cc-webgraph-statistics/topology_stats.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Common Crawl web graph statistics notebook\n",
    "\n",
    "This notebook will guide you through how to:\n",
    "- Load Common Crawl webgraph\n",
    "- Generate statistics about said webgraph\n",
    "- Interact with said webgraph\n",
    "\n",
    "Requirements:\n",
    "- Python\n",
    "- AWS CLI\n",
    "- Maven\n",
    "- Java 11 (or upwards)\n",
    "\n",
    "### Note: Please refer to the [README](README.md) and follow along here, as more details about each step will be provided in the README."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "git clone https://github.com/commoncrawl/cc-webgraph.git\n",
    "\n",
    "cd cc-webgraph\n",
    "\n",
    "mvn package;"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we clone an existing repo, `cc-webgraph` which provides us with some very useful scripts that will help us get the webgraph set up. We run `mvn package` as the Java tools (such as the WebGraph framework) in this repo are compiled and packaged with Maven."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "for f in cc-main-2020-feb-mar-may-domain-t.graph cc-main-2020-feb-mar-may-domain-t.properties \\\n",
    "         cc-main-2020-feb-mar-may-domain.graph cc-main-2020-feb-mar-may-domain.properties \\\n",
    "         cc-main-2020-feb-mar-may-domain.stats \\\n",
    "         cc-main-2020-feb-mar-may-domain-vertices.txt.gz; do\n",
    "   wget https://data.commoncrawl.org/projects/hyperlinkgraph/cc-main-2020-feb-mar-may/domain/$f\n",
    "   # alternatively, on AWS:\n",
    "   # aws s3 cp s3://commoncrawl/projects/hyperlinkgraph/cc-main-2020-feb-mar-may/domain/$f .\n",
    "done"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This will download a representation of the domain level graph (`cc-main-2020-feb-mar-may-domain.graph`), a file that contains useful properties of said graph (`cc-main-2020-feb-mar-may-domain.properties`) which is required to load the graph, a file that contains more statistics about the graph (`cc-main-2020-feb-mar-may-domain.stats`) such as min indegree/outdegrees, number of dangling nodes, etc, and a file that represents the vertices in the graph (`cc-main-2020-feb-mar-may-domain-vertices.txt.gz`). Transpose of the graph is also downloaded (which is any file name that ends with `-t`). More information can be found on the [blog](https://commoncrawl.org/2020/06/host-and-domain-level-web-graphs-febmarmay-2020/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "WG=\"java -Xmx12g -cp target/cc-webgraph-0.1-SNAPSHOT-jar-with-dependencies.jar\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This shell variable `WG` is the path to our jar file that will help us run the java commands. We have to allocate 12 gigabytes of heap space with `-Xmx12g` in order to run some of the more computationally expensive methods."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "$WG it.unimi.dsi.webgraph.BVGraph -O -L cc-main-2020-feb-mar-may-domain\n",
    "$WG it.unimi.dsi.webgraph.BVGraph -O -L cc-main-2020-feb-mar-may-domain-t"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will now generate the offset files of our webgraph. Offset files are used to determine where each successor list is stored (successor lists are what makes up the graph file, each node is represented by a successor list). More information on the [offset files](http://webgraph.di.unimi.it/docs/it/unimi/dsi/webgraph/BVGraph.html) section."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "$WG it.unimi.dsi.webgraph.algo.ConnectedComponents -m --renumber --sizes -t cc-main-2020-feb-mar-may-domain-t cc-main-2020-feb-mar-may-domain\n",
    "$WG it.unimi.dsi.webgraph.algo.StronglyConnectedComponents --renumber --sizes cc-main-2020-feb-mar-may-domain"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can then generate the connected component files, which will produce `*.wcc`, `*.wccsizes` which contains information about weakly connected components, and similarly the files `*.scc` and `*.sccsizes` contain information about strongly connected components. More info on the [weakly connected components](http://webgraph.di.unimi.it/docs/it/unimi/dsi/webgraph/algo/ConnectedComponents.html) and the [strongly connected components](http://webgraph.di.unimi.it/docs/it/unimi/dsi/webgraph/algo/StronglyConnectedComponents.html) classes, and about the concept [itself](http://pages.di.unipi.it/marino/pythonads/Graphs/StronglyConnectedComponents.html#:~:text=To%20transform%20the%20World%20Wide,connecting%20one%20vertex%20to%20another.&text=One%20graph%20algorithm%20that%20can,connected%20components%20algorithm%20(SCC).)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "\n",
    "$WG it.unimi.dsi.webgraph.Stats --save-degrees cc-main-2020-feb-mar-may-domain"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now finally, we may generate some statistics with the Stats class of WebGraph. The --save-degrees flag generates `*.outdegrees` and `*.indegrees`  files which contain number of degrees per node (so the first line of the outdegrees file contains the number of outdegrees that node of id 0). `*.outdegree` and `*.indegree` files will be generated which contains frequency distribution of degrees (so line 0 of .indegree contains the number of nodes with 0 incoming edges). Also, if connected component files are present, Stats will genenerate some information about the connected components. More information about the [Stats class](http://webgraph.di.unimi.it/docs/it/unimi/dsi/webgraph/Stats.html)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Interactive webgraph"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At this point, you now have the required environment to interact with the web graph itself! Read the seperate [document](interactive_notebook.ipynb) to learn how to get it started."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Statistics visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be using `pandas` and `matplotlib` to manipulate and visualize some interesting statistics about the webgraph!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "graph_name = \"./cc-main-2020-feb-mar-may-domain\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Specify the path to the graph of interest. This may vary, depending on which crawl you would like to investigate. If you followed the instructions exactly, the graph files should be right in the current directory."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#List of extensions that should be plotted\n",
    "metric_extensions = ['indegree','outdegree', 'sccdistr']\n",
    "weakly_cc = \"-connected-components-distrib.txt.gz\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, `metric_extensions` is a list of extensions of the files that we will be plotting. All of these files have the same format, which is a list of frequency values, where the i'th value has the number of nodes with indegree/outdegree of i.\n",
    "\n",
    "`weakly_cc` contains information about strongly and weakly connected components, which is formatted differently (list of tuples containing <frequency,size,percent coverage of the graph>."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Use pandas to manipulate frequency tables\n",
    "def generate_df(metric):\n",
    "    df = pd.read_table(graph_name + \".\" + metric,header=None,names=['frequency'])\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Helper function that uses `pandas` to read in our files as dataframes, for faster processing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Collect and clean frequency tables of indegree, outdegree and sccdistr\n",
    "#Note that 0 must be removed as this will be plotted on a log log graph\n",
    "df_list = []\n",
    "for metric in metric_extensions:\n",
    "    generated_df = generate_df(metric)\n",
    "    generated_df = generated_df.iloc[lambda x: x.index != 0]\n",
    "    generated_df = generated_df[generated_df['frequency'] > 0]\n",
    "    df_list.append([generated_df,metric])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We load up all of our data into seperate dataframes, and remove the values of \"0\" due to the fact that we will be plotting on a loglog graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load up all connected component distribution, which has unique header info so must be done seperately\n",
    "df = pd.read_table(graph_name + weakly_cc, header=None, names=['size','percent_coverage']).sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Remove header\n",
    "df = df.iloc[1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Convert data to int rather than string so it may be plotted\n",
    "df.index = df.index.astype(int)\n",
    "df['size'] = df['size'].astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will load up the connected component file seperately, due to the fact that it is formatted differently."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Add processed connected component distribution table to list of all tables that will be plotted\n",
    "df_list.append([df,'allccdistr'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt_list = [0]*4\n",
    "fig, ((plt_list[0], plt_list[1]), (plt_list[2], plt_list[3])) = plt.subplots(2, 2)\n",
    "fig.suptitle('Common Crawl Web Graph statistics')\n",
    "fig.set_figheight(15)\n",
    "fig.set_figwidth(15)\n",
    "graph_number = 0\n",
    "for df_pair in df_list:\n",
    "    df = df_pair[0]\n",
    "    metric = df_pair[1]\n",
    "    #We will plot this on a log log graph\n",
    "    plt_list[graph_number].semilogy([1,10,100,1000,10000,100000,1000000], color='white')\n",
    "    plt_list[graph_number].semilogx([1,10,100,1000,10000,100000,1000000], color='white')\n",
    "    \n",
    "    #Based on how our data is formatted we will plot on different x and y\n",
    "    if metric == \"sccdistr\":\n",
    "        plt_list[graph_number].scatter(df['frequency'], df.index)\n",
    "    elif metric == \"allccdistr\":\n",
    "        plt_list[graph_number].scatter(df.index, df['size'])\n",
    "    else:\n",
    "        plt_list[graph_number].scatter(df.index, df['frequency'])\n",
    "    \n",
    "    plt_list[graph_number].set(xlabel = \"Number of nodes\", ylabel = metric + \" size\")\n",
    "    graph_number += 1\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, we may visualize these dataframes. Iterating over the list of dataframes that we have created, we will plot them on a loglog graph one by one, and for each different suffix (sccdistr, allcdistr, outdegree, indegree) we may or may not plot them differently. "
   ]
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}