Immoscout Crawler.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Immoscout Crawler\n",
    "\n",
    "Loads Immoscout results from url, parse and display as text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The soup extension is already loaded. To reload it, use:\n",
      "  %reload_ext soup\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "import requests\n",
    "import time\n",
    "from datetime import datetime\n",
    "from IPython.display import display_pretty, clear_output\n",
    "%load_ext soup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "configure_ipython_beautifulsoup(show_html=False, show_css=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "urls = list()\n",
    "urls.append(\"http://www.immobilienscout24.de/Suche/S-2/Wohnung-Miete/Umkreissuche/Berlin/10967/228521/2509075/M_fcllenhoffstra_dfe/-/3/3,00-/75,00-/EURO--1100,00/-/-/false/-/-/-/-/-/-/-/-1940/-/-/-/-/1-\")\n",
    "urls.append(\"http://www.immobilienscout24.de/Suche/S-2/Wohnung-Miete/Umkreissuche/Berlin/10967/228521/2509075/M_fcllenhoffstra_dfe/-/3/4,00-/95,00-/EURO--1400,00/-/-/false/-/-/-/-/-/-/-/-1940/-/-/-/-/1-\")\n",
    "urls.append(\"http://www.immobilienscout24.de/Suche/S-2/Wohnung-Miete/Umkreissuche/Berlin/10551/223304/2512943/Bredowstra_dfe/-/1/3,00-/80,00-/EURO--1100,00/-/-/false/-/-/-/-/-/-/-/-1940/-/-/-/-/1-?enteredFrom=result_list\")\n",
    "urls.append(\"http://www.immobilienscout24.de/Suche/S-2/Wohnung-Miete/Umkreissuche/Berlin/10829/225175/2508329/Hohenfriedbergstra_dfe/-/1/3,00-/80,00-/EURO--1100,00/-/-/false/-/-/-/-/-/-/-/-1940/-/-/-/-/1-?enteredFrom=result_list\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "known_ids = set([\"84672385\",\"84607255\",\"65881497\",\"81156001\",\"74122894\",\"84441189\",\"84296913\",\"83082603\",\n",
    "             \"83968655\",\"75090852\",\"84487728\",\"84459163\",\"84441205\",\"73972731\",\"83966216\",\"83256999\",\n",
    "             \"84666872\",\"54065226\",\"84468125\",\"83875423\",\"83819962\",\"83947419\",\"83544107\",\"82957811\",\n",
    "             \"81024598\",\"55481227\",\"64694875\",\"75825617\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Find the ad's link in its entry\n",
    "get_link = lambda res: \"http://www.immobilienscout24.de{}\".format(res.find('a', {\"class\": \"logo_link\"}).get('href'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Return a list of ad texts from parsing a given url\n",
    "def result_list(url):\n",
    "    soup = p(url)\n",
    "    results = soup.find_all('div', {\"class\": 'result_entry_box'})\n",
    "    return [\"{}\\n{}\".format(\" \".join(res.stripped_strings).encode('utf-8'), get_link(res)) for i, res in enumerate(results)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "22:42:57.712692\n"
     ]
    }
   ],
   "source": [
    "run = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "try:\n",
    "    while True:\n",
    "        run += 1\n",
    "        print \"Run #{} {}\".format(run, datetime.now().time())\n",
    "        for url in urls:\n",
    "            for res in result_list(url):\n",
    "                id_matches = re.findall(\"\\d{8}\", res)\n",
    "                if id_matches[0] not in known_ids:\n",
    "                    !say \"Yeah! Found a new place, mhkae?\"\n",
    "                    print res\n",
    "                    known_ids.add(res)\n",
    "            time.sleep(1)\n",
    "        time.sleep(180)\n",
    "except Exception:\n",
    "    !say \"Oh crap! It's broken\"\n",
    "    print \"\\n\\nKnown ids: {}\\n\\n\".format(known_ids)\n",
    "    raise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}