Fix pre-built installation instructions (#248)

* Remove outdated comment about ownership handling

* Update pre-built installation instructions

* Add Getting Started documentation

* Update docs/src/getting-started.md

Co-authored-by: Glenn Moynihan <glennmoy@gmail.com>

---------

Co-authored-by: Glenn Moynihan <glennmoy@gmail.com>

docs/make.jl

@@ -4,7 +4,8 @@ using Documenter
 pages = ["API Documentation" => "index.md",
          "Installation" => "installation.md",
          "Developer Guide" => "developer-guide.md",
-         "Building Artifacts" => "building-artifacts.md"]
+         "Building Artifacts" => "building-artifacts.md"
+         "Getting Started" => "getting-started.md"]
 
 makedocs(; modules=[Ray],
          format=Documenter.HTML(; prettyurls=get(ENV, "CI", nothing) == "true"),

docs/src/getting-started.md

@@ -0,0 +1,41 @@
+# Getting Started
+
+To get started with using Ray.jl you'll need to start by creating a Ray cluster. You can start by creating a local cluster by using:
+
+```sh
+ray start --head
+```
+
+After that you can run your Ray application in Julia. These applications start with the following commands to connect the Julia process to the Ray cluster:
+
+```julia
+using Ray
+Ray.init()
+```
+
+A trivial Ray.jl application could look like:
+
+```julia
+using Ray
+Ray.init()
+
+# Define the square task.
+square = x -> x * x
+
+# Launch four parallel square tasks. Each task returns an object reference.
+refs = [Ray.submit_task(square, (i,)) for i in 0:3]
+
+# Retrieve results.
+Ray.get.(refs)
+# [0, 1, 4, 9]
+```
+
+Once you have finished running Ray applications you can shut down your cluster by using:
+
+```sh
+ray stop
+```
+
+## Further Reading
+
+The [official Ray Core documentation](https://docs.ray.io/en/latest/ray-core/walkthrough.html) is a great resource for learning more about Ray. That documentation primarily provides Python examples which use a similar syntax to that of Ray.jl. Remember that `Actor`s are not yet supported in Ray.jl.

docs/src/installation.md

@@ -1,8 +1,13 @@
 # Installation
 
-Ultimately we aim to make installing Ray.jl as simple as running `Pkg.add("Ray")`. However, at the moment there are some manual steps are required to install Ray.jl. For users with machines using the Linux x86_64 or macOS aarch64 (Apple Silicon) platforms we have provided pre-built binaries for each Ray.jl release.
+Ultimately we aim to make installing Ray.jl as simple as running `Pkg.add("Ray")`. However, at the moment there are some manual steps are required to install Ray.jl. For users with machines using the Linux x86_64 or macOS aarch64 (Apple Silicon) platforms we have provided pre-built binary dependencies for each Ray.jl release.
 
-To install these dependencies and Ray.jl run the following:
+Additionally, the pre-built dependencies requires you are using the following versions of Julia and Python:
+
+- Julia 1.8 - 1.9 (Linux x86_64 / macOS aarch64)
+- Python 3.7 - 3.11
+
+To install the pre-built version of Ray.jl run:
 
 ```sh
 # Install the Ray CLI
@@ -15,8 +20,10 @@ ARCH=$(uname -m)
 RELEASE="ray-2.5.1+1"
 pip install -U "ray[default] @ https://github.com/beacon-biosignals/ray/releases/download/$RELEASE/${RELEASE%+*}-${PYTHON}-${OS}_${ARCH}.whl" "pydantic<2"
 
-# Install the Julia packages "ray_julia_jll" and "Ray"
-TAG="v0.1.0" julia -e 'using Pkg; Pkg.add(PackageSpec(url="https://github.com/beacon-biosignals/Ray.jl", rev=ENV["TAG"]))'
+ray start --head && ray stop
+
+# Install the Ray.jl Julia package
+julia -e 'using Pkg; Pkg.add(PackageSpec(url="https://github.com/beacon-biosignals/Ray.jl", rev="v0.0.4"))'
 ```
 
 Users attempting to use Ray.jl on other platforms can attempt to [build the package from source](./developer-guide.md).

src/object_ref.jl

@@ -98,9 +98,6 @@ function _get_ownership_info(obj_ref::ObjectRef)
     return owner_address, serialized_object_status
 end
 
-# TODO: this is not currently used pending investigation of how to properly handle ownership
-# see https://github.com/beacon-biosignals/Ray.jl/issues/77#issuecomment-1717675779
-# and https://github.com/beacon-biosignals/Ray.jl/pull/108
 function _register_ownership(obj_ref::ObjectRef, outer_obj_ref::Union{ObjectRef,Nothing},
                              owner_address::ray_jll.Address,
                              serialized_object_status::String)