tfidf-benchmark/NLP-pipeline-cuML.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d07a2c86-23aa-4848-b413-2b351614b4b6",
   "metadata": {},
   "source": [
    "# NLP tf-idf pipeline with cuML + Dask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "639f8a3c-70da-4dd3-8e2e-655452dfa306",
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "from dask_cuda import LocalCUDACluster\n",
    "from dask.distributed import Client, wait\n",
    "import dask_cudf\n",
    "from cuml.feature_extraction.text import HashingVectorizer\n",
    "from cuml.dask.feature_extraction.text import TfidfTransformer\n",
    "import nltk\n",
    "import cupy as cp\n",
    "from tqdm import tqdm\n",
    "import dask"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a29e157-5ffc-48f0-af33-591eb404f149",
   "metadata": {},
   "source": [
    "### Import other utility functions for benchmarking purposes "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cc367b02-8ab3-45b1-aa93-cc55bb2ed560",
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils import SimpleTimer, ResultsLogger, scale_workers\n",
    "from utils import visualize_data_cuml as visualize_data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d850d8cf-a343-476e-9c47-a0718e6e04bf",
   "metadata": {},
   "source": [
    "## Setting up the Dask cuda cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a1fb9b87-94af-49f9-b00d-682305b7c802",
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Create a local CUDA cluster\n",
    "cluster = LocalCUDACluster(CUDA_VISIBLE_DEVICES=\"0,1,2,3,4,5,6,7\", local_directory=\"/raid/tfidf_benchmark/dask\", rmm_pool_size=\"28GB\")\n",
    "client = Client(cluster)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a9b987af-61e4-4693-bfc1-5ba6cb96ec24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "            <div>\n",
       "                <div style=\"\n",
       "                    width: 24px;\n",
       "                    height: 24px;\n",
       "                    background-color: #e1e1e1;\n",
       "                    border: 3px solid #9D9D9D;\n",
       "                    border-radius: 5px;\n",
       "                    position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                    <h3 style=\"margin-bottom: 0px;\">Client</h3>\n",
       "                    <p style=\"color: #9D9D9D; margin-bottom: 0px;\">Client-be9d8fc9-fab3-11eb-be1a-d8c4973b6ee2</p>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                    \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\"><strong>Connection method:</strong> Cluster object</td>\n",
       "                    <td style=\"text-align: left;\"><strong>Cluster type:</strong> LocalCUDACluster</td>\n",
       "                </tr>\n",
       "                \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>Dashboard: </strong>\n",
       "                        <a href=\"http://127.0.0.1:8787/status\">http://127.0.0.1:8787/status</a>\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\"></td>\n",
       "                </tr>\n",
       "                \n",
       "                    </table>\n",
       "                    \n",
       "                <details>\n",
       "                <summary style=\"margin-bottom: 20px;\"><h3 style=\"display: inline;\">Cluster Info</h3></summary>\n",
       "                \n",
       "            <div class=\"jp-RenderedHTMLCommon jp-RenderedHTML jp-mod-trusted jp-OutputArea-output\">\n",
       "                <div style=\"\n",
       "                    width: 24px;\n",
       "                    height: 24px;\n",
       "                    background-color: #e1e1e1;\n",
       "                    border: 3px solid #9D9D9D;\n",
       "                    border-radius: 5px;\n",
       "                    position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                    <h3 style=\"margin-bottom: 0px; margin-top: 0px;\">LocalCUDACluster</h3>\n",
       "                    <p style=\"color: #9D9D9D; margin-bottom: 0px;\">2a33d89a</p>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                    \n",
       "            <tr>\n",
       "                <td style=\"text-align: left;\"><strong>Status:</strong> running</td>\n",
       "                <td style=\"text-align: left;\"><strong>Using processes:</strong> True</td>\n",
       "            </tr>\n",
       "        \n",
       "            <tr>\n",
       "                <td style=\"text-align: left;\">\n",
       "                    <strong>Dashboard:</strong> <a href=\"http://127.0.0.1:8787/status\">http://127.0.0.1:8787/status</a>\n",
       "                </td>\n",
       "                <td style=\"text-align: left;\"><strong>Workers:</strong> 8</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                <td style=\"text-align: left;\">\n",
       "                    <strong>Total threads:</strong>\n",
       "                    8\n",
       "                </td>\n",
       "                <td style=\"text-align: left;\">\n",
       "                    <strong>Total memory:</strong>\n",
       "                    0.98 TiB\n",
       "                </td>\n",
       "            </tr>\n",
       "        \n",
       "                    </table>\n",
       "                    <details>\n",
       "                    <summary style=\"margin-bottom: 20px;\"><h3 style=\"display: inline;\">Scheduler Info</h3></summary>\n",
       "                    \n",
       "        <div style=\"\">\n",
       "            \n",
       "            <div>\n",
       "                <div style=\"\n",
       "                    width: 24px;\n",
       "                    height: 24px;\n",
       "                    background-color: #FFF7E5;\n",
       "                    border: 3px solid #FF6132;\n",
       "                    border-radius: 5px;\n",
       "                    position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                    <h3 style=\"margin-bottom: 0px;\">Scheduler</h3>\n",
       "                    <p style=\"color: #9D9D9D; margin-bottom: 0px;\">Scheduler-508ec9ed-cec9-43ef-a1ea-c27da2a6b190</p>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm:</strong> tcp://127.0.0.1:38077</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Workers:</strong> 8</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard:</strong> <a href=\"http://127.0.0.1:8787/status\">http://127.0.0.1:8787/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Total threads:</strong>\n",
       "                                8\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Started:</strong>\n",
       "                                Just now\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Total memory:</strong>\n",
       "                                0.98 TiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                    </table>\n",
       "                </div>\n",
       "            </div>\n",
       "        \n",
       "            <details style=\"margin-left: 48px;\">\n",
       "            <summary style=\"margin-bottom: 20px;\"><h3 style=\"display: inline;\">Workers</h3></summary>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 0</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:38115</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:44863/status\">http://127.0.0.1:44863/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:37435</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-z7eb47og\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 1</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:43831</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:41455/status\">http://127.0.0.1:41455/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:34591</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-gza8qslj\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 2</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:38911</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:39571/status\">http://127.0.0.1:39571/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:42789</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-awu3pgw9\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 3</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:42863</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:40547/status\">http://127.0.0.1:40547/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:36657</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-k6vv1iug\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 4</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:36311</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:46697/status\">http://127.0.0.1:46697/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:33871</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-04sr3n00\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 5</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:38423</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:42397/status\">http://127.0.0.1:42397/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:36607</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-ae483wfu\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 6</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:40897</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:35157/status\">http://127.0.0.1:35157/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:38785</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-t4nxpla5\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            <div style=\"margin-bottom: 20px;\">\n",
       "                <div style=\"width: 24px;\n",
       "                            height: 24px;\n",
       "                            background-color: #DBF5FF;\n",
       "                            border: 3px solid #4CC9FF;\n",
       "                            border-radius: 5px;\n",
       "                            position: absolute;\"> </div>\n",
       "                <div style=\"margin-left: 48px;\">\n",
       "                <details>\n",
       "                    <summary>\n",
       "                        <h4 style=\"margin-bottom: 0px; display: inline;\">Worker: 7</h4>\n",
       "                    </summary>\n",
       "                    <table style=\"width: 100%; text-align: left;\">\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Comm: </strong> tcp://127.0.0.1:39483</td>\n",
       "                            <td style=\"text-align: left;\"><strong>Total threads: </strong> 1</td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Dashboard: </strong>\n",
       "                                <a href=\"http://127.0.0.1:35569/status\">http://127.0.0.1:35569/status</a>\n",
       "                            </td>\n",
       "                            <td style=\"text-align: left;\">\n",
       "                                <strong>Memory: </strong>\n",
       "                                125.97 GiB\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td style=\"text-align: left;\"><strong>Nanny: </strong> tcp://127.0.0.1:36465</td>\n",
       "                            <td style=\"text-align: left;\"></td>\n",
       "                        </tr>\n",
       "                        <tr>\n",
       "                            <td colspan=\"2\" style=\"text-align: left;\">\n",
       "                                <strong>Local directory: </strong>\n",
       "                                /raid/tfidf_benchmark/dask/dask-worker-space/worker-keo4v8xp\n",
       "                            </td>\n",
       "                        </tr>\n",
       "                        \n",
       "                <tr>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU: </strong>Tesla V100-SXM2-32GB\n",
       "                    </td>\n",
       "                    <td style=\"text-align: left;\">\n",
       "                        <strong>GPU memory: </strong>\n",
       "                        31.75 GiB\n",
       "                    </td>\n",
       "                </tr>\n",
       "                \n",
       "                        \n",
       "                    </table>\n",
       "                </details>\n",
       "                </div>\n",
       "            </div>\n",
       "            \n",
       "            </details>\n",
       "        </div>\n",
       "        \n",
       "                    </details>\n",
       "                </div>\n",
       "            </div>\n",
       "        \n",
       "                </details>\n",
       "                \n",
       "                </div>\n",
       "            </div>\n",
       "        "
      ],
      "text/plain": [
       "<Client: 'tcp://127.0.0.1:38077' processes=8 threads=8, memory=0.98 TiB>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "client"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33abd114-7d65-4081-8a04-580e4a5785d8",
   "metadata": {},
   "source": [
    "## Benchmarking Code"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6697d77b-999d-4737-b9ff-f017dbba3868",
   "metadata": {},
   "source": [
    "### Helper Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "90a8f7cb-8837-4ac0-8e21-3ed06665c6f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /home/nfs/anirband/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "nltk.download('stopwords')\n",
    "STOPWORDS = nltk.corpus.stopwords.words('english')\n",
    "PUNCTUATIONS = ['!', '\"', '#', '$', '%', '&', '(', ')', '*', '+', '-', '.',\n",
    "                '/', '\\\\', ':', ';', '<', '=', '>', '?', '@', '[', ']', '^',\n",
    "                '_', '`', '{', '|', '}', '\\t','\\n', \"'\", \",\", '~' , '—']\n",
    "\n",
    "def read_data(client, parquet_path):\n",
    "    data = dask_cudf.read_parquet(parquet_path, columns=[\"review_body\"], split_row_groups=3)\n",
    "    data = client.persist(data)\n",
    "    wait(data)\n",
    "    print(data.shape[0].compute())\n",
    "    return data\n",
    "\n",
    "\n",
    "def text_preprocessor(data, client, column_name, PUNCTUATIONS,STOPWORDS):\n",
    "    data = data[data[column_name].notnull()]\n",
    "    data[column_name] = (data[column_name]\n",
    "                         .str.lower()\n",
    "                         .str.replace_tokens(\n",
    "                             PUNCTUATIONS, [\" \"]*len(PUNCTUATIONS))\n",
    "                         .str.replace_tokens(STOPWORDS, \"\")\n",
    "                         .str.normalize_spaces()\n",
    "                         .str.strip())\n",
    "    data = client.persist(data)\n",
    "    wait(data)\n",
    "    print(data.shape[0].compute())\n",
    "    return data\n",
    "\n",
    "\n",
    "def hashing_vectorizer(data, client, column_name):\n",
    "    vectorizer = HashingVectorizer(stop_words=None, preprocessor=None)\n",
    "    # Meta is an empty dataframe matches the dtypes and columns of the output\n",
    "    meta = dask.array.from_array(cp.sparse.csr_matrix(cp.zeros(1, dtype=cp.float32)))\n",
    "    hashing_vectorized = data[column_name].map_partitions(vectorizer.fit_transform, meta=meta).astype(cp.float32)\n",
    "    hashing_vectorized = client.persist(hashing_vectorized)\n",
    "    wait(hashing_vectorized)\n",
    "    hashing_vectorized.compute_chunk_sizes()\n",
    "    print(hashing_vectorized.shape)\n",
    "    return hashing_vectorized\n",
    "\n",
    "\n",
    "def tfidf_transformer(data, client):\n",
    "    multi_gpu_transformer = TfidfTransformer(client=client)\n",
    "    result = multi_gpu_transformer.fit_transform(data)\n",
    "    result = client.persist(result)\n",
    "    wait(result)\n",
    "    result.compute_chunk_sizes()\n",
    "    print(result.shape)\n",
    "    return result\n",
    "\n",
    "\n",
    "def execute_full_pipeline(n, i, client, parquet_path, worker_counts=[1],\n",
    "                            result_path=\"./results.csv\"):\n",
    "    sample_record = {\"overall\": 0, \"data_read\": 0, \"hashing_vectorizer\": 0,\n",
    "                    \"tfidf_transformer\": 0, \"data_preprocessing\": 0, \"nrows\": 0}\n",
    "    client.restart()\n",
    "    with SimpleTimer() as timer:\n",
    "        data = read_data(client, parquet_path)\n",
    "    sample_record[\"data_read\"] = timer.elapsed/1e9\n",
    "\n",
    "    with SimpleTimer() as timer:\n",
    "        data = text_preprocessor(data, client, \"review_body\", PUNCTUATIONS, STOPWORDS)\n",
    "    sample_record[\"data_preprocessing\"] = timer.elapsed/1e9\n",
    "\n",
    "    with SimpleTimer() as timer:\n",
    "        hashing_vectorized = hashing_vectorizer(data, client, \"review_body\")\n",
    "    sample_record[\"hashing_vectorizer\"] = timer.elapsed/1e9\n",
    "\n",
    "    with SimpleTimer() as timer:\n",
    "        result = tfidf_transformer(hashing_vectorized, client)\n",
    "    sample_record[\"tfidf_transformer\"] = timer.elapsed/1e9\n",
    "\n",
    "    print(f\"Workers:{n}, Sample Run:{i}, Finished loading data in {sample_record['data_read']}s\")\n",
    "    print(f\"Workers:{n}, Sample Run:{i}, Finished preprocessing data in {sample_record['data_preprocessing']}s\")\n",
    "    print(f\"Workers:{n}, Sample Run:{i}, Finished fitting HashVectorizer in {sample_record['hashing_vectorizer']}s\")\n",
    "    print(f\"Workers:{n}, Sample Run:{i}, Finished fitting IDF Transformer in {sample_record['tfidf_transformer']}s\")\n",
    "\n",
    "    return data, result, sample_record"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "897b3081-f947-4f87-8443-315c53053612",
   "metadata": {},
   "outputs": [],
   "source": [
    "def performance_numbers(client, parquet_path, worker_counts=[1], samples=1, result_path=\"./results.csv\"):\n",
    "    \"\"\"\n",
    "    Main function to perform the performance sweep\n",
    "    \"\"\"\n",
    "    results_logger = ResultsLogger(result_path)\n",
    "    for n in worker_counts: \n",
    "        scale_workers(client, n)\n",
    "        \n",
    "        for i in tqdm(range(samples)): \n",
    "            with SimpleTimer() as overalltimer:\n",
    "                data, result, sample_record = execute_full_pipeline(n, i, client, parquet_path, worker_counts=[1], result_path=result_path)\n",
    "            sample_record[\"overall\"]=overalltimer.elapsed/1e9\n",
    "            sample_record[\"nrows\"]=data.shape[0].compute()\n",
    "            sample_record[\"n_workers\"]=n\n",
    "            sample_record[\"sample_run\"]=i\n",
    "            print(f\"Workers:{n}, Sample Run:{i}, Finished executing full pipeline in {overalltimer.elapsed/1e9}s\")\n",
    "            results_logger.log(sample_record)\n",
    "    results_logger.write()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "9fa939d6-3158-4577-b7ef-6fb4f8bec54b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#parquet_path = 's3://amazon-reviews-pds/parquet/product_category=Camera/*.parquet'\n",
    "dataset = \"Books\"\n",
    "#parquet_path = f'./data/product_category={dataset}'\n",
    "parquet_path = f\"/raid/amazon_reviews_dataset/product_category={dataset}\"\n",
    "samples = 5\n",
    "worker_counts = [2,4,6,8]\n",
    "result_path = f\"./results/result_poc_nlp_dask_{dataset}_persist.csv\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "4da2c546-c633-4f77-9220-e0067add959c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                                                                                                                                                      | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:53:16.146309] [Delayed('_merge_stats_to_model-813893b2-626b-4561-b9a4-726917b940f7')]\n",
      "[I] [07:53:16.150658] [Delayed('_merge_stats_to_model-e56a729c-dc17-4ab9-8c9b-ec46ecca3cc6')]\n",
      "[I] [07:53:16.210694] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-813893b2-626b-4561-b9a4-726917b940f7>]\n",
      "[I] [07:53:16.210928] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-e56a729c-dc17-4ab9-8c9b-ec46ecca3cc6>]\n",
      "[I] [07:53:16.216555] [Delayed('_merge_stats_to_model-06d2bc0c-545b-4c04-bcc5-144fc65e305a')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██████████████████████████████████▊                                                                                                                                           | 1/5 [00:44<02:57, 44.41s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:2, Sample Run:0, Finished loading data in 5.0895142s\n",
      "Workers:2, Sample Run:0, Finished preprocessing data in 18.540178819s\n",
      "Workers:2, Sample Run:0, Finished fitting HashVectorizer in 16.29362533s\n",
      "Workers:2, Sample Run:0, Finished fitting IDF Transformer in 1.15394484s\n",
      "Workers:2, Sample Run:0, Finished executing full pipeline in 44.365918109s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:53:59.010989] [Delayed('_merge_stats_to_model-8e1d3a17-5770-463d-9084-e8453939046b')]\n",
      "[I] [07:53:59.015094] [Delayed('_merge_stats_to_model-39ddeb09-cfc6-4e82-80ef-132d5d18fab6')]\n",
      "[I] [07:53:59.078411] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-8e1d3a17-5770-463d-9084-e8453939046b>]\n",
      "[I] [07:53:59.078614] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-39ddeb09-cfc6-4e82-80ef-132d5d18fab6>]\n",
      "[I] [07:53:59.085929] [Delayed('_merge_stats_to_model-674d1b17-e3db-4271-8d48-3ea8e6d653db')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 40%|█████████████████████████████████████████████████████████████████████▌                                                                                                        | 2/5 [01:27<02:10, 43.50s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:2, Sample Run:1, Finished loading data in 3.18178013s\n",
      "Workers:2, Sample Run:1, Finished preprocessing data in 18.522294681s\n",
      "Workers:2, Sample Run:1, Finished fitting HashVectorizer in 16.099082448s\n",
      "Workers:2, Sample Run:1, Finished fitting IDF Transformer in 1.194648918s\n",
      "Workers:2, Sample Run:1, Finished executing full pipeline in 42.820689024s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:54:42.644564] [Delayed('_merge_stats_to_model-db17d6f7-5045-4a84-8d2e-2c4cd7cb8999')]\n",
      "[I] [07:54:42.649168] [Delayed('_merge_stats_to_model-84ae04d3-fe17-4291-a274-c95e661fa74f')]\n",
      "[I] [07:54:42.716200] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-db17d6f7-5045-4a84-8d2e-2c4cd7cb8999>]\n",
      "[I] [07:54:42.716478] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-84ae04d3-fe17-4291-a274-c95e661fa74f>]\n",
      "[I] [07:54:42.723606] [Delayed('_merge_stats_to_model-2a09a71a-6203-49d6-8949-6f338607be7d')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 60%|████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                                                     | 3/5 [02:10<01:27, 43.58s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:2, Sample Run:2, Finished loading data in 3.57763331s\n",
      "Workers:2, Sample Run:2, Finished preprocessing data in 18.595843886s\n",
      "Workers:2, Sample Run:2, Finished fitting HashVectorizer in 16.244207544s\n",
      "Workers:2, Sample Run:2, Finished fitting IDF Transformer in 1.166985276s\n",
      "Workers:2, Sample Run:2, Finished executing full pipeline in 43.618034356s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:55:27.027576] [Delayed('_merge_stats_to_model-ee57678c-e36b-4880-878a-4bd09ebf0f2b')]\n",
      "[I] [07:55:27.032369] [Delayed('_merge_stats_to_model-ace82e14-8ca8-4cac-aa74-576cb688e7e2')]\n",
      "[I] [07:55:27.099419] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ee57678c-e36b-4880-878a-4bd09ebf0f2b>]\n",
      "[I] [07:55:27.099681] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ace82e14-8ca8-4cac-aa74-576cb688e7e2>]\n",
      "[I] [07:55:27.105389] [Delayed('_merge_stats_to_model-96f4cf4d-d035-4d04-8e34-e7ce57ea40e2')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 80%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████▏                                  | 4/5 [02:55<00:43, 43.88s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:2, Sample Run:3, Finished loading data in 3.287220726s\n",
      "Workers:2, Sample Run:3, Finished preprocessing data in 18.502948631s\n",
      "Workers:2, Sample Run:3, Finished fitting HashVectorizer in 16.263773818s\n",
      "Workers:2, Sample Run:3, Finished fitting IDF Transformer in 1.217355487s\n",
      "Workers:2, Sample Run:3, Finished executing full pipeline in 44.294787975s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:56:09.983289] [Delayed('_merge_stats_to_model-2aec9ab0-529b-4ba9-b988-34a35633c758')]\n",
      "[I] [07:56:09.988659] [Delayed('_merge_stats_to_model-56da8331-6d4a-4b88-a58d-f6b104e29c77')]\n",
      "[I] [07:56:10.056251] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-2aec9ab0-529b-4ba9-b988-34a35633c758>]\n",
      "[I] [07:56:10.056324] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-56da8331-6d4a-4b88-a58d-f6b104e29c77>]\n",
      "[I] [07:56:10.062236] [Delayed('_merge_stats_to_model-31aff4f1-39a1-40e1-a258-52f3b8578048')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [03:38<00:00, 43.65s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:2, Sample Run:4, Finished loading data in 3.247898192s\n",
      "Workers:2, Sample Run:4, Finished preprocessing data in 18.469599377s\n",
      "Workers:2, Sample Run:4, Finished fitting HashVectorizer in 15.872357573s\n",
      "Workers:2, Sample Run:4, Finished fitting IDF Transformer in 1.178282708s\n",
      "Workers:2, Sample Run:4, Finished executing full pipeline in 42.909815838s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "  0%|                                                                                                                                                                                      | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:56:42.947363] [Delayed('_merge_stats_to_model-96cbe44c-7000-4e5f-bba7-14c088da2de9')]\n",
      "[I] [07:56:42.950105] [Delayed('_merge_stats_to_model-1db309f7-dabb-4ae7-8d7e-006e5f835588')]\n",
      "[I] [07:56:42.952554] [Delayed('_merge_stats_to_model-4720589a-3abc-4907-82d5-746ca4419a25')]\n",
      "[I] [07:56:42.954965] [Delayed('_merge_stats_to_model-b5632d84-5f8d-4004-8c4e-b2a1eb816544')]\n",
      "[I] [07:56:43.007107] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-96cbe44c-7000-4e5f-bba7-14c088da2de9>]\n",
      "[I] [07:56:43.007344] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-1db309f7-dabb-4ae7-8d7e-006e5f835588>]\n",
      "[I] [07:56:43.007483] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-4720589a-3abc-4907-82d5-746ca4419a25>]\n",
      "[I] [07:56:43.007664] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-b5632d84-5f8d-4004-8c4e-b2a1eb816544>]\n",
      "[I] [07:56:43.014283] [Delayed('_merge_stats_to_model-77365840-5237-4ff3-8426-af2e3bc04ee9')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██████████████████████████████████▊                                                                                                                                           | 1/5 [00:29<01:57, 29.50s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:4, Sample Run:0, Finished loading data in 2.641331591s\n",
      "Workers:4, Sample Run:0, Finished preprocessing data in 10.322448244s\n",
      "Workers:4, Sample Run:0, Finished fitting HashVectorizer in 9.648911977s\n",
      "Workers:4, Sample Run:0, Finished fitting IDF Transformer in 2.511276988s\n",
      "Workers:4, Sample Run:0, Finished executing full pipeline in 29.454382971s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:57:15.518722] [Delayed('_merge_stats_to_model-d174b47b-0418-4c9c-a8c0-9bbd85cd027c')]\n",
      "[I] [07:57:15.521291] [Delayed('_merge_stats_to_model-6c34e345-b358-4cb5-a5fc-c24f977294c6')]\n",
      "[I] [07:57:15.523830] [Delayed('_merge_stats_to_model-82f117e9-5d3b-4231-9ad5-b4eb69e6c7ff')]\n",
      "[I] [07:57:15.526343] [Delayed('_merge_stats_to_model-72b882f1-f9ac-4f2d-a144-a63959eb2955')]\n",
      "[I] [07:57:15.579719] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-d174b47b-0418-4c9c-a8c0-9bbd85cd027c>]\n",
      "[I] [07:57:15.579991] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-6c34e345-b358-4cb5-a5fc-c24f977294c6>]\n",
      "[I] [07:57:15.580128] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-82f117e9-5d3b-4231-9ad5-b4eb69e6c7ff>]\n",
      "[I] [07:57:15.580255] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-72b882f1-f9ac-4f2d-a144-a63959eb2955>]\n",
      "[I] [07:57:15.587570] [Delayed('_merge_stats_to_model-a9d1e77f-6045-4a0c-bd82-fd4afce45940')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 40%|█████████████████████████████████████████████████████████████████████▌                                                                                                        | 2/5 [01:02<01:33, 31.31s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:4, Sample Run:1, Finished loading data in 2.597419993s\n",
      "Workers:4, Sample Run:1, Finished preprocessing data in 11.947546368s\n",
      "Workers:4, Sample Run:1, Finished fitting HashVectorizer in 10.239333623s\n",
      "Workers:4, Sample Run:1, Finished fitting IDF Transformer in 2.550335588s\n",
      "Workers:4, Sample Run:1, Finished executing full pipeline in 32.528956023s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:57:45.345187] [Delayed('_merge_stats_to_model-44ef49bc-c599-4cc1-b25b-f64e06b7dd08')]\n",
      "[I] [07:57:45.348170] [Delayed('_merge_stats_to_model-273f1fb1-f669-4334-9d2d-c4f928f0f5ca')]\n",
      "[I] [07:57:45.350588] [Delayed('_merge_stats_to_model-ac92f192-a15b-4557-b457-38c19f9b70b9')]\n",
      "[I] [07:57:45.353405] [Delayed('_merge_stats_to_model-4f2c82a0-57cb-4514-a6a9-dcdf41b4ee38')]\n",
      "[I] [07:57:45.406716] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-44ef49bc-c599-4cc1-b25b-f64e06b7dd08>]\n",
      "[I] [07:57:45.406904] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-273f1fb1-f669-4334-9d2d-c4f928f0f5ca>]\n",
      "[I] [07:57:45.407040] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ac92f192-a15b-4557-b457-38c19f9b70b9>]\n",
      "[I] [07:57:45.407171] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-4f2c82a0-57cb-4514-a6a9-dcdf41b4ee38>]\n",
      "[I] [07:57:45.413968] [Delayed('_merge_stats_to_model-f99d51fd-b167-499c-a615-240d7c00e592')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 60%|████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                                                     | 3/5 [01:34<01:03, 31.71s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:4, Sample Run:2, Finished loading data in 2.394929625s\n",
      "Workers:4, Sample Run:2, Finished preprocessing data in 10.916926605s\n",
      "Workers:4, Sample Run:2, Finished fitting HashVectorizer in 9.417120038s\n",
      "Workers:4, Sample Run:2, Finished fitting IDF Transformer in 4.860107799s\n",
      "Workers:4, Sample Run:2, Finished executing full pipeline in 32.149036668s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:58:17.295292] [Delayed('_merge_stats_to_model-3fb9dc77-bc1d-4a07-96b2-81eb60b14ac3')]\n",
      "[I] [07:58:17.298308] [Delayed('_merge_stats_to_model-40130cb2-fa4b-46e6-85e6-407e7e26d910')]\n",
      "[I] [07:58:17.300615] [Delayed('_merge_stats_to_model-e412e0b6-b0ef-4ae6-a05b-c92139adce04')]\n",
      "[I] [07:58:17.302832] [Delayed('_merge_stats_to_model-16c3220e-662f-461d-8cb2-33971f49adf2')]\n",
      "[I] [07:58:17.354257] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3fb9dc77-bc1d-4a07-96b2-81eb60b14ac3>]\n",
      "[I] [07:58:17.354497] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-40130cb2-fa4b-46e6-85e6-407e7e26d910>]\n",
      "[I] [07:58:17.354641] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-e412e0b6-b0ef-4ae6-a05b-c92139adce04>]\n",
      "[I] [07:58:17.354772] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-16c3220e-662f-461d-8cb2-33971f49adf2>]\n",
      "[I] [07:58:17.361411] [Delayed('_merge_stats_to_model-5739c26c-ceb4-4c9d-a66d-8b84db3a4b95')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 80%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████▏                                  | 4/5 [02:03<00:30, 30.87s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:4, Sample Run:3, Finished loading data in 2.440834652s\n",
      "Workers:4, Sample Run:3, Finished preprocessing data in 10.32372482s\n",
      "Workers:4, Sample Run:3, Finished fitting HashVectorizer in 9.590961617s\n",
      "Workers:4, Sample Run:3, Finished fitting IDF Transformer in 2.505987025s\n",
      "Workers:4, Sample Run:3, Finished executing full pipeline in 29.542107875s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:58:47.258046] [Delayed('_merge_stats_to_model-f6a3a406-5f0e-4304-b68b-c4a08d849127')]\n",
      "[I] [07:58:47.260448] [Delayed('_merge_stats_to_model-f849a19a-07df-4fdb-ad20-9c0ffe775b74')]\n",
      "[I] [07:58:47.262855] [Delayed('_merge_stats_to_model-681f11a4-e7e5-467b-86bb-0d15e6bd3322')]\n",
      "[I] [07:58:47.265087] [Delayed('_merge_stats_to_model-ea85497f-30bb-4159-a47c-dd9188bedd45')]\n",
      "[I] [07:58:47.315972] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-f6a3a406-5f0e-4304-b68b-c4a08d849127>]\n",
      "[I] [07:58:47.316130] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-f849a19a-07df-4fdb-ad20-9c0ffe775b74>]\n",
      "[I] [07:58:47.316358] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-681f11a4-e7e5-467b-86bb-0d15e6bd3322>]\n",
      "[I] [07:58:47.316496] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ea85497f-30bb-4159-a47c-dd9188bedd45>]\n",
      "[I] [07:58:47.323139] [Delayed('_merge_stats_to_model-68259de5-53dd-4fc9-b9ba-178bc3274cd1')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [02:35<00:00, 31.09s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:4, Sample Run:4, Finished loading data in 2.423161772s\n",
      "Workers:4, Sample Run:4, Finished preprocessing data in 10.310600325s\n",
      "Workers:4, Sample Run:4, Finished fitting HashVectorizer in 9.241648226s\n",
      "Workers:4, Sample Run:4, Finished fitting IDF Transformer in 4.612093404s\n",
      "Workers:4, Sample Run:4, Finished executing full pipeline in 31.554940618s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "  0%|                                                                                                                                                                                      | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:59:18.920150] [Delayed('_merge_stats_to_model-dd4b8909-da16-4a71-b3c2-c42b727d2d17')]\n",
      "[I] [07:59:18.922450] [Delayed('_merge_stats_to_model-a871eaa3-b7ee-45ec-8876-4bf63748d2bc')]\n",
      "[I] [07:59:18.924634] [Delayed('_merge_stats_to_model-865788ea-9b0c-495f-9d78-888e16223e0b')]\n",
      "[I] [07:59:18.926857] [Delayed('_merge_stats_to_model-cbaa0cbc-f0d2-4a30-b525-40bcc25ea5cc')]\n",
      "[I] [07:59:18.929482] [Delayed('_merge_stats_to_model-1da0f8bf-ea35-4dc3-9b3c-7c0a3af604c8')]\n",
      "[I] [07:59:18.931628] [Delayed('_merge_stats_to_model-7ecf9e8e-6053-453c-bcea-f8cc0e9bb2e7')]\n",
      "[I] [07:59:18.985757] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-dd4b8909-da16-4a71-b3c2-c42b727d2d17>]\n",
      "[I] [07:59:18.986018] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-a871eaa3-b7ee-45ec-8876-4bf63748d2bc>]\n",
      "[I] [07:59:18.986155] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-865788ea-9b0c-495f-9d78-888e16223e0b>]\n",
      "[I] [07:59:18.986292] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-cbaa0cbc-f0d2-4a30-b525-40bcc25ea5cc>]\n",
      "[I] [07:59:18.986406] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-1da0f8bf-ea35-4dc3-9b3c-7c0a3af604c8>]\n",
      "[I] [07:59:18.986519] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-7ecf9e8e-6053-453c-bcea-f8cc0e9bb2e7>]\n",
      "[I] [07:59:18.993834] [Delayed('_merge_stats_to_model-3b3356eb-f4fb-439d-87a1-40b610835557')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██████████████████████████████████▊                                                                                                                                           | 1/5 [00:25<01:43, 26.00s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:6, Sample Run:0, Finished loading data in 2.144618946s\n",
      "Workers:6, Sample Run:0, Finished preprocessing data in 7.733729737s\n",
      "Workers:6, Sample Run:0, Finished fitting HashVectorizer in 7.513520817s\n",
      "Workers:6, Sample Run:0, Finished fitting IDF Transformer in 3.511341176s\n",
      "Workers:6, Sample Run:0, Finished executing full pipeline in 25.95266292s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [07:59:44.608318] [Delayed('_merge_stats_to_model-3117fff8-cf8b-4f3e-9155-fa2b0d77ace3')]\n",
      "[I] [07:59:44.610392] [Delayed('_merge_stats_to_model-c61e9cc4-dcaa-495f-972e-6167405a5dde')]\n",
      "[I] [07:59:44.614565] [Delayed('_merge_stats_to_model-b8532553-bbef-4028-8852-ded59585cc8d')]\n",
      "[I] [07:59:44.616651] [Delayed('_merge_stats_to_model-88fcea21-2102-4525-ac9b-41fc4e022688')]\n",
      "[I] [07:59:44.619014] [Delayed('_merge_stats_to_model-4d94604e-0f65-4960-ae4e-15d397f181d2')]\n",
      "[I] [07:59:44.621048] [Delayed('_merge_stats_to_model-ef4fd664-0f56-4c27-9d7d-31f36f9a8861')]\n",
      "[I] [07:59:44.677896] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3117fff8-cf8b-4f3e-9155-fa2b0d77ace3>]\n",
      "[I] [07:59:44.678153] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-c61e9cc4-dcaa-495f-972e-6167405a5dde>]\n",
      "[I] [07:59:44.678292] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-b8532553-bbef-4028-8852-ded59585cc8d>]\n",
      "[I] [07:59:44.678441] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-88fcea21-2102-4525-ac9b-41fc4e022688>]\n",
      "[I] [07:59:44.678567] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-4d94604e-0f65-4960-ae4e-15d397f181d2>]\n",
      "[I] [07:59:44.678693] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ef4fd664-0f56-4c27-9d7d-31f36f9a8861>]\n",
      "[I] [07:59:44.687351] [Delayed('_merge_stats_to_model-a1cfa8f7-cbb0-4b31-ba6a-35f029a98d04')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 40%|█████████████████████████████████████████████████████████████████████▌                                                                                                        | 2/5 [00:50<01:16, 25.34s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:6, Sample Run:1, Finished loading data in 2.325429719s\n",
      "Workers:6, Sample Run:1, Finished preprocessing data in 7.538877269s\n",
      "Workers:6, Sample Run:1, Finished fitting HashVectorizer in 7.108562342s\n",
      "Workers:6, Sample Run:1, Finished fitting IDF Transformer in 2.65666405s\n",
      "Workers:6, Sample Run:1, Finished executing full pipeline in 24.842991045s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:00:09.000642] [Delayed('_merge_stats_to_model-58d88b2a-d35e-4205-9701-9749966b07bf')]\n",
      "[I] [08:00:09.003341] [Delayed('_merge_stats_to_model-28586453-4642-4762-946f-97694c32f718')]\n",
      "[I] [08:00:09.005708] [Delayed('_merge_stats_to_model-87f27f6c-8c27-4a87-ab77-bae7c463ae81')]\n",
      "[I] [08:00:09.007929] [Delayed('_merge_stats_to_model-06506dc9-db2a-4cd0-a9aa-54782210cf45')]\n",
      "[I] [08:00:09.010179] [Delayed('_merge_stats_to_model-9740d5fd-cc3f-4221-ae2e-d0ec614769f8')]\n",
      "[I] [08:00:09.011470] [Delayed('_merge_stats_to_model-3345641d-d0ca-44d9-bff5-caebe478d128')]\n",
      "[I] [08:00:09.067347] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-58d88b2a-d35e-4205-9701-9749966b07bf>]\n",
      "[I] [08:00:09.067598] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-28586453-4642-4762-946f-97694c32f718>]\n",
      "[I] [08:00:09.067808] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-87f27f6c-8c27-4a87-ab77-bae7c463ae81>]\n",
      "[I] [08:00:09.067956] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-06506dc9-db2a-4cd0-a9aa-54782210cf45>]\n",
      "[I] [08:00:09.068076] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-9740d5fd-cc3f-4221-ae2e-d0ec614769f8>]\n",
      "[I] [08:00:09.068198] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3345641d-d0ca-44d9-bff5-caebe478d128>]\n",
      "[I] [08:00:09.075887] [Delayed('_merge_stats_to_model-338d0bd8-0668-4418-8e0c-49e7ef50d61f')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 60%|████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                                                     | 3/5 [01:14<00:48, 24.42s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:6, Sample Run:2, Finished loading data in 2.151522119s\n",
      "Workers:6, Sample Run:2, Finished preprocessing data in 7.164732339s\n",
      "Workers:6, Sample Run:2, Finished fitting HashVectorizer in 6.911168291s\n",
      "Workers:6, Sample Run:2, Finished fitting IDF Transformer in 1.805155971s\n",
      "Workers:6, Sample Run:2, Finished executing full pipeline in 23.274068614s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:00:32.697567] [Delayed('_merge_stats_to_model-341b4659-d97b-43af-8840-7b894575abe6')]\n",
      "[I] [08:00:32.699991] [Delayed('_merge_stats_to_model-14713259-7f5c-4b2c-9b58-8210b510502c')]\n",
      "[I] [08:00:32.702234] [Delayed('_merge_stats_to_model-721a06f0-8092-4038-89fc-1c889fcf7881')]\n",
      "[I] [08:00:32.705799] [Delayed('_merge_stats_to_model-309c39b6-0e8d-489f-bcc6-30716d0b3a31')]\n",
      "[I] [08:00:32.708448] [Delayed('_merge_stats_to_model-4cd630bd-87c3-45b9-bac8-1373d5781731')]\n",
      "[I] [08:00:32.710758] [Delayed('_merge_stats_to_model-c1bffb05-4cf4-4da9-bfe4-93964667dff3')]\n",
      "[I] [08:00:32.766644] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-341b4659-d97b-43af-8840-7b894575abe6>]\n",
      "[I] [08:00:32.766930] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-14713259-7f5c-4b2c-9b58-8210b510502c>]\n",
      "[I] [08:00:32.767074] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-721a06f0-8092-4038-89fc-1c889fcf7881>]\n",
      "[I] [08:00:32.767218] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-309c39b6-0e8d-489f-bcc6-30716d0b3a31>]\n",
      "[I] [08:00:32.767343] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-4cd630bd-87c3-45b9-bac8-1373d5781731>]\n",
      "[I] [08:00:32.767468] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-c1bffb05-4cf4-4da9-bfe4-93964667dff3>]\n",
      "[I] [08:00:32.775170] [Delayed('_merge_stats_to_model-efe49004-d73f-4561-adb8-5b055ef6b979')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 80%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████▏                                  | 4/5 [01:39<00:24, 24.67s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:6, Sample Run:3, Finished loading data in 2.244197676s\n",
      "Workers:6, Sample Run:3, Finished preprocessing data in 7.310196311s\n",
      "Workers:6, Sample Run:3, Finished fitting HashVectorizer in 6.985161746s\n",
      "Workers:6, Sample Run:3, Finished fitting IDF Transformer in 2.968798225s\n",
      "Workers:6, Sample Run:3, Finished executing full pipeline in 25.013885772s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:00:58.554466] [Delayed('_merge_stats_to_model-ea4587f3-dea6-467d-b745-ffc151dfee1c')]\n",
      "[I] [08:00:58.557117] [Delayed('_merge_stats_to_model-f927271d-60d3-494b-8f82-6880f7ddcab4')]\n",
      "[I] [08:00:58.559146] [Delayed('_merge_stats_to_model-02ea80f6-e27a-4048-821b-f550d5870005')]\n",
      "[I] [08:00:58.560508] [Delayed('_merge_stats_to_model-5813a05b-fdd1-4162-8733-152eeb6eb82d')]\n",
      "[I] [08:00:58.562638] [Delayed('_merge_stats_to_model-b78440f4-f3be-4b4b-8911-3e5696d60c5f')]\n",
      "[I] [08:00:58.564866] [Delayed('_merge_stats_to_model-be99921b-0610-4149-9238-b5f729785eb4')]\n",
      "[I] [08:00:58.617827] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ea4587f3-dea6-467d-b745-ffc151dfee1c>]\n",
      "[I] [08:00:58.618061] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-f927271d-60d3-494b-8f82-6880f7ddcab4>]\n",
      "[I] [08:00:58.618196] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-02ea80f6-e27a-4048-821b-f550d5870005>]\n",
      "[I] [08:00:58.618318] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-5813a05b-fdd1-4162-8733-152eeb6eb82d>]\n",
      "[I] [08:00:58.618440] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-b78440f4-f3be-4b4b-8911-3e5696d60c5f>]\n",
      "[I] [08:00:58.618556] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-be99921b-0610-4149-9238-b5f729785eb4>]\n",
      "[I] [08:00:58.625397] [Delayed('_merge_stats_to_model-5bb75c96-bf95-4b0c-83a6-d9cafe6d8cc7')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [02:05<00:00, 25.12s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:6, Sample Run:4, Finished loading data in 2.088568303s\n",
      "Workers:6, Sample Run:4, Finished preprocessing data in 7.58727199s\n",
      "Workers:6, Sample Run:4, Finished fitting HashVectorizer in 7.462833005s\n",
      "Workers:6, Sample Run:4, Finished fitting IDF Transformer in 3.419484624s\n",
      "Workers:6, Sample Run:4, Finished executing full pipeline in 26.259453617s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "  0%|                                                                                                                                                                                      | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:01:26.561953] [Delayed('_merge_stats_to_model-3f8625a0-4881-45d5-82eb-2278315188c9')]\n",
      "[I] [08:01:26.563389] [Delayed('_merge_stats_to_model-2ad6b606-0c42-4607-8cd3-a8f9662ed876')]\n",
      "[I] [08:01:26.564715] [Delayed('_merge_stats_to_model-cff71b5c-10d1-44ba-900e-e1faa337c461')]\n",
      "[I] [08:01:26.566148] [Delayed('_merge_stats_to_model-3a82cd52-65c7-4605-9f2a-b18fed6c1456')]\n",
      "[I] [08:01:26.567456] [Delayed('_merge_stats_to_model-bc66e1b2-5a66-4ac9-963b-771daed765c4')]\n",
      "[I] [08:01:26.568907] [Delayed('_merge_stats_to_model-a55996b6-28a5-440e-b01a-d3e6aa91d9da')]\n",
      "[I] [08:01:26.570294] [Delayed('_merge_stats_to_model-3fc3125f-b2d0-4203-a493-fadbea0e7cbf')]\n",
      "[I] [08:01:26.571830] [Delayed('_merge_stats_to_model-2745af1e-fbc0-47ad-86ea-a7b97638bbf7')]\n",
      "[I] [08:01:26.619331] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3f8625a0-4881-45d5-82eb-2278315188c9>]\n",
      "[I] [08:01:26.619585] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-2ad6b606-0c42-4607-8cd3-a8f9662ed876>]\n",
      "[I] [08:01:26.619751] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-cff71b5c-10d1-44ba-900e-e1faa337c461>]\n",
      "[I] [08:01:26.619934] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3a82cd52-65c7-4605-9f2a-b18fed6c1456>]\n",
      "[I] [08:01:26.620074] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-bc66e1b2-5a66-4ac9-963b-771daed765c4>]\n",
      "[I] [08:01:26.620197] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-a55996b6-28a5-440e-b01a-d3e6aa91d9da>]\n",
      "[I] [08:01:26.620469] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3fc3125f-b2d0-4203-a493-fadbea0e7cbf>]\n",
      "[I] [08:01:26.620597] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-2745af1e-fbc0-47ad-86ea-a7b97638bbf7>]\n",
      "[I] [08:01:26.629997] [Delayed('_merge_stats_to_model-7db2dd98-3315-4837-a179-beccb3ad34fd')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 20%|██████████████████████████████████▊                                                                                                                                           | 1/5 [00:21<01:25, 21.40s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:8, Sample Run:0, Finished loading data in 2.106481397s\n",
      "Workers:8, Sample Run:0, Finished preprocessing data in 5.670550447s\n",
      "Workers:8, Sample Run:0, Finished fitting HashVectorizer in 6.212106907s\n",
      "Workers:8, Sample Run:0, Finished fitting IDF Transformer in 1.93128143s\n",
      "Workers:8, Sample Run:0, Finished executing full pipeline in 21.351185165s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:01:48.766231] [Delayed('_merge_stats_to_model-b4d03902-891d-4af6-a5dc-011eab654c02')]\n",
      "[I] [08:01:48.767709] [Delayed('_merge_stats_to_model-33b18500-16da-45b9-af81-bfc4c411dee2')]\n",
      "[I] [08:01:48.769085] [Delayed('_merge_stats_to_model-725aa57e-7ed7-4dd8-b698-75ad32ef1956')]\n",
      "[I] [08:01:48.770511] [Delayed('_merge_stats_to_model-96d2b721-b2e0-4b32-9eb5-ff5424c0dbe3')]\n",
      "[I] [08:01:48.771744] [Delayed('_merge_stats_to_model-06a3e5ab-9716-4f4b-82dd-dfe7760f1ce6')]\n",
      "[I] [08:01:48.772997] [Delayed('_merge_stats_to_model-ad787bc3-5a0d-4cec-9792-0ab47a786ff7')]\n",
      "[I] [08:01:48.774226] [Delayed('_merge_stats_to_model-57e8a6ba-bc5b-4fc9-a1c0-8b6051c11cfe')]\n",
      "[I] [08:01:48.775920] [Delayed('_merge_stats_to_model-5894730f-b702-4fd7-889a-89185a179cae')]\n",
      "[I] [08:01:48.817455] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-b4d03902-891d-4af6-a5dc-011eab654c02>]\n",
      "[I] [08:01:48.817702] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-33b18500-16da-45b9-af81-bfc4c411dee2>]\n",
      "[I] [08:01:48.817844] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-725aa57e-7ed7-4dd8-b698-75ad32ef1956>]\n",
      "[I] [08:01:48.818012] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-96d2b721-b2e0-4b32-9eb5-ff5424c0dbe3>]\n",
      "[I] [08:01:48.818129] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-06a3e5ab-9716-4f4b-82dd-dfe7760f1ce6>]\n",
      "[I] [08:01:48.818247] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ad787bc3-5a0d-4cec-9792-0ab47a786ff7>]\n",
      "[I] [08:01:48.818360] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-57e8a6ba-bc5b-4fc9-a1c0-8b6051c11cfe>]\n",
      "[I] [08:01:48.818624] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-5894730f-b702-4fd7-889a-89185a179cae>]\n",
      "[I] [08:01:48.826151] [Delayed('_merge_stats_to_model-c6448c12-c8e5-4450-a1ed-ee2dad4a67be')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 40%|█████████████████████████████████████████████████████████████████████▌                                                                                                        | 2/5 [00:43<01:05, 21.74s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:8, Sample Run:1, Finished loading data in 2.079169233s\n",
      "Workers:8, Sample Run:1, Finished preprocessing data in 5.384161838s\n",
      "Workers:8, Sample Run:1, Finished fitting HashVectorizer in 5.744474666s\n",
      "Workers:8, Sample Run:1, Finished fitting IDF Transformer in 1.697279404s\n",
      "Workers:8, Sample Run:1, Finished executing full pipeline in 21.922310933s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:02:11.169684] [Delayed('_merge_stats_to_model-52d7f1a7-efb7-451f-ab48-3d20dc129e12')]\n",
      "[I] [08:02:11.172257] [Delayed('_merge_stats_to_model-806c2a09-6d70-4287-abf8-6417ab22bb93')]\n",
      "[I] [08:02:11.173933] [Delayed('_merge_stats_to_model-bcb3c149-e2c0-47bf-bf88-b77df9974202')]\n",
      "[I] [08:02:11.177166] [Delayed('_merge_stats_to_model-b5c90890-08bb-4e60-bea4-195480dfc7c4')]\n",
      "[I] [08:02:11.178581] [Delayed('_merge_stats_to_model-a0454257-c1cf-4b50-86bc-02c7661427f4')]\n",
      "[I] [08:02:11.180098] [Delayed('_merge_stats_to_model-319b83be-6086-4603-a072-5568380f5bb8')]\n",
      "[I] [08:02:11.181465] [Delayed('_merge_stats_to_model-6d3d1fb7-ec91-42e4-abce-e775964a99aa')]\n",
      "[I] [08:02:11.182795] [Delayed('_merge_stats_to_model-caa3eced-6799-4463-8f43-2835ae61d104')]\n",
      "[I] [08:02:11.466361] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-52d7f1a7-efb7-451f-ab48-3d20dc129e12>]\n",
      "[I] [08:02:11.466677] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-806c2a09-6d70-4287-abf8-6417ab22bb93>]\n",
      "[I] [08:02:11.466842] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-bcb3c149-e2c0-47bf-bf88-b77df9974202>]\n",
      "[I] [08:02:11.466982] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-b5c90890-08bb-4e60-bea4-195480dfc7c4>]\n",
      "[I] [08:02:11.467146] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-a0454257-c1cf-4b50-86bc-02c7661427f4>]\n",
      "[I] [08:02:11.467289] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-319b83be-6086-4603-a072-5568380f5bb8>]\n",
      "[I] [08:02:11.467574] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-6d3d1fb7-ec91-42e4-abce-e775964a99aa>]\n",
      "[I] [08:02:11.467726] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-caa3eced-6799-4463-8f43-2835ae61d104>]\n",
      "[I] [08:02:11.478331] [Delayed('_merge_stats_to_model-b190cc01-ad8c-4aa4-b212-88f31cf72416')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 60%|████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                                                     | 3/5 [01:06<00:44, 22.40s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:8, Sample Run:2, Finished loading data in 2.201501475s\n",
      "Workers:8, Sample Run:2, Finished preprocessing data in 5.841224353s\n",
      "Workers:8, Sample Run:2, Finished fitting HashVectorizer in 6.222125601s\n",
      "Workers:8, Sample Run:2, Finished fitting IDF Transformer in 2.514887639s\n",
      "Workers:8, Sample Run:2, Finished executing full pipeline in 23.133723633s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:02:33.078308] [Delayed('_merge_stats_to_model-76e9f355-20e6-4a6a-a5f9-72a9af617cd4')]\n",
      "[I] [08:02:33.080683] [Delayed('_merge_stats_to_model-c17d1b1e-b4de-45a4-a86e-250c7f4fead9')]\n",
      "[I] [08:02:33.082033] [Delayed('_merge_stats_to_model-3f6f3806-48a4-4fe2-bb31-a438d3bbdd2b')]\n",
      "[I] [08:02:33.083294] [Delayed('_merge_stats_to_model-49e447e7-3561-4caf-a93c-fbe38bbb74e3')]\n",
      "[I] [08:02:33.084440] [Delayed('_merge_stats_to_model-909cad06-0d6b-43ad-a2da-19b32c412ffd')]\n",
      "[I] [08:02:33.085830] [Delayed('_merge_stats_to_model-e78c48f2-f9b3-4594-b61b-27c4fc4decfe')]\n",
      "[I] [08:02:33.087025] [Delayed('_merge_stats_to_model-e31224f2-4998-40ab-abfd-ba6295f343a2')]\n",
      "[I] [08:02:33.088449] [Delayed('_merge_stats_to_model-ba3b4928-ad32-417c-805a-fe5f057a535b')]\n",
      "[I] [08:02:33.134861] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-76e9f355-20e6-4a6a-a5f9-72a9af617cd4>]\n",
      "[I] [08:02:33.135097] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-c17d1b1e-b4de-45a4-a86e-250c7f4fead9>]\n",
      "[I] [08:02:33.135228] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-3f6f3806-48a4-4fe2-bb31-a438d3bbdd2b>]\n",
      "[I] [08:02:33.135347] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-49e447e7-3561-4caf-a93c-fbe38bbb74e3>]\n",
      "[I] [08:02:33.135462] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-909cad06-0d6b-43ad-a2da-19b32c412ffd>]\n",
      "[I] [08:02:33.135617] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-e78c48f2-f9b3-4594-b61b-27c4fc4decfe>]\n",
      "[I] [08:02:33.135754] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-e31224f2-4998-40ab-abfd-ba6295f343a2>]\n",
      "[I] [08:02:33.136220] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ba3b4928-ad32-417c-805a-fe5f057a535b>]\n",
      "[I] [08:02:33.145305] [Delayed('_merge_stats_to_model-b587beea-abc5-4959-9715-7da6b732c21c')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 80%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████▏                                  | 4/5 [01:27<00:21, 21.96s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:8, Sample Run:3, Finished loading data in 2.134167161s\n",
      "Workers:8, Sample Run:3, Finished preprocessing data in 5.299706862s\n",
      "Workers:8, Sample Run:3, Finished fitting HashVectorizer in 5.77923249s\n",
      "Workers:8, Sample Run:3, Finished fitting IDF Transformer in 1.849184448s\n",
      "Workers:8, Sample Run:3, Finished executing full pipeline in 21.235994979s\n",
      "20726160\n",
      "20725971\n",
      "# nnz: 0\n",
      "(20725971, 1048576)\n",
      "[I] [08:02:56.382959] [Delayed('_merge_stats_to_model-ad2d2b9f-8ac2-4c56-a9d5-161f3d7a7e2b')]\n",
      "[I] [08:02:56.384456] [Delayed('_merge_stats_to_model-cdc1e4b5-8619-4f68-bde2-d3cf88298500')]\n",
      "[I] [08:02:56.385664] [Delayed('_merge_stats_to_model-770fdac2-e04b-49de-8e55-70a7468c59bf')]\n",
      "[I] [08:02:56.386969] [Delayed('_merge_stats_to_model-6e5469f3-99d2-4bbe-b80a-6a2aac787b8d')]\n",
      "[I] [08:02:56.388278] [Delayed('_merge_stats_to_model-8f9f6ee1-ed42-4327-85da-93db87d32e8a')]\n",
      "[I] [08:02:56.389532] [Delayed('_merge_stats_to_model-0a20b3f1-8da1-4e86-9dd2-1c57150049ab')]\n",
      "[I] [08:02:56.390699] [Delayed('_merge_stats_to_model-edb17fd4-3791-418a-9d03-3edb67014e5b')]\n",
      "[I] [08:02:56.392101] [Delayed('_merge_stats_to_model-158612f0-c81a-4df1-86b0-6a0f4b563d1c')]\n",
      "[I] [08:02:56.434685] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-ad2d2b9f-8ac2-4c56-a9d5-161f3d7a7e2b>]\n",
      "[I] [08:02:56.434927] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-cdc1e4b5-8619-4f68-bde2-d3cf88298500>]\n",
      "[I] [08:02:56.435061] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-770fdac2-e04b-49de-8e55-70a7468c59bf>]\n",
      "[I] [08:02:56.435180] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-6e5469f3-99d2-4bbe-b80a-6a2aac787b8d>]\n",
      "[I] [08:02:56.435297] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-8f9f6ee1-ed42-4327-85da-93db87d32e8a>]\n",
      "[I] [08:02:56.435428] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-0a20b3f1-8da1-4e86-9dd2-1c57150049ab>]\n",
      "[I] [08:02:56.435575] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-edb17fd4-3791-418a-9d03-3edb67014e5b>]\n",
      "[I] [08:02:56.436033] [<Future: finished, type: cuml.TfidfTransformer, key: _merge_stats_to_model-158612f0-c81a-4df1-86b0-6a0f4b563d1c>]\n",
      "[I] [08:02:56.443886] [Delayed('_merge_stats_to_model-220245a5-d07f-47d5-802b-7913b91e6d7f')]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [01:51<00:00, 22.32s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20725971, 1048576)\n",
      "Workers:8, Sample Run:4, Finished loading data in 2.192760121s\n",
      "Workers:8, Sample Run:4, Finished preprocessing data in 6.155102677s\n",
      "Workers:8, Sample Run:4, Finished fitting HashVectorizer in 6.345779296s\n",
      "Workers:8, Sample Run:4, Finished fitting IDF Transformer in 2.301130579s\n",
      "Workers:8, Sample Run:4, Finished executing full pipeline in 23.69564001s\n",
      "CPU times: user 37.2 s, sys: 12 s, total: 49.2 s\n",
      "Wall time: 10min 31s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "performance_numbers(client, parquet_path=parquet_path, worker_counts=worker_counts, samples=samples, result_path=result_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3bd627f9-0878-4296-bf73-151ec1b1bec3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data, melt_data = visualize_data(result_path)        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "36dc1a20-d436-4362-af91-75d4b6e829b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "groupby = data.groupby(\"n_workers\").agg(['mean', 'std', 'count'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c0ac4d3f-5d2f-4941-a984-adad8d15edb8",
   "metadata": {},
   "outputs": [
    {
     "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">overall</th>\n",
       "      <th colspan=\"3\" halign=\"left\">data_read</th>\n",
       "      <th colspan=\"3\" halign=\"left\">hashing_vectorizer</th>\n",
       "      <th colspan=\"3\" halign=\"left\">tfidf_transformer</th>\n",
       "      <th colspan=\"3\" halign=\"left\">data_preprocessing</th>\n",
       "      <th colspan=\"3\" halign=\"left\">nrows</th>\n",
       "      <th colspan=\"3\" halign=\"left\">sample_run</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>...</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>n_workers</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>43.601849</td>\n",
       "      <td>0.733715</td>\n",
       "      <td>5</td>\n",
       "      <td>3.676809</td>\n",
       "      <td>0.804110</td>\n",
       "      <td>5</td>\n",
       "      <td>16.154609</td>\n",
       "      <td>0.174656</td>\n",
       "      <td>5</td>\n",
       "      <td>1.182243</td>\n",
       "      <td>...</td>\n",
       "      <td>5</td>\n",
       "      <td>18.526173</td>\n",
       "      <td>0.046918</td>\n",
       "      <td>5</td>\n",
       "      <td>20725971</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1.581139</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31.045885</td>\n",
       "      <td>1.455146</td>\n",
       "      <td>5</td>\n",
       "      <td>2.499536</td>\n",
       "      <td>0.111701</td>\n",
       "      <td>5</td>\n",
       "      <td>9.627595</td>\n",
       "      <td>0.377252</td>\n",
       "      <td>5</td>\n",
       "      <td>3.407960</td>\n",
       "      <td>...</td>\n",
       "      <td>5</td>\n",
       "      <td>10.764249</td>\n",
       "      <td>0.710378</td>\n",
       "      <td>5</td>\n",
       "      <td>20725971</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1.581139</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>25.068612</td>\n",
       "      <td>1.169822</td>\n",
       "      <td>5</td>\n",
       "      <td>2.190867</td>\n",
       "      <td>0.093679</td>\n",
       "      <td>5</td>\n",
       "      <td>7.196249</td>\n",
       "      <td>0.276245</td>\n",
       "      <td>5</td>\n",
       "      <td>2.872289</td>\n",
       "      <td>...</td>\n",
       "      <td>5</td>\n",
       "      <td>7.466962</td>\n",
       "      <td>0.227338</td>\n",
       "      <td>5</td>\n",
       "      <td>20725971</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1.581139</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>22.267771</td>\n",
       "      <td>1.096895</td>\n",
       "      <td>5</td>\n",
       "      <td>2.142816</td>\n",
       "      <td>0.053349</td>\n",
       "      <td>5</td>\n",
       "      <td>6.060744</td>\n",
       "      <td>0.278152</td>\n",
       "      <td>5</td>\n",
       "      <td>2.058753</td>\n",
       "      <td>...</td>\n",
       "      <td>5</td>\n",
       "      <td>5.670149</td>\n",
       "      <td>0.347658</td>\n",
       "      <td>5</td>\n",
       "      <td>20725971</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1.581139</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             overall                 data_read                  \\\n",
       "                mean       std count      mean       std count   \n",
       "n_workers                                                        \n",
       "2          43.601849  0.733715     5  3.676809  0.804110     5   \n",
       "4          31.045885  1.455146     5  2.499536  0.111701     5   \n",
       "6          25.068612  1.169822     5  2.190867  0.093679     5   \n",
       "8          22.267771  1.096895     5  2.142816  0.053349     5   \n",
       "\n",
       "          hashing_vectorizer                 tfidf_transformer  ...        \\\n",
       "                        mean       std count              mean  ... count   \n",
       "n_workers                                                       ...         \n",
       "2                  16.154609  0.174656     5          1.182243  ...     5   \n",
       "4                   9.627595  0.377252     5          3.407960  ...     5   \n",
       "6                   7.196249  0.276245     5          2.872289  ...     5   \n",
       "8                   6.060744  0.278152     5          2.058753  ...     5   \n",
       "\n",
       "          data_preprocessing                     nrows            sample_run  \\\n",
       "                        mean       std count      mean  std count       mean   \n",
       "n_workers                                                                      \n",
       "2                  18.526173  0.046918     5  20725971  0.0     5          2   \n",
       "4                  10.764249  0.710378     5  20725971  0.0     5          2   \n",
       "6                   7.466962  0.227338     5  20725971  0.0     5          2   \n",
       "8                   5.670149  0.347658     5  20725971  0.0     5          2   \n",
       "\n",
       "                           \n",
       "                std count  \n",
       "n_workers                  \n",
       "2          1.581139     5  \n",
       "4          1.581139     5  \n",
       "6          1.581139     5  \n",
       "8          1.581139     5  \n",
       "\n",
       "[4 rows x 21 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "groupby"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "aec87629-1843-4b99-97d3-c33baea7dcc4",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.close()\n",
    "cluster.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df34f385-6187-4753-adc3-cd9bc2fe90b5",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}