extra-03-root-in-jupyter.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ROOT in Jupyter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "ROOT can be used in Jupyter notebooks, both in Python and C++. In this course we will focus only on Python, but for people interested in ROOT C++ notebooks some examples can be found [here](https://swan-gallery.web.cern.ch/root_primer/).\n",
    "\n",
    "There are some specificities and extra features available when running ROOT from a notebook, and that's what will be covered in this section!\n",
    "\n",
    "In Python, everything starts with importing ROOT as we already saw in [02-root-python-cpp.ipynb](02-root-python-cpp.ipynb)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "import ROOT"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Quickly checking the docs\n",
    "\n",
    "To take a quick peek at a function signature, we can use the notebook's ? feature:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "?ROOT.TH1D.FillRandom"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## ROOT Graphics in a notebook\n",
    "\n",
    "There are two modes in which we can visualize ROOT plots in a notebook\n",
    "* Static images (default)\n",
    "* Interactive JavaScript displays\n",
    "\n",
    "To see an example, let's first create a histogram and fill it with random numbers following a Gaussian distribution:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "h = ROOT.TH1D(\"my_histo\", \"Example histogram\", 100, -4, 4)\n",
    "h.FillRandom(\"gaus\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Now let's plot the histogram. For that purpose, we first need to create a canvas, then draw the histogram and finally draw the canvas. See how the output is a static image."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "c = ROOT.TCanvas()\n",
    "h.Draw()\n",
    "c.Draw() # don't forget to draw the canvas too!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Interactive graphics\n",
    "\n",
    "As mentioned before, we can also display interactive ROOT graphics in a notebook. To do that, we can use the `%jsroot` magic and turn on the JavaScript mode."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%jsroot on"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "If we draw again the same canvas after activating JS graphics, this time the histogram plot is interactive! Try to hover over it with the mouse and zoom in and out."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "c.Draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Combining Python and C++ in the same notebook\n",
    "\n",
    "As we saw in [02-root-python-cpp.ipynb](02-root-python-cpp.ipynb), PyROOT lets us define C++ and call it from Python right away."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "ROOT.gInterpreter.ProcessLine(\"void print_integer(int i) { std::cout << i << std::endl; }\")\n",
    "\n",
    "ROOT.print_integer(7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "In a notebook, we can use the `%%cpp` magic to do the same. If `%%cpp` is present in a cell, its whole content is interpreted and executed as C++, and it has C++ syntax coloring!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%cpp\n",
    "void print_integer_2(int i) {\n",
    "   std::cout << i << std::endl;\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "The function we just defined in the previous (C++) cell can be now invoked from the next (Python) cell:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "ROOT.print_integer_2(7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Note that, besides defining new C++ entities, we can also run C++ statements."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%cpp\n",
    "std::cout << \"Hello from C++!\" << std::endl;\n",
    "\n",
    "print_integer_2(7);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Redefining C++ entities\n",
    "\n",
    "In C++, redefinition of entities such as variables is not allowed. However, ROOT provides a cool redefinition feature that is especially useful in interactive interfaces like notebooks. This feature works for types, functions and variables defined in the global namespace.\n",
    "\n",
    "You can check it by running the next cells, even multiple times!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%cpp\n",
    "int foo = 0;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%cpp\n",
    "class foo {};"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%cpp\n",
    "void foo() {}"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}