modulinstallation-in-jupyter.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exkurs: Installation von Modulen im Jupyter Notebook\n",
    "\n",
    "Es ist möglich, fehlende Module direkt aus einem Jupyter Notebook heraus zu installieren.\n",
    "\n",
    "Dafür bedient man sich bestimmter Kommandos, die es ermöglichen, auf die Systemebene zuzugreifen.\n",
    "\n",
    "Im Folgenden wird dies kurz demonstriert."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Wir nehmen an, wir wollen das Modul 'numpy' in unserem Code nutzen, stellen aber beim Import fest, dass dieses nicht vorhanden ist:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "ModuleNotFoundError",
     "evalue": "No module named 'numpy'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-d4cdadb62aa7>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mimport\u001b[0m \u001b[0mnumpy\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mModuleNotFoundError\u001b[0m: No module named 'numpy'"
     ]
    }
   ],
   "source": [
    "import numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Zunächst müssen wir das Modul `sys` für den Systemzugriff importieren:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dieses Paket bietet uns Zugriff auf Variablen, die wir für den Aufruf von Python oder Conda brauchen können:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'/home/mandy/anaconda3/bin/python'}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Variable, die auf die Python-Executable zeigt:\n",
    "{sys.executable}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'/home/mandy/anaconda3'}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Variable, die auf das Verzeichnis-Präfix der aktiven Python-Installation zeigt:\n",
    "{sys.prefix}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nun gibt es zwei Optionen, das fehlende Paket zu installieren.\n",
    "\n",
    "Die erste Variante ruft das Paketinstallationsprogramm `pip` auf:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting numpy\n",
      "  Using cached https://files.pythonhosted.org/packages/9b/af/4fc72f9d38e43b092e91e5b8cb9956d25b2e3ff8c75aed95df5569e4734e/numpy-1.17.4-cp37-cp37m-manylinux1_x86_64.whl\n",
      "Installing collected packages: numpy\n",
      "Successfully installed numpy-1.17.4\n"
     ]
    }
   ],
   "source": [
    "# entpricht /path/to/python -m pip install numpy\n",
    "!{sys.executable} -m pip install numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Verwendet man conda, ist auch der folgende Aufruf möglich:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting package metadata (repodata.json): done\n",
      "Solving environment: done\n",
      "\n",
      "## Package Plan ##\n",
      "\n",
      "  environment location: /home/mandy/anaconda3\n",
      "\n",
      "  added / updated specs:\n",
      "    - numpy\n",
      "\n",
      "\n",
      "The following packages will be downloaded:\n",
      "\n",
      "    package                    |            build\n",
      "    ---------------------------|-----------------\n",
      "    numpy-1.17.3               |   py37hd14ec0e_0           4 KB\n",
      "    numpy-base-1.17.3          |   py37hde5b4d6_0         5.2 MB\n",
      "    ------------------------------------------------------------\n",
      "                                           Total:         5.2 MB\n",
      "\n",
      "The following NEW packages will be INSTALLED:\n",
      "\n",
      "  blas               pkgs/main/linux-64::blas-1.0-mkl\n",
      "  intel-openmp       pkgs/main/linux-64::intel-openmp-2019.4-243\n",
      "  mkl                pkgs/main/linux-64::mkl-2019.4-243\n",
      "  mkl-service        pkgs/main/linux-64::mkl-service-2.3.0-py37he904b0f_0\n",
      "  mkl_fft            pkgs/main/linux-64::mkl_fft-1.0.15-py37ha843d7b_0\n",
      "  mkl_random         pkgs/main/linux-64::mkl_random-1.1.0-py37hd6b4f25_0\n",
      "  numpy              pkgs/main/linux-64::numpy-1.17.3-py37hd14ec0e_0\n",
      "  numpy-base         pkgs/main/linux-64::numpy-base-1.17.3-py37hde5b4d6_0\n",
      "\n",
      "\n",
      "\n",
      "Downloading and Extracting Packages\n",
      "numpy-base-1.17.3    | 5.2 MB    | ##################################### | 100% \n",
      "numpy-1.17.3         | 4 KB      | ##################################### | 100% \n",
      "Preparing transaction: done\n",
      "Verifying transaction: done\n",
      "Executing transaction: done\n"
     ]
    }
   ],
   "source": [
    "!conda install --yes --prefix {sys.prefix} numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nach der Installation funktioniert der Import:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conda-Spezialfall: Pakete, die nicht im Main Channel von Anaconda verfügbar sind\n",
    "\n",
    "Sollte ein Paket nicht im Main Channel von Anaconda verfügbar sein, muss man zusätzlich noch über ein Argument den alternativen Channel angeben. Den Namen des Channels findet man über die Suche des Paketes auf https://anaconda.org, z.B. https://anaconda.org/search?q=langdetect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting package metadata (repodata.json): done\n",
      "Solving environment: done\n",
      "\n",
      "## Package Plan ##\n",
      "\n",
      "  environment location: /home/mandy/anaconda3\n",
      "\n",
      "  added / updated specs:\n",
      "    - langdetect\n",
      "\n",
      "\n",
      "The following packages will be downloaded:\n",
      "\n",
      "    package                    |            build\n",
      "    ---------------------------|-----------------\n",
      "    certifi-2019.9.11          |           py37_0         147 KB  conda-forge\n",
      "    conda-4.7.12               |           py37_0         3.0 MB  conda-forge\n",
      "    ------------------------------------------------------------\n",
      "                                           Total:         3.1 MB\n",
      "\n",
      "The following NEW packages will be INSTALLED:\n",
      "\n",
      "  langdetect         conda-forge/noarch::langdetect-1.0.7-py_0\n",
      "\n",
      "The following packages will be SUPERSEDED by a higher-priority channel:\n",
      "\n",
      "  ca-certificates    pkgs/main::ca-certificates-2019.10.16~ --> conda-forge::ca-certificates-2019.9.11-hecc5488_0\n",
      "  certifi                                         pkgs/main --> conda-forge\n",
      "  conda                                           pkgs/main --> conda-forge\n",
      "  openssl              pkgs/main::openssl-1.1.1d-h7b6447c_3 --> conda-forge::openssl-1.1.1d-h516909a_0\n",
      "\n",
      "\n",
      "\n",
      "Downloading and Extracting Packages\n",
      "conda-4.7.12         | 3.0 MB    | ##################################### | 100% \n",
      "certifi-2019.9.11    | 147 KB    | ##################################### | 100% \n",
      "Preparing transaction: done\n",
      "Verifying transaction: done\n",
      "Executing transaction: done\n"
     ]
    }
   ],
   "source": [
    "!conda install --yes --prefix {sys.prefix} -c conda-forge langdetect"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}