github.com/NVIDIA/aistore@v1.3.23-0.20240517131212-7df6609be51d/python/examples/dask/dask-aistore-demo.ipynb (about)

     1  {
     2   "cells": [
     3    {
     4     "cell_type": "markdown",
     5     "metadata": {},
     6     "source": [
     7      "## **AIStore w/ Dask**"
     8     ]
     9    },
    10    {
    11     "cell_type": "markdown",
    12     "metadata": {},
    13     "source": [
    14      "[Dask](https://www.dask.org/) is a new and popular open-source Python library for parallel computing. Dask provides this parallel functionality for a single machine up to a distributed cluster."
    15     ]
    16    },
    17    {
    18     "cell_type": "markdown",
    19     "metadata": {},
    20     "source": [
    21      "#### **Pandas v. Dask**\n",
    22      "\n",
    23      "For data scientists, [Pandas](https://pandas.pydata.org/) has long been the preferred tool for analyzing and manipulating data with Python. However, Dask provides similar capabilities, with the notable exception being Dask's *scalability* with dataset size. \n",
    24      "\n",
    25      "[Dask DataFrames](https://docs.dask.org/en/stable/dataframe.html) are essentially, a large, virtual dataframe divided along the index into multiple Pandas DataFrames. In fact, the `dask.DataFrame` API is a [subset](https://docs.dask.org/en/stable/dataframe.html#dask-dataframe-copies-the-pandas-dataframe-api) of `pd.DataFrame` API and should be familiar to those already familiar with [Pandas DataFrames](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.html).\n",
    26      "\n",
    27      "> For more information on Dask DataFrame API use, please refer to the [Dask DataFrame API reference](https://docs.dask.org/en/stable/dataframe-api.html).\n",
    28      "\n",
    29      "Dask `DataFrames` support data access via HTTP(s) protocol and AIStore clusters support data access via [REST API](https://aiatscale.org/docs/http-api). REST API allows for data on AIStore to be read, written, and otherwise operated on via HTTP(s) verbs."
    30     ]
    31    },
    32    {
    33     "cell_type": "markdown",
    34     "metadata": {},
    35     "source": [
    36      "### <ins>**Getting Started**"
    37     ]
    38    },
    39    {
    40     "cell_type": "markdown",
    41     "metadata": {},
    42     "source": [
    43      "Install Dask with `pip` as follows:\n",
    44      "\n",
    45      "```console\n",
    46      "$ python -m pip install dask\n",
    47      "```\n",
    48      "\n",
    49      "Install the latest AIStore release either with `conda` or `pip`:\n",
    50      "\n",
    51      "```console\n",
    52      "$ conda install aistore\n",
    53      "```\n",
    54      "\n",
    55      "```console\n",
    56      "$ pip install aistore\n",
    57      "```"
    58     ]
    59    },
    60    {
    61     "cell_type": "markdown",
    62     "metadata": {},
    63     "source": [
    64      "Start by deploying an AIStore cluster. The following demonstrations will be utilizing a [Minikube (Kubernetes) deployment](https://github.com/NVIDIA/aistore/blob/master/deploy/dev/k8s/README.md) of AIStore. \n",
    65      "\n",
    66      "> For information on AIStore deployment options, refer [here](https://github.com/NVIDIA/aistore/blob/master/deploy/README.md).\n",
    67      "\n",
    68      "Once deployed, import the `aistore` package and initialize a `Client`:"
    69     ]
    70    },
    71    {
    72     "cell_type": "code",
    73     "execution_count": 4,
    74     "metadata": {},
    75     "outputs": [
    76      {
    77       "data": {
    78        "text/plain": [
    79         "[BucketEntry(name='zillow.csv', size=842, checksum='8b118b6e12d5b4b1', atime='15 Aug 22 15:13 UTC', version='', target_url='', copies=0, flags=64)]"
    80        ]
    81       },
    82       "execution_count": 4,
    83       "metadata": {},
    84       "output_type": "execute_result"
    85      }
    86     ],
    87     "source": [
    88      "from aistore import Client\n",
    89      "\n",
    90      "AIS_ENDPOINT = \"http://192.168.49.2:8080\"\n",
    91      "\n",
    92      "# Initialize AIStore client\n",
    93      "client = Client(AIS_ENDPOINT)\n",
    94      "\n",
    95      "# Load sample data to AIStore for following demonstrations (Local to AIStore)\n",
    96      "client.bucket(\"dask-demo-bucket\").create()\n",
    97      "client.bucket(\"dask-demo-bucket\").object(\"zillow.csv\").put(\"./data/zillow.csv\")\n",
    98      "\n",
    99      "# Verify that the object is in bucket\n",
   100      "client.bucket(\"dask-demo-bucket\").list_objects().get_entries()"
   101     ]
   102    },
   103    {
   104     "cell_type": "markdown",
   105     "metadata": {},
   106     "source": [
   107      "### <ins>**ETL** (**E**xtract-**T**ransform-**L**oad)"
   108     ]
   109    },
   110    {
   111     "cell_type": "markdown",
   112     "metadata": {},
   113     "source": [
   114      "**Note:** ETL processes using Dask with data on AIStore *are* possible, but has limitations as Dask does *not* currently support AIStore as a recognized storage provider. Refer to the AIS-ETL service [here](https://github.com/NVIDIA/aistore/blob/master/docs/etl.md), which offers both *offline* and *inline* custom transformations, as well as flexibility to the scope of those transformations (bucket-specific or object(s)-specific), and more."
   115     ]
   116    },
   117    {
   118     "cell_type": "markdown",
   119     "metadata": {},
   120     "source": [
   121      "#### **Extract:**"
   122     ]
   123    },
   124    {
   125     "cell_type": "markdown",
   126     "metadata": {},
   127     "source": [
   128      "Initialize the Dask [`DataFrame`](https://docs.dask.org/en/stable/dataframe.html) with a direct HTTP(s) address to data residing on AIStore:"
   129     ]
   130    },
   131    {
   132     "cell_type": "code",
   133     "execution_count": 5,
   134     "metadata": {},
   135     "outputs": [],
   136     "source": [
   137      "import dask.dataframe as dd\n",
   138      "\n",
   139      "\n",
   140      "def read_csv_ais(bck_name: str, obj_name: str):\n",
   141      "    return dd.read_csv(f\"{AIS_ENDPOINT}/v1/objects/{bck_name}/{obj_name}\")\n",
   142      "\n",
   143      "\n",
   144      "# Initialize DataFrame\n",
   145      "df = read_csv_ais(bck_name=\"dask-demo-bucket\", obj_name=\"zillow.csv\")"
   146     ]
   147    },
   148    {
   149     "cell_type": "markdown",
   150     "metadata": {},
   151     "source": [
   152      "Dask `DataFrames` are split into smaller partitions and partitions are loaded into memory *as needed*. The method `head()` allows us to see just the first few lines of the data:"
   153     ]
   154    },
   155    {
   156     "cell_type": "code",
   157     "execution_count": 6,
   158     "metadata": {},
   159     "outputs": [
   160      {
   161       "data": {
   162        "text/html": [
   163         "<div>\n",
   164         "<style scoped>\n",
   165         "    .dataframe tbody tr th:only-of-type {\n",
   166         "        vertical-align: middle;\n",
   167         "    }\n",
   168         "\n",
   169         "    .dataframe tbody tr th {\n",
   170         "        vertical-align: top;\n",
   171         "    }\n",
   172         "\n",
   173         "    .dataframe thead th {\n",
   174         "        text-align: right;\n",
   175         "    }\n",
   176         "</style>\n",
   177         "<table border=\"1\" class=\"dataframe\">\n",
   178         "  <thead>\n",
   179         "    <tr style=\"text-align: right;\">\n",
   180         "      <th></th>\n",
   181         "      <th>Index</th>\n",
   182         "      <th>\"Living Space (sq ft)\"</th>\n",
   183         "      <th>\"Beds\"</th>\n",
   184         "      <th>\"Baths\"</th>\n",
   185         "      <th>\"Zip\"</th>\n",
   186         "      <th>\"Year\"</th>\n",
   187         "      <th>\"List Price ($)\"</th>\n",
   188         "    </tr>\n",
   189         "  </thead>\n",
   190         "  <tbody>\n",
   191         "    <tr>\n",
   192         "      <th>0</th>\n",
   193         "      <td>1</td>\n",
   194         "      <td>2222</td>\n",
   195         "      <td>3</td>\n",
   196         "      <td>3.5</td>\n",
   197         "      <td>32312</td>\n",
   198         "      <td>1981</td>\n",
   199         "      <td>250000</td>\n",
   200         "    </tr>\n",
   201         "    <tr>\n",
   202         "      <th>1</th>\n",
   203         "      <td>2</td>\n",
   204         "      <td>1628</td>\n",
   205         "      <td>3</td>\n",
   206         "      <td>2.0</td>\n",
   207         "      <td>32308</td>\n",
   208         "      <td>2009</td>\n",
   209         "      <td>185000</td>\n",
   210         "    </tr>\n",
   211         "    <tr>\n",
   212         "      <th>2</th>\n",
   213         "      <td>3</td>\n",
   214         "      <td>3824</td>\n",
   215         "      <td>5</td>\n",
   216         "      <td>4.0</td>\n",
   217         "      <td>32312</td>\n",
   218         "      <td>1954</td>\n",
   219         "      <td>399000</td>\n",
   220         "    </tr>\n",
   221         "    <tr>\n",
   222         "      <th>3</th>\n",
   223         "      <td>4</td>\n",
   224         "      <td>1137</td>\n",
   225         "      <td>3</td>\n",
   226         "      <td>2.0</td>\n",
   227         "      <td>32309</td>\n",
   228         "      <td>1993</td>\n",
   229         "      <td>150000</td>\n",
   230         "    </tr>\n",
   231         "    <tr>\n",
   232         "      <th>4</th>\n",
   233         "      <td>5</td>\n",
   234         "      <td>3560</td>\n",
   235         "      <td>6</td>\n",
   236         "      <td>4.0</td>\n",
   237         "      <td>32309</td>\n",
   238         "      <td>1973</td>\n",
   239         "      <td>315000</td>\n",
   240         "    </tr>\n",
   241         "  </tbody>\n",
   242         "</table>\n",
   243         "</div>"
   244        ],
   245        "text/plain": [
   246         "   Index   \"Living Space (sq ft)\"   \"Beds\"   \"Baths\"   \"Zip\"   \"Year\"  \\\n",
   247         "0      1                     2222        3       3.5   32312     1981   \n",
   248         "1      2                     1628        3       2.0   32308     2009   \n",
   249         "2      3                     3824        5       4.0   32312     1954   \n",
   250         "3      4                     1137        3       2.0   32309     1993   \n",
   251         "4      5                     3560        6       4.0   32309     1973   \n",
   252         "\n",
   253         "    \"List Price ($)\"  \n",
   254         "0             250000  \n",
   255         "1             185000  \n",
   256         "2             399000  \n",
   257         "3             150000  \n",
   258         "4             315000  "
   259        ]
   260       },
   261       "execution_count": 6,
   262       "metadata": {},
   263       "output_type": "execute_result"
   264      }
   265     ],
   266     "source": [
   267      "# View first partition of data\n",
   268      "df.head()"
   269     ]
   270    },
   271    {
   272     "cell_type": "code",
   273     "execution_count": 7,
   274     "metadata": {},
   275     "outputs": [
   276      {
   277       "data": {
   278        "text/html": [
   279         "<div>\n",
   280         "<style scoped>\n",
   281         "    .dataframe tbody tr th:only-of-type {\n",
   282         "        vertical-align: middle;\n",
   283         "    }\n",
   284         "\n",
   285         "    .dataframe tbody tr th {\n",
   286         "        vertical-align: top;\n",
   287         "    }\n",
   288         "\n",
   289         "    .dataframe thead th {\n",
   290         "        text-align: right;\n",
   291         "    }\n",
   292         "</style>\n",
   293         "<table border=\"1\" class=\"dataframe\">\n",
   294         "  <thead>\n",
   295         "    <tr style=\"text-align: right;\">\n",
   296         "      <th></th>\n",
   297         "      <th>Index</th>\n",
   298         "      <th>\"Living Space (sq ft)\"</th>\n",
   299         "      <th>\"Beds\"</th>\n",
   300         "      <th>\"Baths\"</th>\n",
   301         "      <th>\"Zip\"</th>\n",
   302         "      <th>\"Year\"</th>\n",
   303         "      <th>\"List Price ($)\"</th>\n",
   304         "    </tr>\n",
   305         "  </thead>\n",
   306         "  <tbody>\n",
   307         "    <tr>\n",
   308         "      <th>0</th>\n",
   309         "      <td>1</td>\n",
   310         "      <td>2222</td>\n",
   311         "      <td>3</td>\n",
   312         "      <td>3.5</td>\n",
   313         "      <td>32312</td>\n",
   314         "      <td>1981</td>\n",
   315         "      <td>250000</td>\n",
   316         "    </tr>\n",
   317         "    <tr>\n",
   318         "      <th>1</th>\n",
   319         "      <td>2</td>\n",
   320         "      <td>1628</td>\n",
   321         "      <td>3</td>\n",
   322         "      <td>2.0</td>\n",
   323         "      <td>32308</td>\n",
   324         "      <td>2009</td>\n",
   325         "      <td>185000</td>\n",
   326         "    </tr>\n",
   327         "  </tbody>\n",
   328         "</table>\n",
   329         "</div>"
   330        ],
   331        "text/plain": [
   332         "   Index   \"Living Space (sq ft)\"   \"Beds\"   \"Baths\"   \"Zip\"   \"Year\"  \\\n",
   333         "0      1                     2222        3       3.5   32312     1981   \n",
   334         "1      2                     1628        3       2.0   32308     2009   \n",
   335         "\n",
   336         "    \"List Price ($)\"  \n",
   337         "0             250000  \n",
   338         "1             185000  "
   339        ]
   340       },
   341       "execution_count": 7,
   342       "metadata": {},
   343       "output_type": "execute_result"
   344      }
   345     ],
   346     "source": [
   347      "# View first two lines of data\n",
   348      "df.head(2)"
   349     ]
   350    },
   351    {
   352     "cell_type": "markdown",
   353     "metadata": {},
   354     "source": [
   355      "#### **Transform:**"
   356     ]
   357    },
   358    {
   359     "cell_type": "markdown",
   360     "metadata": {},
   361     "source": [
   362      "As mentioned before, Dask `DataFrames` are built on and in parallel with the Pandas API. The `dask.dataframe` API is very similar to the [Pandas](https://docs.dask.org/en/stable/dataframe.html#dask-dataframe-copies-the-pandas-dataframe-api) API and should be very familiar to those already familiar with Pandas:"
   363     ]
   364    },
   365    {
   366     "cell_type": "code",
   367     "execution_count": 8,
   368     "metadata": {},
   369     "outputs": [
   370      {
   371       "name": "stdout",
   372       "output_type": "stream",
   373       "text": [
   374        "Index(['Index', ' \"Living Space (sq ft)\"', ' \"Beds\"', ' \"Baths\"', ' \"Zip\"',\n",
   375        "       ' \"Year\"', ' \"List Price ($)\"'],\n",
   376        "      dtype='object')\n"
   377       ]
   378      },
   379      {
   380       "data": {
   381        "text/html": [
   382         "<div>\n",
   383         "<style scoped>\n",
   384         "    .dataframe tbody tr th:only-of-type {\n",
   385         "        vertical-align: middle;\n",
   386         "    }\n",
   387         "\n",
   388         "    .dataframe tbody tr th {\n",
   389         "        vertical-align: top;\n",
   390         "    }\n",
   391         "\n",
   392         "    .dataframe thead th {\n",
   393         "        text-align: right;\n",
   394         "    }\n",
   395         "</style>\n",
   396         "<table border=\"1\" class=\"dataframe\">\n",
   397         "  <thead>\n",
   398         "    <tr style=\"text-align: right;\">\n",
   399         "      <th></th>\n",
   400         "      <th>index</th>\n",
   401         "      <th>Index</th>\n",
   402         "      <th>\"Living Space (sq ft)\"</th>\n",
   403         "      <th>\"Beds\"</th>\n",
   404         "      <th>\"Baths\"</th>\n",
   405         "      <th>\"Zip\"</th>\n",
   406         "      <th>\"Year\"</th>\n",
   407         "      <th>\"List Price ($)\"</th>\n",
   408         "    </tr>\n",
   409         "  </thead>\n",
   410         "  <tbody>\n",
   411         "    <tr>\n",
   412         "      <th>0</th>\n",
   413         "      <td>1</td>\n",
   414         "      <td>2</td>\n",
   415         "      <td>1628</td>\n",
   416         "      <td>3</td>\n",
   417         "      <td>2.0</td>\n",
   418         "      <td>32308</td>\n",
   419         "      <td>2009</td>\n",
   420         "      <td>185000</td>\n",
   421         "    </tr>\n",
   422         "    <tr>\n",
   423         "      <th>1</th>\n",
   424         "      <td>9</td>\n",
   425         "      <td>10</td>\n",
   426         "      <td>1997</td>\n",
   427         "      <td>3</td>\n",
   428         "      <td>3.0</td>\n",
   429         "      <td>32311</td>\n",
   430         "      <td>2006</td>\n",
   431         "      <td>295000</td>\n",
   432         "    </tr>\n",
   433         "    <tr>\n",
   434         "      <th>2</th>\n",
   435         "      <td>10</td>\n",
   436         "      <td>11</td>\n",
   437         "      <td>2097</td>\n",
   438         "      <td>4</td>\n",
   439         "      <td>3.0</td>\n",
   440         "      <td>32311</td>\n",
   441         "      <td>2016</td>\n",
   442         "      <td>290000</td>\n",
   443         "    </tr>\n",
   444         "    <tr>\n",
   445         "      <th>3</th>\n",
   446         "      <td>14</td>\n",
   447         "      <td>15</td>\n",
   448         "      <td>1381</td>\n",
   449         "      <td>3</td>\n",
   450         "      <td>2.0</td>\n",
   451         "      <td>32301</td>\n",
   452         "      <td>2006</td>\n",
   453         "      <td>143000</td>\n",
   454         "    </tr>\n",
   455         "  </tbody>\n",
   456         "</table>\n",
   457         "</div>"
   458        ],
   459        "text/plain": [
   460         "   index  Index   \"Living Space (sq ft)\"   \"Beds\"   \"Baths\"   \"Zip\"   \"Year\"  \\\n",
   461         "0      1      2                     1628        3       2.0   32308     2009   \n",
   462         "1      9     10                     1997        3       3.0   32311     2006   \n",
   463         "2     10     11                     2097        4       3.0   32311     2016   \n",
   464         "3     14     15                     1381        3       2.0   32301     2006   \n",
   465         "\n",
   466         "    \"List Price ($)\"  \n",
   467         "0             185000  \n",
   468         "1             295000  \n",
   469         "2             290000  \n",
   470         "3             143000  "
   471        ]
   472       },
   473       "execution_count": 8,
   474       "metadata": {},
   475       "output_type": "execute_result"
   476      }
   477     ],
   478     "source": [
   479      "# Print columns\n",
   480      "print(df.columns)\n",
   481      "\n",
   482      "# Filter datapoints for houses built AFTER the year 2000\n",
   483      "df_new = df[df[' \"Year\"'] > 2000]\n",
   484      "# Further filter for houses with a price less than $300,000\n",
   485      "df_new_and_budget = df_new[df_new[' \"List Price ($)\"'] < 300000]\n",
   486      "# Reindex\n",
   487      "df_final = df_new_and_budget.reset_index()\n",
   488      "\n",
   489      "# Verify that data has been correctly filtered\n",
   490      "df_final.head()"
   491     ]
   492    },
   493    {
   494     "cell_type": "markdown",
   495     "metadata": {},
   496     "source": [
   497      "Dask also provides other tools that efficiently process a variety of different types of files, such as [images](https://examples.dask.org/applications/image-processing.html). The following snippet demonstrates image pre-processing capabilities with Dask using the [`dask-image`](https://github.com/dask/dask-image) module, similar to image transformations demonstrated with AIS-ETLs [here](https://aiatscale.org/blog/2021/10/22/ais-etl-2):"
   498     ]
   499    },
   500    {
   501     "cell_type": "code",
   502     "execution_count": 29,
   503     "metadata": {},
   504     "outputs": [
   505      {
   506       "data": {
   507        "text/plain": [
   508         "<matplotlib.image.AxesImage at 0x7f0e38725ab0>"
   509        ]
   510       },
   511       "execution_count": 29,
   512       "metadata": {},
   513       "output_type": "execute_result"
   514      },
   515      {
   516       "data": {
   517        "image/png": "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",
   518        "text/plain": [
   519         "<Figure size 432x288 with 1 Axes>"
   520        ]
   521       },
   522       "metadata": {
   523        "needs_background": "light"
   524       },
   525       "output_type": "display_data"
   526      }
   527     ],
   528     "source": [
   529      "from nturl2path import url2pathname\n",
   530      "import imageio.v3\n",
   531      "from io import BytesIO\n",
   532      "import matplotlib.pyplot as plt\n",
   533      "import os\n",
   534      "import requests\n",
   535      "\n",
   536      "try:\n",
   537      "    from skimage.io import imread as sk_imread\n",
   538      "except (AttributeError, ImportError):\n",
   539      "    pass\n",
   540      "\n",
   541      "from dask.array.core import Array\n",
   542      "from dask.base import tokenize\n",
   543      "\n",
   544      "\n",
   545      "def add_leading_dimension(x):\n",
   546      "    return x[None, ...]\n",
   547      "\n",
   548      "\n",
   549      "def url_image_reader(url):\n",
   550      "    response = requests.get(url)\n",
   551      "    byte_content = BytesIO(response.content)\n",
   552      "    image = imageio.v3.imread(byte_content)\n",
   553      "    return image\n",
   554      "\n",
   555      "\n",
   556      "# Modified version of dask.array.image.imread (doesn't use glob)\n",
   557      "def custom_imread(filenames, imread=None, preprocess=None):\n",
   558      "    \"\"\"Read a stack of images into a dask array\n",
   559      "    Parameters\n",
   560      "    ----------\n",
   561      "    filenames: list of strings\n",
   562      "        A list of filename strings, eg: ['myfile._01.png', 'myfile_02.png']\n",
   563      "    imread: function (optional)\n",
   564      "        Optionally provide custom imread function.\n",
   565      "        Function should expect a filename and produce a numpy array.\n",
   566      "        Defaults to ``skimage.io.imread``.\n",
   567      "    preprocess: function (optional)\n",
   568      "        Optionally provide custom function to preprocess the image.\n",
   569      "        Function should expect a numpy array for a single image.\n",
   570      "    Examples\n",
   571      "    --------\n",
   572      "    >>> from dask.array.image import imread\n",
   573      "    >>> im = imread('2015-*-*.png')  # doctest: +SKIP\n",
   574      "    >>> im.shape  # doctest: +SKIP\n",
   575      "    (365, 1000, 1000, 3)\n",
   576      "    Returns\n",
   577      "    -------\n",
   578      "    Dask array of all images stacked along the first dimension.\n",
   579      "    Each separate image file will be treated as an individual chunk.\n",
   580      "    \"\"\"\n",
   581      "    imread = imread or sk_imread\n",
   582      "\n",
   583      "    name = \"imread-%s\" % tokenize(filenames, map(os.path.getmtime, filenames))\n",
   584      "\n",
   585      "    sample = imread(filenames[0])\n",
   586      "    if preprocess:\n",
   587      "        sample = preprocess(sample)\n",
   588      "\n",
   589      "    keys = [(name, i) + (0,) * len(sample.shape) for i in range(len(filenames))]\n",
   590      "    if preprocess:\n",
   591      "        values = [\n",
   592      "            (add_leading_dimension, (preprocess, (imread, fn))) for fn in filenames\n",
   593      "        ]\n",
   594      "    else:\n",
   595      "        values = [(add_leading_dimension, (imread, fn)) for fn in filenames]\n",
   596      "    dsk = dict(zip(keys, values))\n",
   597      "\n",
   598      "    chunks = ((1,) * len(filenames),) + tuple((d,) for d in sample.shape)\n",
   599      "\n",
   600      "    return Array(dsk, name, chunks, sample.dtype)\n",
   601      "\n",
   602      "\n",
   603      "filenames = [f\"{AIS_ENDPOINT}/v1/objects/dask-demo-bucket/sample-image.jpg\"]\n",
   604      "result = custom_imread(filenames, imread=url_image_reader)\n",
   605      "\n",
   606      "\n",
   607      "# Pre-Processing function for image\n",
   608      "def custom_filter(rgb):\n",
   609      "    result = (rgb[..., 0] * 0.2125) + (rgb[..., 1] * 0.7154) + (rgb[..., 2] * 0.0721)\n",
   610      "    return result\n",
   611      "\n",
   612      "\n",
   613      "flower = result[0]\n",
   614      "modified_flower = custom_filter(flower)\n",
   615      "\n",
   616      "plt.imshow(modified_flower)"
   617     ]
   618    },
   619    {
   620     "cell_type": "markdown",
   621     "metadata": {},
   622     "source": [
   623      "#### **Load:**"
   624     ]
   625    },
   626    {
   627     "cell_type": "markdown",
   628     "metadata": {},
   629     "source": [
   630      "Dask *does* support many of the popular cloud storage providers (i.e. S3, GCP, Azure), and supports the direct writing of DataFrames as files to those services:\n",
   631      "\n",
   632      "```python\n",
   633      "# Write DataFrame as CSV to S3\n",
   634      "dd.to_csv(\"s3://dask-demo-bucket/sample.csv\")\n",
   635      "\n",
   636      "# Write DataFrame as JSON to GCP\n",
   637      "dd.to_json(\"gcs://dask-demo-bucket/sample.json\")\n",
   638      "```\n",
   639      "\n",
   640      "> AIStore supports a subset of S3 API and Dask supports [S3-compatible storage services](https://docs.dask.org/en/stable/how-to/connect-to-remote-data.html#using-other-s3-compatible-services). However, Dask uses Boto3 for this and AIStore does not support Boto3 as it does not follow HTTP(s) protocol. For more information on AIStore compatibility with S3, refer [here](https://github.com/NVIDIA/aistore/blob/master/docs/s3compat.md).\n",
   641      "\n",
   642      "While we cannot directly write to AIStore with Dask API (`dd.to_csv(\"ais://bucket/file.csv\")` not supported as of now), we can convert the DataFrame to bytes and move it using AIStore's own API:"
   643     ]
   644    },
   645    {
   646     "cell_type": "code",
   647     "execution_count": null,
   648     "metadata": {},
   649     "outputs": [],
   650     "source": [
   651      "# Convert to Pandas DataFrame\n",
   652      "pd = df_new_and_budget.compute()\n",
   653      "files = pd.to_csv(encoding=\"utf-8\")\n",
   654      "\n",
   655      "client.bucket(\"dask-demo-bucket\").list_objects().get_entries()\n",
   656      "\n",
   657      "client.bucket(\"dask-demo-bucket\").object(\"formatted-zillow.csv\").put(content=files)\n",
   658      "\n",
   659      "print(type(files))\n",
   660      "\n",
   661      "# Verify that the transformed file is now in the AIStore bucket\n",
   662      "client.bucket(\"dask-demo-bucket\").list_objects().get_entries()"
   663     ]
   664    },
   665    {
   666     "cell_type": "markdown",
   667     "metadata": {},
   668     "source": [
   669      "However, the above method of converting a Dask DataFrame to a Pandas DataFrame may not be ideal as it defeats some of the performance advantages of using Dask. Converting a Dask DataFrame to a Pandas DataFrame only makes sense to do if the data can fully fit into memory (i.e. data has been filtered and is now much smaller).\n",
   670      "\n",
   671      "For much larger datasets, [AIS-ETL](https://github.com/NVIDIA/aistore/blob/master/docs/etl.md) may offer better performance while offering similar ETL capabilities to those demonstrated above. Please refer [here](https://github.com/NVIDIA/aistore/blob/master/docs/etl.md) for more information."
   672     ]
   673    },
   674    {
   675     "cell_type": "markdown",
   676     "metadata": {},
   677     "source": [
   678      "### <ins>**Data Analysis** "
   679     ]
   680    },
   681    {
   682     "cell_type": "markdown",
   683     "metadata": {},
   684     "source": [
   685      "> Dask is not always the most efficient tool to be used in certain scenarios. Refer [here](https://docs.dask.org/en/latest/best-practices.html#start-small) for information on the best use-cases for Dask.\n",
   686      "\n",
   687      "While Dask's use-cases with AIStore are quite limited as of now, it is still useful for performing data analysis with optimized memory usage. With Dask DataFrames, data can be analyzed **without** loading the entirety of the data into memory. More specifically, because Dask DataFrames (and most other Dask collections) are [lazy](https://saturncloud.io/blog/a-data-scientist-s-guide-to-lazy-evaluation-with-dask/), computations are executed only when the `dask.distributed.compute` method is called. As a result, specific parts of the data are loaded into memory as needed for the computations and are not loaded otherwise, optimizing memory usage:"
   688     ]
   689    },
   690    {
   691     "cell_type": "code",
   692     "execution_count": null,
   693     "metadata": {},
   694     "outputs": [],
   695     "source": [
   696      "# Computations to be made (not computed until .compute() is called)\n",
   697      "mean_price = df[' \"List Price ($)\"'].mean()\n",
   698      "bed_sum = df[' \"Beds\"'].sum()\n",
   699      "mean_size = df[' \"Living Space (sq ft)\"'].mean()\n",
   700      "\n",
   701      "# Computes above in parallel\n",
   702      "dd.compute({\"price_mean\": mean_price, \"bed_sum\": bed_sum, \"mean_size\": mean_size})"
   703     ]
   704    },
   705    {
   706     "cell_type": "markdown",
   707     "metadata": {},
   708     "source": [
   709      "More examples of data analysis with Dask DataFrames can be found [here](https://examples.dask.org/dataframe.html) under Dask documentation."
   710     ]
   711    },
   712    {
   713     "cell_type": "markdown",
   714     "metadata": {},
   715     "source": [
   716      "### <ins>**References**"
   717     ]
   718    },
   719    {
   720     "attachments": {},
   721     "cell_type": "markdown",
   722     "metadata": {},
   723     "source": [
   724      "* [Dask API](https://docs.dask.org/en/stable/dataframe-api.html)\n",
   725      "* [Pandas API](https://pandas.pydata.org/docs/reference/index.html)\n",
   726      "* [AIStore Python SDK](https://github.com/NVIDIA/aistore/blob/master/docs/python_sdk.md)\n",
   727      "* [AIS-ETL](https://github.com/NVIDIA/aistore/blob/master/docs/etl.md)"
   728     ]
   729    }
   730   ],
   731   "metadata": {
   732    "kernelspec": {
   733     "display_name": "Python 3.10.4 64-bit",
   734     "language": "python",
   735     "name": "python3"
   736    },
   737    "language_info": {
   738     "codemirror_mode": {
   739      "name": "ipython",
   740      "version": 3
   741     },
   742     "file_extension": ".py",
   743     "mimetype": "text/x-python",
   744     "name": "python",
   745     "nbconvert_exporter": "python",
   746     "pygments_lexer": "ipython3",
   747     "version": "3.10.6 (main, Nov 14 2022, 16:10:14) [GCC 11.3.0]"
   748    },
   749    "orig_nbformat": 4,
   750    "vscode": {
   751     "interpreter": {
   752      "hash": "916dbcbb3f70747c44a77c7bcd40155683ae19c65e1c03b4aa3499c5328201f1"
   753     }
   754    }
   755   },
   756   "nbformat": 4,
   757   "nbformat_minor": 2
   758  }