notebooks/book1/02/discrete_prob_dist_plot.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Discrete Probability Distribution Plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import jax\n",
    "import jax.numpy as jnp\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "try:\n",
    "    from probml_utils import savefig, latexify\n",
    "except ModuleNotFoundError:\n",
    "    %pip install -qq git+https://github.com/probml/probml-utils.git\n",
    "    from probml_utils import savefig, latexify"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/patel_zeel/probml-utils/probml_utils/plotting.py:25: UserWarning: LATEXIFY environment variable not set, not latexifying\n",
      "  warnings.warn(\"LATEXIFY environment variable not set, not latexifying\")\n"
     ]
    }
   ],
   "source": [
    "latexify(width_scale_factor=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)\n",
      "/home/patel_zeel/probml-utils/probml_utils/plotting.py:65: UserWarning: set FIG_DIR environment variable to save figures\n",
      "  warnings.warn(\"set FIG_DIR environment variable to save figures\")\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Bar graphs showing a uniform discrete distribution and another with full mass on one value.\n",
    "\n",
    "N = 4\n",
    "\n",
    "\n",
    "def make_graph(probs, N, save_name, fig=None, ax=None):\n",
    "    x = jnp.arange(1, N + 1)\n",
    "    if fig is None:\n",
    "        fig, ax = plt.subplots()\n",
    "    ax.bar(x, probs, align=\"center\")\n",
    "    ax.set_xlim([min(x) - 0.5, max(x) + 0.5])\n",
    "    ax.set_xticks(x)\n",
    "    ax.set_yticks(jnp.linspace(0, 1, N + 1))\n",
    "    ax.set_xlabel(\"$x$\")\n",
    "    ax.set_ylabel(\"$Pr(X=x)$\")\n",
    "    sns.despine()\n",
    "    if len(save_name) > 0:\n",
    "        savefig(save_name)\n",
    "    return fig, ax\n",
    "\n",
    "\n",
    "uniform_probs = jnp.repeat(1.0 / N, N)\n",
    "_, _ = make_graph(\n",
    "    uniform_probs, N, \"uniform_histogram\"\n",
    ")  # Do not add .pdf or .png as it is automatically added by savefig method\n",
    "\n",
    "delta_probs = jnp.array([1, 0, 0, 0])\n",
    "_, _ = make_graph(delta_probs, N, \"delta_histogram\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Demo\n",
    "\n",
    "You can see different examples of discrete distributions by changing the seed in the following demo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7eb88c0e987446f7bf4f990bfde42c7a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(IntSlider(value=5, description='random_state', max=10, min=1), IntSlider(value=6, descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from ipywidgets import interact\n",
    "\n",
    "\n",
    "@interact(random_state=(1, 10), N=(2, 10))\n",
    "def generate_random(random_state, N):\n",
    "    key = jax.random.PRNGKey(random_state)\n",
    "    probs = jax.random.uniform(key, shape=(N,))\n",
    "    probs = probs / jnp.sum(probs)\n",
    "    fig, ax = make_graph(probs, N, \"\")\n",
    "    ax.set_yticks(jnp.linspace(0, 1, 11))"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
  },
  "kernelspec": {
   "display_name": "Python [conda env:pyprobml]",
   "language": "python",
   "name": "conda-env-pyprobml-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}