Kdd_rl.ipynb

kdd_rl.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Kdd_rl.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/taposh/35e99ebe489e7679cd83aec011c08452/kdd_rl.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "R91c9ZDZtc-y",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# **KDD Competition - RL on Hexagon-ml**"
      ]
    },
    {
      "metadata": {
        "id": "_TbWx66Xt0vj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "9e4e7ee7-f4d8-46a8-ebed-d5a8a1c794a5"
      },
      "cell_type": "code",
      "source": [
        "\n",
        "!pip3 uninstall netsapi\n",
        "\n"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Uninstalling netsapi-1.0:\n",
            "  Would remove:\n",
            "    /root/.local/lib/python3.6/site-packages/netsapi-1.0.dist-info/*\n",
            "    /root/.local/lib/python3.6/site-packages/netsapi/*\n",
            "Proceed (y/n)? y\n",
            "  Successfully uninstalled netsapi-1.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "rj8CY7UstXSk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "9f61c556-691b-436d-92b4-bf317a9f7fb2"
      },
      "cell_type": "code",
      "source": [
        "!pip3 install git+https://github.com/slremy/netsapi --user --upgrade\n",
        "  \n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from collections import OrderedDict\n",
        "\n",
        "from netsapi.challenge import *\n",
        "from sys import exit, exc_info, argv\n",
        "\n",
        "class CustomAgent:\n",
        "    def __init__(self, environment, popsize=10):\n",
        "        self.popsize=popsize\n",
        "        self.environment = environment\n",
        "        self.popsize = popsize\n",
        "\n",
        "            \n",
        "        self.episodes = []\n",
        "        self.scores = []\n",
        "        self.policies = []\n",
        "\n",
        "    def generate(self):\n",
        "        best_policy = None\n",
        "        best_reward = -float('Inf')\n",
        "        try:\n",
        "            # select a set of random candidate solutions to be evaluated\n",
        "            policies = np.random.rand(self.popsize, self.environment.policyDimension)\n",
        "            rewards = self.environment.evaluateReward(policies)\n",
        "            best_policy = policies[np.argmax(rewards),:]\n",
        "        \n",
        "        except (KeyboardInterrupt, SystemExit):\n",
        "            print(exc_info())\n",
        "            \n",
        "        return best_policy\n",
        "\n",
        "    def scoringFunction(self):\n",
        "        scores = []\n",
        "        for ii in range(10):\n",
        "            self.environment.reset()\n",
        "            finalresult = self.generate()\n",
        "            self.policies.append(finalresult)\n",
        "            reward = self.environment.evaluateReward(finalresult)\n",
        "            self.scores.append(reward)\n",
        "            self.episodes.append(ii)\n",
        "\n",
        "        return np.mean(self.scores)/np.std(self.scores)\n",
        " \n",
        "        \n",
        "    def create_submissions(self, filename = 'my_submission.csv'):\n",
        "        labels = ['episode_no', 'reward', 'policy']\n",
        "        rewards = np.array(self.scores)\n",
        "        data = { 'episode_no': self.episodes, \n",
        "                 'rewards': rewards,\n",
        "                 'policy': self.policies,\n",
        "               }\n",
        "        submission_file = pd.DataFrame(data)\n",
        "        submission_file.to_csv(filename, index=False)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting git+https://github.com/slremy/netsapi\n",
            "  Cloning https://github.com/slremy/netsapi to /tmp/pip-req-build-vlpi2t_u\n",
            "Building wheels for collected packages: netsapi\n",
            "  Building wheel for netsapi (setup.py) ... \u001b[?25ldone\n",
            "\u001b[?25h  Stored in directory: /tmp/pip-ephem-wheel-cache-8ye6c3l1/wheels/9e/73/c9/86a9cc2460e11b3ce5b0a5ebd2d9d332a68afe0941659967fa\n",
            "Successfully built netsapi\n",
            "Installing collected packages: netsapi\n",
            "  Found existing installation: netsapi 1.0\n",
            "    Uninstalling netsapi-1.0:\n",
            "      Successfully uninstalled netsapi-1.0\n",
            "Successfully installed netsapi-1.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "-v9BQpLrtYKP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "7a2cdaa2-c322-40b6-b2b1-94ce9c1b8997"
      },
      "cell_type": "code",
      "source": [
        "env = ChallengeEnvironment(experimentCount = 1000)\n",
        "a = CustomAgent(env)\n",
        "a.scoringFunction()\n",
        "a.create_submissions(\"test.csv\")"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n",
            "1000  exps left\n",
            "990  exps left\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}