examples/numpy_visdom_logger_example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Example\n",
    "\n",
    "First you have to start the visdom server\n",
    "```\n",
    "python -m visdom.server -port 8080\n",
    "```\n",
    "and go to `http://localhost:8080/`.\n",
    "\n",
    "Then you can run the following code and get a line plot and an image plot.\n",
    "\n",
    "After starting the code you have to select `my_environment` in the environment dropdown in your browser. Maybe you have to refresh the page to see it."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot a value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from trixi.logger import NumpyVisdomLogger as Nvl\n",
    "from time import sleep\n",
    "import numpy as np\n",
    "\n",
    "nvl = Nvl(name=\"my_environment\")\n",
    "\n",
    "x = 0\n",
    "for i in range(50):\n",
    "    y = np.sin(x)\n",
    "    #Add one value to the plot. Setting x-axis values is currently not supported.\n",
    "    nvl.show_value(y, name=\"sin plot\")\n",
    "    x += 0.1\n",
    "    sleep(0.1)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot image and text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from trixi.logger import NumpyVisdomLogger as Nvl\n",
    "from time import sleep\n",
    "import numpy as np\n",
    "from skimage.data import camera, checkerboard, chelsea, astronaut\n",
    "\n",
    "nvl = Nvl(name=\"my_environment\")\n",
    "\n",
    "images = [astronaut(), camera(), checkerboard(), chelsea()]\n",
    "\n",
    "for i in range(50):\n",
    "    img = images[i%len(images)] #use modulo to cycle images\n",
    "    \n",
    "    #color channel has to be in first dimension\n",
    "    if len(img.shape) == 3:\n",
    "        img = img.transpose((2,0,1)) #move color channel from last to first dimension\n",
    "    \n",
    "    nvl.show_image(img, name=\"img plot\", title=\"image title\")\n",
    "    nvl.show_text(\"Iteration index: {}\".format(i), name=\"text plot\")\n",
    "    sleep(0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot image grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from trixi.logger import NumpyVisdomLogger as Nvl\n",
    "from time import sleep\n",
    "import numpy as np\n",
    "from skimage.data import camera, checkerboard, chelsea, astronaut, clock, coffee, coins, horse, moon, rocket\n",
    "\n",
    "nvl = Nvl(name=\"my_environment\")\n",
    "\n",
    "#Prepare images\n",
    "images = [astronaut(), camera(), checkerboard(), chelsea(), clock(), coffee(), coins(), horse(), moon(), rocket()]\n",
    "images_converted = []\n",
    "for img in images:\n",
    "    img = img[:200,:200] #make all images have same dimensions\n",
    "    if len(img.shape) == 3:\n",
    "        img = img[:,:,0]    #if rgb image use only one channel\n",
    "    images_converted.append(img)\n",
    "images_converted = np.array(images_converted)\n",
    "s = images_converted.shape # (10, 200, 200)\n",
    "images_converted = np.reshape(images_converted, (s[0], 1, s[1], s[2])) # images need to have channel dimension in the beginning. New shape: (10, 1, 200, 200)\n",
    "\n",
    "#Show images\n",
    "nvl.show_images(images_converted, name=\"img grid plot\", title=\"image grid title\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot Histogram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from trixi.logger import NumpyVisdomLogger as Nvl\n",
    "from time import sleep\n",
    "import numpy as np\n",
    "import random\n",
    "\n",
    "nvl = Nvl(name=\"my_environment\")\n",
    "\n",
    "values = [random.gauss(0, 1)]\n",
    "for i in range(500):\n",
    "    values.append(random.gauss(0, 1))\n",
    "    nvl.show_histogram(np.array(values), name=\"histogram plot\")\n",
    "    sleep(0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot 3D Histogram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from trixi.logger import NumpyVisdomLogger as Nvl\n",
    "from time import sleep\n",
    "import numpy as np\n",
    "import random\n",
    "\n",
    "nvl = Nvl(name=\"my_environment\")\n",
    "\n",
    "for i in range(5):\n",
    "    new_ar = np.random.normal(0, 1, 2000)\n",
    "    nvl.show_histogram_3d(new_ar, name=\"histogram 3D plot\")  # will multiply values by 100 to make plots better visible\n",
    "    sleep(0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}