ex_03_remo_meth.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A Guided Tour of Ray Core: Remote Classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[*Remote Classes*](https://docs.ray.io/en/latest/walkthrough.html#remote-classes-actors)\n",
    "involve using a `@ray.remote` decorator on a class. \n",
    "\n",
    "This implements an [*actor*](https://patterns.eecs.berkeley.edu/?page_id=258) pattern, with properties: *stateful*, *message-passing semantics*\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, let's start Ray…"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import logging\n",
    "import ray\n",
    "\n",
    "ray.init(\n",
    "    ignore_reinit_error=True,\n",
    "    logging_level=logging.ERROR,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Remote Classes example"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To start, we'll define a class and use the decorator:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "@ray.remote\n",
    "class Counter:\n",
    "    def __init__ (self):\n",
    "        self.value = 0\n",
    "\n",
    "    def increment (self):\n",
    "        self.value += 1\n",
    "        return self.value"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now use this class `Counter` to create an actor:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "counter = Counter.remote()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then call the actor:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "obj_ref = counter.increment.remote()\n",
    "ray.get(obj_ref)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use list comprehension to show the state being maintained in the actor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "f_list = [counter.increment.remote() for _ in range(3)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time \n",
    "\n",
    "print(ray.get(f_list))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Let's use another Actor class and create multiple instances associated with a distinct attribute, such as a name."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from random import randint\n",
    "\n",
    "@ray.remote\n",
    "class GoalsScored:\n",
    "    def __init__ (self, player) -> None:\n",
    "        self._goals = 0\n",
    "        self._player = player\n",
    "\n",
    "    def score (self) -> object:\n",
    "        self._goals += randint(1, 5)\n",
    "        return self._goals\n",
    "       \n",
    "    def player(self) -> str:\n",
    "        return self._player"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define three Actors: Rolando, Neymar, Messi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time \n",
    "\n",
    "ronaldo = GoalsScored.remote(\"Ronaldo\")\n",
    "neymar = GoalsScored.remote(\"Neymar\")\n",
    "messi = GoalsScored.remote(\"Messi\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "ronaldo_ref = ronaldo.score.remote()\n",
    "neymar_ref = neymar.score.remote()\n",
    "messi_ref  = messi.score.remote()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Again, use list comprehension to iterate over each Actor instances, along with object_ref\n",
    "for their scores, maintained by each distincgive actor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for ref, ref_obj in [(ronaldo, ronaldo_ref), (neymar, neymar_ref), (messi, messi_ref) ]:\n",
    "   print(f\"Player: {ray.get(ref.player.remote())} and goals scored: {ray.get(ref_obj)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, shutdown Ray"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ray.shutdown()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## References"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[\"A Universal Modular Actor Formalism for Artificial Intelligence\"](https://www.ijcai.org/Proceedings/73/Papers/027B.pdf)  \n",
    "Carl Hewitt, Peter Bishop, Richard Steiger  \n",
    "*IJCAI* (1973)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}