. autofunction:: ray.method
 
+Inspect the Cluster State
+-------------------------
+
+.. autofunction:: ray.nodes()
+
+.. autofunction:: ray.tasks()
+
+.. autofunction:: ray.objects()
+
+.. autofunction:: ray.timeline()
+
+.. autofunction:: ray.object_transfer_timeline()
+
+.. autofunction:: ray.cluster_resources()
+
+.. autofunction:: ray.available_resources()
+
+.. autofunction:: ray.errors()
+
+
 The Ray Command Line API
 ------------------------
 

doc/source/conf.py

@@ -26,6 +26,7 @@
     "ray.core.generated.ActorCheckpointIdData",
     "ray.core.generated.ClientTableData",
     "ray.core.generated.DriverTableData",
+    "ray.core.generated.EntryType",
     "ray.core.generated.ErrorTableData",
     "ray.core.generated.ErrorType",
     "ray.core.generated.GcsTableEntry",

doc/source/development.rst

@@ -60,7 +60,7 @@ Python script with the following:
 .. code-block:: bash
 
  RAY_RAYLET_GDB=1 RAY_RAYLET_TMUX=1 python
- 
+
 You can then list the ``tmux`` sessions with ``tmux ls`` and attach to the
 appropriate one.
 
@@ -71,17 +71,17 @@ allow core dump files to be written.
 
 Inspecting Redis shards
 ~~~~~~~~~~~~~~~~~~~~~~~
-To inspect Redis, you can use the ``ray.experimental.state.GlobalState`` Python
-API.  The easiest way to do this is to start or connect to a Ray cluster with
-``ray.init()``, then query the API like so:
+To inspect Redis, you can use the global state API. The easiest way to do this
+is to start or connect to a Ray cluster with ``ray.init()``, then query the API
+like so:
 
 .. code-block:: python
 
  ray.init()
- ray.worker.global_state.client_table()
+ ray.nodes()
  # Returns current information about the nodes in the cluster, such as:
  # [{'ClientID': '2a9d2b34ad24a37ed54e4fcd32bf19f915742f5b',
- #   'IsInsertion': True,
+ #   'EntryType': 0,
  #   'NodeManagerAddress': '1.2.3.4',
  #   'NodeManagerPort': 43280,
  #   'ObjectManagerPort': 38062,

doc/source/index.rst

@@ -7,7 +7,7 @@ Ray
     <a href="https://github.com/ray-project/ray"><img style="position: absolute; top: 0; right: 0; border: 0;" src="https://camo.githubusercontent.com/365986a132ccd6a44c23a9169022c0b5c890c387/68747470733a2f2f73332e616d617a6f6e6177732e636f6d2f6769746875622f726962626f6e732f666f726b6d655f72696768745f7265645f6161303030302e706e67" alt="Fork me on GitHub" data-canonical-src="https://s3.amazonaws.com/github/ribbons/forkme_right_red_aa0000.png"></a>
   </embed>
 
-*Ray is a flexible, high-performance distributed execution framework.*
+*Ray is a fast and simple framework for building and running distributed applications.*
 
 
 Ray is easy to install: ``pip install ray``
@@ -98,10 +98,10 @@ Ray comes with libraries that accelerate deep learning and reinforcement learnin
    rllib-models.rst
    rllib-algorithms.rst
    rllib-offline.rst
-   rllib-dev.rst
    rllib-concepts.rst
-   rllib-package-ref.rst
    rllib-examples.rst
+   rllib-dev.rst
+   rllib-package-ref.rst
 
 .. toctree::
    :maxdepth: 1

doc/source/installation.rst

@@ -33,14 +33,14 @@ Here are links to the latest wheels (which are built off of master). To install
 ===================  ===================
 
 
-.. _`Linux Python 3.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp37-cp37m-manylinux1_x86_64.whl
-.. _`Linux Python 3.6`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp36-cp36m-manylinux1_x86_64.whl
-.. _`Linux Python 3.5`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp35-cp35m-manylinux1_x86_64.whl
-.. _`Linux Python 2.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp27-cp27mu-manylinux1_x86_64.whl
-.. _`MacOS Python 3.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp37-cp37m-macosx_10_6_intel.whl
-.. _`MacOS Python 3.6`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp36-cp36m-macosx_10_6_intel.whl
-.. _`MacOS Python 3.5`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp35-cp35m-macosx_10_6_intel.whl
-.. _`MacOS Python 2.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.7.0.dev2-cp27-cp27m-macosx_10_6_intel.whl
+.. _`Linux Python 3.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp37-cp37m-manylinux1_x86_64.whl
+.. _`Linux Python 3.6`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp36-cp36m-manylinux1_x86_64.whl
+.. _`Linux Python 3.5`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp35-cp35m-manylinux1_x86_64.whl
+.. _`Linux Python 2.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp27-cp27mu-manylinux1_x86_64.whl
+.. _`MacOS Python 3.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp37-cp37m-macosx_10_6_intel.whl
+.. _`MacOS Python 3.6`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp36-cp36m-macosx_10_6_intel.whl
+.. _`MacOS Python 3.5`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp35-cp35m-macosx_10_6_intel.whl
+.. _`MacOS Python 2.7`: https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev0-cp27-cp27m-macosx_10_6_intel.whl
 
 
 Building Ray from source

doc/source/internals-overview.rst

@@ -66,32 +66,6 @@ listens for the addition of remote functions to the centralized control state.
 When a new remote function is added, the thread fetches the pickled remote
 function, unpickles it, and can then execute that function.
 
-Notes and limitations
-~~~~~~~~~~~~~~~~~~~~~
-
-- Because we export remote functions as soon as they are defined, that means
-  that remote functions can't close over variables that are defined after the
-  remote function is defined. For example, the following code gives an error.
-
-  .. code-block:: python
-
-    @ray.remote
-    def f(x):
-        return helper(x)
-
-    def helper(x):
-        return x + 1
-
-  If you call ``f.remote(0)``, it will give an error of the form.
-
-  .. code-block:: python
-
-    Traceback (most recent call last):
-        File "<ipython-input-3-12a5beeb2306>", line 3, in f
-    NameError: name 'helper' is not defined
-
-  On the other hand, if ``helper`` is defined before ``f``, then it will work.
-
 Calling a remote function
 -------------------------
 

doc/source/rllib-algorithms.rst

@@ -95,7 +95,7 @@ Asynchronous Proximal Policy Optimization (APPO)
 `[implementation] <https://github.com/ray-project/ray/blob/master/python/ray/rllib/agents/ppo/appo.py>`__
 We include an asynchronous variant of Proximal Policy Optimization (PPO) based on the IMPALA architecture. This is similar to IMPALA but using a surrogate policy loss with clipping. Compared to synchronous PPO, APPO is more efficient in wall-clock time due to its use of asynchronous sampling. Using a clipped loss also allows for multiple SGD passes, and therefore the potential for better sample efficiency compared to IMPALA. V-trace can also be enabled to correct for off-policy samples.
 
-This implementation is currently *experimental*. Consider also using `PPO <rllib-algorithms.html#proximal-policy-optimization-ppo>`__ or `IMPALA <rllib-algorithms.html#importance-weighted-actor-learner-architecture-impala>`__.
+APPO is not always more efficient; it is often better to simply use `PPO <rllib-algorithms.html#proximal-policy-optimization-ppo>`__ or `IMPALA <rllib-algorithms.html#importance-weighted-actor-learner-architecture-impala>`__.
 
 Tuned examples: `PongNoFrameskip-v4 <https://github.com/ray-project/ray/blob/master/python/ray/rllib/tuned_examples/pong-appo.yaml>`__
 
@@ -274,7 +274,7 @@ QMIX Monotonic Value Factorisation (QMIX, VDN, IQN)
 ---------------------------------------------------
 `[paper] <https://arxiv.org/abs/1803.11485>`__ `[implementation] <https://github.com/ray-project/ray/blob/master/python/ray/rllib/agents/qmix/qmix.py>`__ Q-Mix is a specialized multi-agent algorithm. Code here is adapted from https://github.com/oxwhirl/pymarl_alpha  to integrate with RLlib multi-agent APIs. To use Q-Mix, you must specify an agent `grouping <rllib-env.html#grouping-agents>`__ in the environment (see the `two-step game example <https://github.com/ray-project/ray/blob/master/python/ray/rllib/examples/twostep_game.py>`__). Currently, all agents in the group must be homogeneous. The algorithm can be scaled by increasing the number of workers or using Ape-X.
 
-Q-Mix is implemented in `PyTorch <https://github.com/ray-project/ray/blob/master/python/ray/rllib/agents/qmix/qmix_policy_graph.py>`__ and is currently *experimental*.
+Q-Mix is implemented in `PyTorch <https://github.com/ray-project/ray/blob/master/python/ray/rllib/agents/qmix/qmix_policy.py>`__ and is currently *experimental*.
 
 Tuned examples: `Two-step game <https://github.com/ray-project/ray/blob/master/python/ray/rllib/examples/twostep_game.py>`__