{
 "cells": [
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "# Passive Algorithms",
   "id": "eb9fb4a1a68e633"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "AMLGym includes several algorithms for passive learning domain models from full/partial/noisy trajectories.",
   "id": "1ac309b575ed4764"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": "!pip install amlgym",
   "id": "3a2c1c2a35c2e0f2"
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "initial_id",
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from amlgym.benchmarks import get_domain_path, get_trajectories_path\n",
    "from amlgym.algorithms import print_algorithms, get_algorithm\n",
    "print_algorithms()"
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Example 1: passive learning with full observability",
   "id": "bf47f050c2091de8"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Instantiate an algorithm for passive learning domain models in fully observable environments (e.g., [SAM](https://proceedings.kr.org/2021/36/)):",
   "id": "79155f5492207006"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": "learner = get_algorithm('sam')",
   "id": "fd36b0d1fc890673"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Get an empty IPC domain:",
   "id": "b2269aec54a3bbdf"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "from amlgym.util.util import empty_domain\n",
    "\n",
    "domain_path = get_domain_path('blocksworld')\n",
    "\n",
    "domain_empty_path = empty_domain(domain_path)\n",
    "with open(domain_empty_path, 'r') as f:\n",
    "    print(f.read())"
   ],
   "id": "9f69a6f17bb858d8"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Get a set of trajectories associated with the domain:",
   "id": "36f783186cfae5ea"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": "traj_paths = get_trajectories_path('blocksworld')",
   "id": "89539b7fd80ffd2e"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Learn a domain model specified in PDDL:",
   "id": "cf8aec8f3208ce17"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "model = learner.learn(domain_empty_path, traj_paths)\n",
    "print(model)"
   ],
   "id": "61ccca1a6d88ba94"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Example 2: passive learning with partial observability",
   "id": "9743932cbe218fc2"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Instantiate an algorithm for passive learning domain models in partially observable environments (e.g., [OffLAM](https://www.sciencedirect.com/science/article/abs/pii/S0004370224001929)):",
   "id": "bab53c77635504ca"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": "learner = get_algorithm('offlam')",
   "id": "43e369f4736b8347"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Get an empty IPC domain and associated set of trajectories:",
   "id": "539d621d2da27455"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "domain_path = get_domain_path('blocksworld')\n",
    "domain_empty_path = empty_domain(domain_path)\n",
    "traj_paths = get_trajectories_path('blocksworld')"
   ],
   "id": "635700869294d7e9"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Learn a domain model specified in PDDL:",
   "id": "5420c1797cfeea5d"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "model = learner.learn(domain_empty_path, traj_paths)\n",
    "print(model)"
   ],
   "id": "44fe1f41db91c3c4"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Example 3: passive learning with noisy observability",
   "id": "2d2085c200ee951"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Instantiate an algorithm for passive learning domain models in noisy environments (e.g., [NOLAM](https://ojs.aaai.org/index.php/ICAPS/article/view/31493)):",
   "id": "6716e7dab6fea2f5"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": "learner = get_algorithm('nolam', noise=0.1)",
   "id": "5428b802bec37744"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Get an empty IPC domain and associated set of trajectories:",
   "id": "c4e7bae09e022dc4"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "domain_path = get_domain_path('blocksworld')\n",
    "domain_empty_path = empty_domain(domain_path)\n",
    "traj_paths = get_trajectories_path('blocksworld')"
   ],
   "id": "e8fe833ef5eb441"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Learn a domain model specified in PDDL:",
   "id": "767927be0478cf87"
  },
  {
   "metadata": {},
   "cell_type": "code",
   "outputs": [],
   "execution_count": null,
   "source": [
    "model = learner.learn(domain_empty_path, traj_paths)\n",
    "print(model)"
   ],
   "id": "e118d39c95031ebb"
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}