TemplateClusterKEuclid.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "file_extension": ".py",
    "kernelspec": {
      "display_name": "Python 3.7.4 64-bit ('base': conda)",
      "language": "python",
      "name": "python37464bitbasecondaccb10e44a6224ec5a8090d4d6e72e602"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.4"
    },
    "mimetype": "text/x-python",
    "name": "python",
    "npconvert_exporter": "python",
    "pygments_lexer": "ipython3",
    "version": 3,
    "colab": {
      "name": "TemplateClusterKEuclid.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/DevashishX/AbstractClustering/blob/master/TemplateClusterKEuclid.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n6yRKl7Nx8NV",
        "colab_type": "text"
      },
      "source": [
        "#***Set parameters***"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RLla__P2wqVL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "__algo__ = \"KMEANS\" #Name of the algorithm\n",
        "recnum = 3000 #Number of records to be read from files\n",
        "k = 50 #Number of Clusters\n",
        "\n",
        "\n",
        "usesqrt = False #Set value of k to sqrt of recnum\n",
        "randomsample = False #Sampling to be True/False for records which are read"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yFgCtBa6hjJE",
        "colab_type": "text"
      },
      "source": [
        "#Bash Code\n",
        "Git all the Required Files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8xcIo5MshifC",
        "colab_type": "code",
        "outputId": "5b11d6a4-ebb9-48e6-c16c-2c18b0507166",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "!git clone https://github.com/DevashishX/AbstractClustering.git; "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "fatal: destination path 'AbstractClustering' already exists and is not an empty directory.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hJhYEiN2hiV4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp -r AbstractClustering/* ./."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HsYvN3-1hh3V",
        "colab_type": "code",
        "outputId": "65654f47-6837-40e7-b741-24fbec0d1b93",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 660
        }
      },
      "source": [
        "!ls -alFh"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total 84M\n",
            "drwxr-xr-x 1 root root 4.0K Feb 27 13:14 ./\n",
            "drwxr-xr-x 1 root root 4.0K Feb 27 12:02 ../\n",
            "drwxr-xr-x 9 root root 4.0K Feb 27 12:04 AbstractClustering/\n",
            "-rw-r--r-- 1 root root  39M Feb 27 14:52 cleaned.7z\n",
            "-rw-r--r-- 1 root root  90K Feb 27 14:52 ClusterKEuclid.ipynb\n",
            "drwxr-xr-x 1 root root 4.0K Feb 19 17:12 .config/\n",
            "-rw-r--r-- 1 root root 3.7K Feb 27 14:52 convert_repository_json.py\n",
            "drwxr-xr-x 2 root root 4.0K Feb 27 12:04 CORE_files/\n",
            "-rw-r--r-- 1 root root  33K Feb 27 14:52 CORE.html\n",
            "-rw-r--r-- 1 root root  48K Feb 27 14:52 CreateSentenceEmbedding.ipynb\n",
            "drwxr-xr-x 2 root root 4.0K Feb 27 12:04 dataset/\n",
            "-rw-r--r-- 1 root root  95K Feb 27 14:52 Devashish_word2vec_cluster.ipynb\n",
            "-rw-r--r-- 1 root root 978K Feb 27 13:14 downloaded_Mega_Sample_centroid.pkl\n",
            "drwx------ 4 root root 4.0K Feb 27 13:04 drive/\n",
            "-rw-r--r-- 1 root root  27K Feb 27 14:52 example_tags.txt\n",
            "-rw-r--r-- 1 root root  37K Feb 27 14:52 h_cluster.ipynb\n",
            "drwxr-xr-x 2 root root 4.0K Feb 27 12:17 .ipynb_checkpoints/\n",
            "-rw-r--r-- 1 root root 912K Feb 27 14:52 Keras-DEC.ipynb\n",
            "-rw-r--r-- 1 root root 1.1K Feb 27 14:52 LICENSE\n",
            "drwxr-xr-x 2 root root  12K Feb 27 12:04 MegaSentEmbs/\n",
            "drwxr-xr-x 2 root root 4.0K Feb 27 12:04 output/\n",
            "drwxr-xr-x 2 root root  12K Feb 27 12:04 PklCleaned/\n",
            "-rw-r--r-- 1 root root  39M Feb 27 14:53 PklCleaned.7z\n",
            "-rw-r--r-- 1 root root  61K Feb 27 14:53 plot_dbscan.ipynb\n",
            "-rw-r--r-- 1 root root 1.4K Feb 27 14:53 preprocessing_files.py\n",
            "-rw-r--r-- 1 root root  603 Feb 27 14:53 preprocessing.py\n",
            "-rw-r--r-- 1 root root   31 Feb 27 14:53 README.md\n",
            "-rw-r--r-- 1 root root 3.9M Feb 27 12:17 sample_500.pkl\n",
            "drwxr-xr-x 1 root root 4.0K Feb  5 18:37 sample_data/\n",
            "-rw-r--r-- 1 root root  360 Feb 27 14:53 scripts.sh\n",
            "-rw-r--r-- 1 root root 1.1K Feb 27 14:53 sentence_embedding.py\n",
            "-rw-r--r-- 1 root root  25K Feb 27 14:53 splitembs.ipynb\n",
            "-rw-r--r-- 1 root root    0 Feb 27 14:53 test.txt\n",
            "-rw-r--r-- 1 root root  51K Feb 27 14:53 Untitled.ipynb\n",
            "-rw-r--r-- 1 root root  38K Feb 27 14:53 word2vec_cluster.ipynb\n",
            "-rw-r--r-- 1 root root 1.8K Feb 27 14:53 word_embedding.py\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LGy8jsWQx0nS",
        "colab_type": "text"
      },
      "source": [
        "#Actual Code"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "II0jpbbTzD8M",
        "colab_type": "text"
      },
      "source": [
        "#imports and time"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wkS4KGOphHK4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "from sklearn import cluster, datasets\n",
        "import seaborn as sns\n",
        "import os, subprocess\n",
        "import datetime, time"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bFxfXnFh2taT",
        "colab_type": "code",
        "outputId": "d0d94d02-e693-4cde-cd87-2dc51c77b129",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "start_time = time.time()\n",
        "start_datetime = datetime.datetime.now()\n",
        "print(\"Start time: \", start_time)\n",
        "print(\"Start datetime: \", start_datetime)\n",
        "# print(datetime.datetime.now()-start_datetime)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "start_time:  1582829692.078808\n",
            "start_datetime:  2020-02-27 18:54:52.078847\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X2T_-VrUzBVt",
        "colab_type": "text"
      },
      "source": [
        "#File Settings"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fAi1m08zhHLn",
        "colab_type": "code",
        "outputId": "1e03b451-0705-4b84-9e6e-449141acfb97",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "embedDir = \"./MegaSentEmbs/\"\n",
        "dumpdir = \"./dump/\"\n",
        "oldlist = os.listdir(embedDir)\n",
        "filelist = sorted([embedDir+f for f in oldlist if f[-3:]==\"pkl\"])\n",
        "filenum = int(len(filelist))\n",
        "# filenum = 5\n",
        "smalllist = filelist[:filenum]\n",
        "print(\"Length of Smalllist: \", len(smalllist))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Length of Smalllist:  116\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Verw6R-QVK24",
        "colab_type": "text"
      },
      "source": [
        "#Number of Records\n",
        "It is Recommended to Set this at the top parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1KLrsnyYVN0R",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "recnum = recnum"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fN1QgBS0hHL1",
        "colab_type": "text"
      },
      "source": [
        "#Read all the pandas dataframes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "opx4vfnfhHL4",
        "colab_type": "code",
        "outputId": "4de12cfa-fe83-4811-c5aa-b25e965ed72f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "%%time\n",
        "megadf = pd.DataFrame()\n",
        "\n",
        "if randomsample == True:\n",
        "    print(\"randomsample: \", randomsample)\n",
        "    for f in smalllist:\n",
        "        tempdf = pd.read_pickle(f)\n",
        "        megadf = megadf.append(tempdf, ignore_index = True)\n",
        "    megadf = megadf.sample(recnum, random_state=42)\n",
        "else:\n",
        "    print(\"randomsample: \", randomsample)\n",
        "    for f in smalllist:\n",
        "        tempdf = pd.read_pickle(f)\n",
        "        megadf = megadf.append(tempdf, ignore_index = True)\n",
        "        if megadf.shape[0] >= recnum:\n",
        "            megadf = megadf[:recnum]\n",
        "            break\n",
        "\n",
        "print(\"megadf.shape: \", megadf.shape)\n",
        "print(\"megadf.head:\\n\", megadf.head())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "randomsample:  False\n",
            "megadf.shape:  (3000, 4)\n",
            "megadf.head:\n",
            "        id  ...                                          embedding\n",
            "0  395086  ...  [0.1048434150393569, -0.022277361643991286, -0...\n",
            "1  395088  ...  [0.52842, -0.1687455, 0.40093833333333334, -0....\n",
            "2  395097  ...  [0.1048434150393569, -0.022277361643991286, -0...\n",
            "3  395102  ...  [0.07919400708734191, 0.42536576316588975, -0....\n",
            "4  395105  ...  [0.16442311594651932, 0.30571671712325205, -0....\n",
            "\n",
            "[5 rows x 4 columns]\n",
            "CPU times: user 143 ms, sys: 18.9 ms, total: 162 ms\n",
            "Wall time: 165 ms\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AH7_54wfhHMH",
        "colab_type": "code",
        "outputId": "e234ab86-7a01-47b6-cb23-c073e10fbf3b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "predata = megadf[\"embedding\"]\n",
        "data = np.matrix(predata.to_list())\n",
        "print(data.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(3000, 250)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AHw4QXjahHMV",
        "colab_type": "text"
      },
      "source": [
        "#Number of Clusters\n",
        "It is Recommended to Set this at the top parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NDTDoVfhhHMa",
        "colab_type": "code",
        "outputId": "6834227e-c968-4fe6-a288-3e502641167a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "if usesqrt == True:\n",
        "    print(\"usesqrt: \", usesqrt)\n",
        "    sqrt_k = int(np.sqrt(data.shape[0]))\n",
        "    k = int(sqrt_k)\n",
        "else:\n",
        "    print(\"usesqrt: \", usesqrt)\n",
        "    k = k\n",
        "print(\"k: \", k)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "usesqrt:  False\n",
            "k:  50\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KGn8OXrj2CJQ",
        "colab_type": "text"
      },
      "source": [
        "#Clustering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bQYesl8uhHMk",
        "colab_type": "code",
        "outputId": "6b0bad20-a06c-40f3-a8c5-322bed5d81ce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "%%time\n",
        "print(\"Starting Clustering Process\")\n",
        "kmeans = cluster.KMeans(n_clusters=k, verbose=1, n_jobs=-1)\n",
        "kmeans.fit(data)\n",
        "print(\"Done\")\n",
        "print(\"k_means.fit(data) Done!\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Starting Clustering Process\n",
            "done\n",
            "k_means.fit(data) Done!\n",
            "CPU times: user 53.3 ms, sys: 85.5 ms, total: 139 ms\n",
            "Wall time: 7.75 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CArLi48A6ABy",
        "colab_type": "text"
      },
      "source": [
        "###Saving the output data into vars"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QIfHjjLShHM8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "centroids = kmeans.cluster_centers_\n",
        "labels = kmeans.labels_\n",
        "megadf[\"clusterlabel\"]=labels\n",
        "# megadf[\"centroids\"] = centroids\n",
        "# centroidDF = pd.DataFrame(centroids)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "katJKlYX3DVA",
        "colab_type": "text"
      },
      "source": [
        "#Plotting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5RdhXBYhhHNG",
        "colab_type": "code",
        "outputId": "ff87a9c6-03d2-4527-d0b0-710176d17b6a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "plt.figure(figsize=(16,16))\n",
        "snsplot = sns.countplot(\"clusterlabel\", data=megadf)\n",
        "snsplot.xaxis.label.set_size(20)\n",
        "snsplot.yaxis.label.set_size(20)\n",
        "plt.title(\n",
        "        \"{} with k={} records={} features={}\".format(__algo__, k, data.shape[0], data.shape[1]),\n",
        "        fontdict = {'fontsize' : 30}\n",
        "    )"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'KMEANS with k=50 records=3000 features=250')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 143
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7wAAAO5CAYAAAA6jimEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLgm8AAAAdEngBQAAoEsCLwAAAF36/z9fdrDxY20AAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1152x1152 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Auz4Iqgh4Run",
        "colab_type": "text"
      },
      "source": [
        "#Saving Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l0RRWXwvQ9jz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "snspltname = \"{}_K{}_R{}_F{}_plt.png\".format(__algo__, k, data.shape[0], data.shape[1])\n",
        "snsplot.figure.savefig(snspltname)\n",
        "os.environ[\"SAVE_DF\"] = snspltname\n",
        "!cp $SAVE_DF ./drive/My\\ Drive/ColabOutput/."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZLHiM-gclArn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "clusterdfname = \"{}_K{}_R{}_F{}.pkl\".format(__algo__, k, data.shape[0], data.shape[1])\n",
        "megadf.to_pickle(clusterdfname)\n",
        "os.environ[\"SAVE_DF\"] = clusterdfname\n",
        "!cp $SAVE_DF ./drive/My\\ Drive/ColabOutput/."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N8MsneQ_vx4H",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "centroidDF = pd.DataFrame(centroids)\n",
        "centroidDFname = \"{}_K{}_R{}_F{}_Centroids.pkl\".format(__algo__, k, data.shape[0], data.shape[1])\n",
        "centroidDF.to_pickle(centroidDFname)\n",
        "os.environ[\"SAVE_DF\"] = centroidDFname\n",
        "!cp $SAVE_DF ./drive/My\\ Drive/ColabOutput/."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YGmej0Kl5m49",
        "colab_type": "text"
      },
      "source": [
        "###local machine dtale code\n",
        "commented out"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Mp6kiUWhHNR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# # import dtale\n",
        "# s = megadf[megadf[\"clusterlabel\"] == 48].loc[:, [\"title\", \"abstract\"]]\n",
        "# # megadf.columns\n",
        "# # dtale.show(s)\n",
        "# s"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4z2jpCjU5b-1",
        "colab_type": "text"
      },
      "source": [
        "###Sampling the clusters\n",
        "commented out"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MoZTzcR3hHNZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# sample = pd.DataFrame()\n",
        "# for n in range(k):\n",
        "#     temp = megadf[megadf[\"clusterlabel\"] == n].loc[:, :]\n",
        "#     sample = sample.append(temp.head(50), ignore_index=True)\n",
        "#     print(\"Cluster Number: \", n)\n",
        "    # print(temp.head(100).loc[:, [\"id\", \"title\", \"abstract\", \"clusterlabel\"]])\n",
        "    "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x35sOox64Lme",
        "colab_type": "text"
      },
      "source": [
        "#Total Time to Exec"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Fg2D3Fsww94W",
        "colab_type": "code",
        "outputId": "e52fb623-13bd-49a0-e8d0-bb44220cc774",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "end_time = time.time()\n",
        "end_datetime = datetime.datetime.now()\n",
        "# print(\"end_time:\",  end_time)\n",
        "print(\"end_datetime: \", end_datetime)\n",
        "print(\"done! {}\".format(round(end_time-start_time, 2)))\n",
        "print(\"Time taken: \", str(end_datetime-start_datetime))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "end_datetime:  2020-02-27 18:55:10.713255\n",
            "done! 18.63\n",
            "Time taken:  0:00:18.634408\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P0kYelyC5NqY",
        "colab_type": "text"
      },
      "source": [
        "#Open dataframe in colab"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VvsuuCMaQEj4",
        "colab_type": "code",
        "outputId": "5b11f4b6-275c-441c-ebab-130b39dccf5a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "sub = megadf.loc[:, [\"title\", \"abstract\", \"clusterlabel\"]]\n",
        "sub.tail()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>title</th>\n",
              "      <th>abstract</th>\n",
              "      <th>clusterlabel</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2995</th>\n",
              "      <td>Defining the clinical role of the mammalian ta...</td>\n",
              "      <td>[lung, cancer, biggest, cancer, killer, patien...</td>\n",
              "      <td>21</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2996</th>\n",
              "      <td>Bio-functionalised nanoparticles for enzyme se...</td>\n",
              "      <td>[inorganic, nanoparticles, accompanying, diver...</td>\n",
              "      <td>21</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2997</th>\n",
              "      <td>Lithological mapping on the Antarctic Peninsul...</td>\n",
              "      <td>[antarctic, peninsula, larger, uk, limited, ge...</td>\n",
              "      <td>28</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2998</th>\n",
              "      <td>WWOX, tumour suppressor and modifier gene, as ...</td>\n",
              "      <td>[wwox, tumour, suppressor, gene, demonstrated,...</td>\n",
              "      <td>21</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2999</th>\n",
              "      <td>Improving efficiency in drive lines : an exper...</td>\n",
              "      <td>[research, concern, improvement, power, effici...</td>\n",
              "      <td>35</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                  title  ... clusterlabel\n",
              "2995  Defining the clinical role of the mammalian ta...  ...           21\n",
              "2996  Bio-functionalised nanoparticles for enzyme se...  ...           21\n",
              "2997  Lithological mapping on the Antarctic Peninsul...  ...           28\n",
              "2998  WWOX, tumour suppressor and modifier gene, as ...  ...           21\n",
              "2999  Improving efficiency in drive lines : an exper...  ...           35\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 149
        }
      ]
    }
  ]
}