{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "name": "2021-06-19-recsys20-tutorial-feature-engineering-part-1.ipynb", "provenance": [], "collapsed_sections": [], "mount_file_id": "1bnZOC6wVT95f8hS9bWb6x03D_FnxaYWL", "authorship_tag": "ABX9TyPmPXgo7GPE2xbCVlv6wGUY" }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" } }, "cells": [ { "cell_type": "markdown", "metadata": { "id": "Qp9CyeCjAP6I" }, "source": [ "# Recsys'20 Feature Engineering Tutorial Part 1\n", "> RecSys'20 tutorial on feature engineering on a large retail dataset\n", "\n", "- toc: true\n", "- badges: true\n", "- comments: true\n", "- categories: [features, recsys, cudf, retail, bigdata]\n", "- image: " ] }, { "cell_type": "markdown", "metadata": { "id": "glqcrXsxGNq1" }, "source": [ "### Data download" ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "nMLPnhbYEm8A", "outputId": "594c3c96-9665-4eb1-964b-6e2c1d3ac780" }, "source": [ "!pip install -q -U kaggle\n", "!pip install --upgrade --force-reinstall --no-deps kaggle\n", "!mkdir ~/.kaggle\n", "!cp /content/drive/MyDrive/kaggle.json ~/.kaggle/\n", "!chmod 600 ~/.kaggle/kaggle.json\n", "\n", "!kaggle datasets download -d mkechinov/ecommerce-behavior-data-from-multi-category-store" ], "execution_count": 1, "outputs": [ { "output_type": "stream", "text": [ "Collecting kaggle\n", "\u001b[?25l Downloading https://files.pythonhosted.org/packages/3a/e7/3bac01547d2ed3d308ac92a0878fbdb0ed0f3d41fb1906c319ccbba1bfbc/kaggle-1.5.12.tar.gz (58kB)\n", "\r\u001b[K |█████▋ | 10kB 6.6MB/s eta 0:00:01\r\u001b[K |███████████▏ | 20kB 10.2MB/s eta 0:00:01\r\u001b[K |████████████████▊ | 30kB 13.1MB/s eta 0:00:01\r\u001b[K |██████████████████████▎ | 40kB 11.6MB/s eta 0:00:01\r\u001b[K |███████████████████████████▉ | 51kB 6.1MB/s eta 0:00:01\r\u001b[K |████████████████████████████████| 61kB 2.9MB/s \n", "\u001b[?25hBuilding wheels for collected packages: kaggle\n", " Building wheel for kaggle (setup.py) ... \u001b[?25l\u001b[?25hdone\n", " Created wheel for kaggle: filename=kaggle-1.5.12-cp37-none-any.whl size=73053 sha256=845e8bef0aaf19c5d7a8d3d66934eb5e52047f126c5e82fe7dcb5a6c1f107b0a\n", " Stored in directory: /root/.cache/pip/wheels/a1/6a/26/d30b7499ff85a4a4593377a87ecf55f7d08af42f0de9b60303\n", "Successfully built kaggle\n", "Installing collected packages: kaggle\n", " Found existing installation: kaggle 1.5.12\n", " Uninstalling kaggle-1.5.12:\n", " Successfully uninstalled kaggle-1.5.12\n", "Successfully installed kaggle-1.5.12\n", "Downloading ecommerce-behavior-data-from-multi-category-store.zip to /content\n", "100% 4.29G/4.29G [01:24<00:00, 50.2MB/s]\n", "100% 4.29G/4.29G [01:24<00:00, 54.3MB/s]\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "2Tac1tdjBvqs", "outputId": "aef95b25-7997-4245-8491-b58692b814fc" }, "source": [ "!gdown --id 1qZIwMbMgMmgDC5EoMdJ8aI9lQPsWA3-P\n", "!gdown --id 1x5ohrrZNhWQN4Q-zww0RmXOwctKHH9PT\n", "!gdown --id 1-Rov9fFtGJqb7_ePc6qH-Rhzxn0cIcKB\n", "!gdown --id 1zr_RXpGvOWN2PrWI6itWL8HnRsCpyqz8\n", "!gdown --id 1g5WoIgLe05UMdREbxAjh0bEFgVCjA1UL" ], "execution_count": 2, "outputs": [ { "output_type": "stream", "text": [ "Downloading...\n", "From: https://drive.google.com/uc?id=1qZIwMbMgMmgDC5EoMdJ8aI9lQPsWA3-P\n", "To: /content/2019-Dec.csv.gz\n", "2.95GB [01:08, 42.9MB/s]\n", "Downloading...\n", "From: https://drive.google.com/uc?id=1x5ohrrZNhWQN4Q-zww0RmXOwctKHH9PT\n", "To: /content/2020-Jan.csv.gz\n", "2.39GB [00:58, 40.7MB/s]\n", "Downloading...\n", "From: https://drive.google.com/uc?id=1-Rov9fFtGJqb7_ePc6qH-Rhzxn0cIcKB\n", "To: /content/2020-Feb.csv.gz\n", "2.35GB [00:56, 41.7MB/s]\n", "Downloading...\n", "From: https://drive.google.com/uc?id=1zr_RXpGvOWN2PrWI6itWL8HnRsCpyqz8\n", "To: /content/2020-Mar.csv.gz\n", "2.42GB [00:40, 60.3MB/s]\n", "Downloading...\n", "From: https://drive.google.com/uc?id=1g5WoIgLe05UMdREbxAjh0bEFgVCjA1UL\n", "To: /content/2020-Apr.csv.gz\n", "2.93GB [00:44, 66.3MB/s]\n" ], "name": "stdout" } ] }, { "cell_type": "markdown", "metadata": { "id": "cbqqSo1mGRFU" }, "source": [ "### Data extract" ] }, { "cell_type": "code", "metadata": { "id": "yEDIXNVkGuCW" }, "source": [ "import os\n", "import gc\n", "import glob\n", "import pandas as pd" ], "execution_count": 3, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "eWEtIxIXDlfW", "outputId": "51f4a064-f076-40e9-a89b-a7f48bb67465" }, "source": [ "!unzip /content/ecommerce-behavior-data-from-multi-category-store.zip\n", "!rm /content/ecommerce-behavior-data-from-multi-category-store.zip" ], "execution_count": 4, "outputs": [ { "output_type": "stream", "text": [ "Archive: /content/ecommerce-behavior-data-from-multi-category-store.zip\n", " inflating: 2019-Nov.csv \n", " inflating: 2019-Oct.csv \n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "IOxOL1pJGpWM", "outputId": "2caa1dbf-462e-40fc-c8fe-fdfb41fcebcf" }, "source": [ "list_gz_files = glob.glob('/content/*.gz')\n", "list_gz_files" ], "execution_count": 5, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "['/content/2020-Jan.csv.gz',\n", " '/content/2020-Mar.csv.gz',\n", " '/content/2019-Dec.csv.gz',\n", " '/content/2020-Apr.csv.gz',\n", " '/content/2020-Feb.csv.gz']" ] }, "metadata": { "tags": [] }, "execution_count": 5 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "PVC7SGvzHFm1", "outputId": "8ee095e2-ad18-4d92-8a0b-013ff484933c" }, "source": [ "for file in list_gz_files:\n", " print(file)\n", " !gunzip $file" ], "execution_count": 6, "outputs": [ { "output_type": "stream", "text": [ "/content/2020-Jan.csv.gz\n", "/content/2020-Mar.csv.gz\n", "/content/2019-Dec.csv.gz\n", "/content/2020-Apr.csv.gz\n", "/content/2020-Feb.csv.gz\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 479 }, "id": "teiuKEb1QzcQ", "outputId": "c64a375c-41c4-4040-e92a-49e80d875932" }, "source": [ "snapshot = pd.read_csv(\"/content/2019-Oct.csv\").sample(1000)\n", "snapshot.head()" ], "execution_count": 7, "outputs": [ { "output_type": "execute_result", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
event_timeevent_typeproduct_idcategory_idcategory_codebrandpriceuser_iduser_session
197046812019-10-15 12:13:56 UTCview13073482053013558920217191computers.notebookacer694.975152632771f854496-4dae-4114-af89-4f027ba31e56
416614712019-10-31 09:41:21 UTCview127091692053013553559896355NaNmichelin68.21512784291caa01956-1d21-4966-8767-3fa5013ef13a
146364832019-10-12 06:19:13 UTCview113000592053013555531219353electronics.telephonetexet17.46519299950d67e54c3-88e4-4df2-bc37-d8c7cf656a6a
412728372019-10-31 03:29:40 UTCview115002732053013554625249641NaNjbl94.71543868618a162e395-7b69-49aa-9d45-f05a60e19641
35274622019-10-03 20:29:36 UTCview10047772053013555631882655electronics.smartphonexiaomi135.0152083467198d5af9f-c141-4b79-bdd1-dc9bd1d9a7cb
\n", "
" ], "text/plain": [ " event_time ... user_session\n", "19704681 2019-10-15 12:13:56 UTC ... 1f854496-4dae-4114-af89-4f027ba31e56\n", "41661471 2019-10-31 09:41:21 UTC ... caa01956-1d21-4966-8767-3fa5013ef13a\n", "14636483 2019-10-12 06:19:13 UTC ... d67e54c3-88e4-4df2-bc37-d8c7cf656a6a\n", "41272837 2019-10-31 03:29:40 UTC ... a162e395-7b69-49aa-9d45-f05a60e19641\n", "3527462 2019-10-03 20:29:36 UTC ... 98d5af9f-c141-4b79-bdd1-dc9bd1d9a7cb\n", "\n", "[5 rows x 9 columns]" ] }, "metadata": { "tags": [] }, "execution_count": 7 } ] }, { "cell_type": "code", "metadata": { "id": "Jq2EU5X4RAgG" }, "source": [ "!mkdir -p /content/data/tmp\n", "snapshot.to_csv('/content/data/tmp/snapshot.csv', index=False)" ], "execution_count": 8, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "CYgxGyUMHmDQ", "outputId": "72be5f21-5abd-4169-ca86-432920d660f7" }, "source": [ "gc.collect()" ], "execution_count": 9, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "123" ] }, "metadata": { "tags": [] }, "execution_count": 9 } ] }, { "cell_type": "markdown", "metadata": { "id": "CD_eROhoNxEu" }, "source": [ "### Data transform" ] }, { "cell_type": "code", "metadata": { "id": "yTtSL-aeN464" }, "source": [ "import os\n", "import gc\n", "import glob\n", "import pandas as pd\n", "from pathlib import Path" ], "execution_count": 1, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "bD1X1HtbXxqU" }, "source": [ "file = '/content/data/tmp/snapshot.csv'" ], "execution_count": 11, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "LlHEnxnZR056", "colab": { "base_uri": "https://localhost:8080/", "height": 374 }, "outputId": "6db3f1c9-8543-4270-e3e2-4d05defa1923" }, "source": [ "df_tmp = pd.read_csv(file)\n", "df_tmp['session_purchase'] = df_tmp['user_session'] + '_' + df_tmp['product_id'].astype(str)\n", "df_purchase = df_tmp[df_tmp['event_type']=='purchase']\n", "df_cart = df_tmp[df_tmp['event_type']=='cart']\n", "df_purchase = df_purchase[df_purchase['session_purchase'].isin(df_cart['session_purchase'])]\n", "df_cart = df_cart[~(df_cart['session_purchase'].isin(df_purchase['session_purchase']))]\n", "df_cart['target'] = 0\n", "df_purchase['target'] = 1\n", "df = pd.concat([df_cart, df_purchase])\n", "df = df.drop('category_id', axis=1)\n", "df = df.drop('session_purchase', axis=1)\n", "df.head()" ], "execution_count": 12, "outputs": [ { "output_type": "execute_result", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
event_timeevent_typeproduct_idcategory_codebrandpriceuser_iduser_sessiontarget
552019-10-01 14:17:59 UTCcart4804055electronics.audio.headphoneapple189.10522564858c3f614f6-bdcb-46d7-ad93-18d03b0e11ba0
752019-10-10 10:16:18 UTCcart1004708electronics.smartphonehuawei153.96516882359b513996d-0335-4337-bb4d-09c3c14ee2cf0
1332019-10-18 08:22:44 UTCcart2701288appliances.kitchen.refrigeratorssamsung849.265127625672dd77291-d54e-4d3c-a25a-2c915df9fff90
1402019-10-10 04:24:55 UTCcart1004767electronics.smartphonesamsung250.93558658709d6c93da2-f4ee-42a1-951a-45fc52bbbf820
1632019-10-12 20:19:27 UTCcart1004750electronics.smartphonesamsung196.80549588267d63ac7ef-d492-4034-b846-351fd2a5c32b0
\n", "
" ], "text/plain": [ " event_time ... target\n", "55 2019-10-01 14:17:59 UTC ... 0\n", "75 2019-10-10 10:16:18 UTC ... 0\n", "133 2019-10-18 08:22:44 UTC ... 0\n", "140 2019-10-10 04:24:55 UTC ... 0\n", "163 2019-10-12 20:19:27 UTC ... 0\n", "\n", "[5 rows x 9 columns]" ] }, "metadata": { "tags": [] }, "execution_count": 12 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 479 }, "id": "H_rxW2oGRkIh", "outputId": "1830c320-50e1-41cf-bbbf-bfe8887306d6" }, "source": [ "df[['cat_0', 'cat_1', 'cat_2']] = df['category_code'].str.split(\"\\.\", n = 3, expand = True).fillna('NA')\n", "# df[['cat_0', 'cat_1', 'cat_2', 'cat_3']] = df['category_code'].str.split(\"\\.\", n = 3, expand = True).fillna('NA')\n", "df['brand'] = df['brand'].fillna('NA')\n", "df = df.drop('category_code', axis=1)\n", "df['timestamp'] = pd.to_datetime(df['event_time'].str.replace(' UTC', ''))\n", "df['ts_hour'] = df['timestamp'].dt.hour\n", "df['ts_minute'] = df['timestamp'].dt.minute\n", "df['ts_weekday'] = df['timestamp'].dt.weekday\n", "df['ts_day'] = df['timestamp'].dt.day\n", "df['ts_month'] = df['timestamp'].dt.month\n", "df['ts_year'] = df['timestamp'].dt.year\n", "df.head()" ], "execution_count": 13, "outputs": [ { "output_type": "execute_result", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
event_timeevent_typeproduct_idbrandpriceuser_iduser_sessiontargetcat_0cat_1cat_2timestampts_hourts_minutets_weekdayts_dayts_monthts_year
552019-10-01 14:17:59 UTCcart4804055apple189.10522564858c3f614f6-bdcb-46d7-ad93-18d03b0e11ba0electronicsaudioheadphone2019-10-01 14:17:59141711102019
752019-10-10 10:16:18 UTCcart1004708huawei153.96516882359b513996d-0335-4337-bb4d-09c3c14ee2cf0electronicssmartphoneNA2019-10-10 10:16:181016310102019
1332019-10-18 08:22:44 UTCcart2701288samsung849.265127625672dd77291-d54e-4d3c-a25a-2c915df9fff90applianceskitchenrefrigerators2019-10-18 08:22:44822418102019
1402019-10-10 04:24:55 UTCcart1004767samsung250.93558658709d6c93da2-f4ee-42a1-951a-45fc52bbbf820electronicssmartphoneNA2019-10-10 04:24:55424310102019
1632019-10-12 20:19:27 UTCcart1004750samsung196.80549588267d63ac7ef-d492-4034-b846-351fd2a5c32b0electronicssmartphoneNA2019-10-12 20:19:272019512102019
\n", "
" ], "text/plain": [ " event_time event_type product_id ... ts_day ts_month ts_year\n", "55 2019-10-01 14:17:59 UTC cart 4804055 ... 1 10 2019\n", "75 2019-10-10 10:16:18 UTC cart 1004708 ... 10 10 2019\n", "133 2019-10-18 08:22:44 UTC cart 2701288 ... 18 10 2019\n", "140 2019-10-10 04:24:55 UTC cart 1004767 ... 10 10 2019\n", "163 2019-10-12 20:19:27 UTC cart 1004750 ... 12 10 2019\n", "\n", "[5 rows x 18 columns]" ] }, "metadata": { "tags": [] }, "execution_count": 13 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "M95wOkSINI_K", "outputId": "8b637a53-be8f-4c29-a286-d5207cf1b41e" }, "source": [ "list_files = glob.glob('/content/*.csv')\n", "list_files" ], "execution_count": 14, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "['/content/2020-Apr.csv',\n", " '/content/2019-Nov.csv',\n", " '/content/2020-Jan.csv',\n", " '/content/2019-Dec.csv',\n", " '/content/2020-Mar.csv',\n", " '/content/2019-Oct.csv',\n", " '/content/2020-Feb.csv']" ] }, "metadata": { "tags": [] }, "execution_count": 14 } ] }, { "cell_type": "code", "metadata": { "id": "AAVAUULwNriQ" }, "source": [ "def process_files(df_tmp, chunkname):\n", " df_tmp['session_purchase'] = df_tmp['user_session'] + '_' + df_tmp['product_id'].astype(str)\n", " df_purchase = df_tmp[df_tmp['event_type']=='purchase']\n", " df_cart = df_tmp[df_tmp['event_type']=='cart']\n", " df_purchase = df_purchase[df_purchase['session_purchase'].isin(df_cart['session_purchase'])]\n", " df_cart = df_cart[~(df_cart['session_purchase'].isin(df_purchase['session_purchase']))]\n", " df_cart['target'] = 0\n", " df_purchase['target'] = 1\n", " df = pd.concat([df_cart, df_purchase])\n", " df = df.drop('category_id', axis=1)\n", " df = df.drop('session_purchase', axis=1)\n", " # df[['cat_0', 'cat_1', 'cat_2', 'cat_3']] = df['category_code'].str.split(\"\\.\", n = 3, expand = True).fillna('NA')\n", " df['brand'] = df['brand'].fillna('NA')\n", " # df = df.drop('category_code', axis=1)\n", " df['timestamp'] = pd.to_datetime(df['event_time'].str.replace(' UTC', ''))\n", " df['ts_hour'] = df['timestamp'].dt.hour\n", " df['ts_minute'] = df['timestamp'].dt.minute\n", " df['ts_weekday'] = df['timestamp'].dt.weekday\n", " df['ts_day'] = df['timestamp'].dt.day\n", " df['ts_month'] = df['timestamp'].dt.month\n", " df['ts_year'] = df['timestamp'].dt.year\n", " df.to_csv(chunkname, index=False)" ], "execution_count": 15, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "l8_QGBIzWxcO" }, "source": [ "base_path_silver = \"/content/data/silver\"\n", "!mkdir -p $base_path_silver" ], "execution_count": 2, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "rlpfP0TRdsUK", "outputId": "1d002eaf-c8d7-475b-a63b-21a17e6edd1b" }, "source": [ "for idx, chunk in enumerate(list_files[:2]):\n", " chunkname = os.path.join(base_path_silver, Path(chunk).stem + '-' + str(idx) + '.csv')\n", " print(chunkname)" ], "execution_count": 24, "outputs": [ { "output_type": "stream", "text": [ "/content/data/silver/2020-Apr-0.csv\n", "/content/data/silver/2019-Nov-1.csv\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "RiN1K5yxOSik", "outputId": "7ee593d1-0673-444e-8ff2-64ca109a30f7" }, "source": [ "chunksize = 10 ** 6\n", "\n", "for file in list_files:\n", " print(file)\n", " for idx, chunk in enumerate(pd.read_csv(file, chunksize=chunksize)):\n", " chunkname = os.path.join(base_path_silver, Path(file).stem + '-' + str(idx) + '.csv')\n", " print(chunkname)\n", " if not os.path.exists(chunkname):\n", " process_files(chunk, chunkname)" ], "execution_count": 26, "outputs": [ { "output_type": "stream", "text": [ "/content/2020-Apr.csv\n", "/content/data/silver/2020-Apr-0.csv\n", "/content/data/silver/2020-Apr-1.csv\n", "/content/data/silver/2020-Apr-2.csv\n", "/content/data/silver/2020-Apr-3.csv\n", "/content/data/silver/2020-Apr-4.csv\n", "/content/data/silver/2020-Apr-5.csv\n", "/content/data/silver/2020-Apr-6.csv\n", "/content/data/silver/2020-Apr-7.csv\n", "/content/data/silver/2020-Apr-8.csv\n", "/content/data/silver/2020-Apr-9.csv\n", "/content/data/silver/2020-Apr-10.csv\n", "/content/data/silver/2020-Apr-11.csv\n", "/content/data/silver/2020-Apr-12.csv\n", "/content/data/silver/2020-Apr-13.csv\n", "/content/data/silver/2020-Apr-14.csv\n", "/content/data/silver/2020-Apr-15.csv\n", "/content/data/silver/2020-Apr-16.csv\n", "/content/data/silver/2020-Apr-17.csv\n", "/content/data/silver/2020-Apr-18.csv\n", "/content/data/silver/2020-Apr-19.csv\n", "/content/data/silver/2020-Apr-20.csv\n", "/content/data/silver/2020-Apr-21.csv\n", "/content/data/silver/2020-Apr-22.csv\n", "/content/data/silver/2020-Apr-23.csv\n", "/content/data/silver/2020-Apr-24.csv\n", "/content/data/silver/2020-Apr-25.csv\n", "/content/data/silver/2020-Apr-26.csv\n", "/content/data/silver/2020-Apr-27.csv\n", "/content/data/silver/2020-Apr-28.csv\n", "/content/data/silver/2020-Apr-29.csv\n", "/content/data/silver/2020-Apr-30.csv\n", "/content/data/silver/2020-Apr-31.csv\n", "/content/data/silver/2020-Apr-32.csv\n", "/content/data/silver/2020-Apr-33.csv\n", "/content/data/silver/2020-Apr-34.csv\n", "/content/data/silver/2020-Apr-35.csv\n", "/content/data/silver/2020-Apr-36.csv\n", "/content/data/silver/2020-Apr-37.csv\n", "/content/data/silver/2020-Apr-38.csv\n", "/content/data/silver/2020-Apr-39.csv\n", "/content/data/silver/2020-Apr-40.csv\n", "/content/data/silver/2020-Apr-41.csv\n", "/content/data/silver/2020-Apr-42.csv\n", "/content/data/silver/2020-Apr-43.csv\n", "/content/data/silver/2020-Apr-44.csv\n", "/content/data/silver/2020-Apr-45.csv\n", "/content/data/silver/2020-Apr-46.csv\n", "/content/data/silver/2020-Apr-47.csv\n", "/content/data/silver/2020-Apr-48.csv\n", "/content/data/silver/2020-Apr-49.csv\n", "/content/data/silver/2020-Apr-50.csv\n", "/content/data/silver/2020-Apr-51.csv\n", "/content/data/silver/2020-Apr-52.csv\n", "/content/data/silver/2020-Apr-53.csv\n", "/content/data/silver/2020-Apr-54.csv\n", "/content/data/silver/2020-Apr-55.csv\n", "/content/data/silver/2020-Apr-56.csv\n", "/content/data/silver/2020-Apr-57.csv\n", "/content/data/silver/2020-Apr-58.csv\n", "/content/data/silver/2020-Apr-59.csv\n", "/content/data/silver/2020-Apr-60.csv\n", "/content/data/silver/2020-Apr-61.csv\n", "/content/data/silver/2020-Apr-62.csv\n", "/content/data/silver/2020-Apr-63.csv\n", "/content/data/silver/2020-Apr-64.csv\n", "/content/data/silver/2020-Apr-65.csv\n", "/content/data/silver/2020-Apr-66.csv\n", "/content/2019-Nov.csv\n", "/content/data/silver/2019-Nov-0.csv\n", "/content/data/silver/2019-Nov-1.csv\n", "/content/data/silver/2019-Nov-2.csv\n", "/content/data/silver/2019-Nov-3.csv\n", "/content/data/silver/2019-Nov-4.csv\n", "/content/data/silver/2019-Nov-5.csv\n", "/content/data/silver/2019-Nov-6.csv\n", "/content/data/silver/2019-Nov-7.csv\n", "/content/data/silver/2019-Nov-8.csv\n", "/content/data/silver/2019-Nov-9.csv\n", "/content/data/silver/2019-Nov-10.csv\n", "/content/data/silver/2019-Nov-11.csv\n", "/content/data/silver/2019-Nov-12.csv\n", "/content/data/silver/2019-Nov-13.csv\n", "/content/data/silver/2019-Nov-14.csv\n", "/content/data/silver/2019-Nov-15.csv\n", "/content/data/silver/2019-Nov-16.csv\n", "/content/data/silver/2019-Nov-17.csv\n", "/content/data/silver/2019-Nov-18.csv\n", "/content/data/silver/2019-Nov-19.csv\n", "/content/data/silver/2019-Nov-20.csv\n", "/content/data/silver/2019-Nov-21.csv\n", "/content/data/silver/2019-Nov-22.csv\n", "/content/data/silver/2019-Nov-23.csv\n", "/content/data/silver/2019-Nov-24.csv\n", "/content/data/silver/2019-Nov-25.csv\n", "/content/data/silver/2019-Nov-26.csv\n", "/content/data/silver/2019-Nov-27.csv\n", "/content/data/silver/2019-Nov-28.csv\n", "/content/data/silver/2019-Nov-29.csv\n", "/content/data/silver/2019-Nov-30.csv\n", "/content/data/silver/2019-Nov-31.csv\n", "/content/data/silver/2019-Nov-32.csv\n", "/content/data/silver/2019-Nov-33.csv\n", "/content/data/silver/2019-Nov-34.csv\n", "/content/data/silver/2019-Nov-35.csv\n", "/content/data/silver/2019-Nov-36.csv\n", "/content/data/silver/2019-Nov-37.csv\n", "/content/data/silver/2019-Nov-38.csv\n", "/content/data/silver/2019-Nov-39.csv\n", "/content/data/silver/2019-Nov-40.csv\n", "/content/data/silver/2019-Nov-41.csv\n", "/content/data/silver/2019-Nov-42.csv\n", "/content/data/silver/2019-Nov-43.csv\n", "/content/data/silver/2019-Nov-44.csv\n", "/content/data/silver/2019-Nov-45.csv\n", "/content/data/silver/2019-Nov-46.csv\n", "/content/data/silver/2019-Nov-47.csv\n", "/content/data/silver/2019-Nov-48.csv\n", "/content/data/silver/2019-Nov-49.csv\n", "/content/data/silver/2019-Nov-50.csv\n", "/content/data/silver/2019-Nov-51.csv\n", "/content/data/silver/2019-Nov-52.csv\n", "/content/data/silver/2019-Nov-53.csv\n", "/content/data/silver/2019-Nov-54.csv\n", "/content/data/silver/2019-Nov-55.csv\n", "/content/data/silver/2019-Nov-56.csv\n", "/content/data/silver/2019-Nov-57.csv\n", "/content/data/silver/2019-Nov-58.csv\n", "/content/data/silver/2019-Nov-59.csv\n", "/content/data/silver/2019-Nov-60.csv\n", "/content/data/silver/2019-Nov-61.csv\n", "/content/data/silver/2019-Nov-62.csv\n", "/content/data/silver/2019-Nov-63.csv\n", "/content/data/silver/2019-Nov-64.csv\n", "/content/data/silver/2019-Nov-65.csv\n", "/content/data/silver/2019-Nov-66.csv\n", "/content/data/silver/2019-Nov-67.csv\n", "/content/2020-Jan.csv\n", "/content/data/silver/2020-Jan-0.csv\n", "/content/data/silver/2020-Jan-1.csv\n", "/content/data/silver/2020-Jan-2.csv\n", "/content/data/silver/2020-Jan-3.csv\n", "/content/data/silver/2020-Jan-4.csv\n", "/content/data/silver/2020-Jan-5.csv\n", "/content/data/silver/2020-Jan-6.csv\n", "/content/data/silver/2020-Jan-7.csv\n", "/content/data/silver/2020-Jan-8.csv\n", "/content/data/silver/2020-Jan-9.csv\n", "/content/data/silver/2020-Jan-10.csv\n", "/content/data/silver/2020-Jan-11.csv\n", "/content/data/silver/2020-Jan-12.csv\n", "/content/data/silver/2020-Jan-13.csv\n", "/content/data/silver/2020-Jan-14.csv\n", "/content/data/silver/2020-Jan-15.csv\n", "/content/data/silver/2020-Jan-16.csv\n", "/content/data/silver/2020-Jan-17.csv\n", "/content/data/silver/2020-Jan-18.csv\n", "/content/data/silver/2020-Jan-19.csv\n", "/content/data/silver/2020-Jan-20.csv\n", "/content/data/silver/2020-Jan-21.csv\n", "/content/data/silver/2020-Jan-22.csv\n", "/content/data/silver/2020-Jan-23.csv\n", "/content/data/silver/2020-Jan-24.csv\n", "/content/data/silver/2020-Jan-25.csv\n", "/content/data/silver/2020-Jan-26.csv\n", "/content/data/silver/2020-Jan-27.csv\n", "/content/data/silver/2020-Jan-28.csv\n", "/content/data/silver/2020-Jan-29.csv\n", "/content/data/silver/2020-Jan-30.csv\n", "/content/data/silver/2020-Jan-31.csv\n", "/content/data/silver/2020-Jan-32.csv\n", "/content/data/silver/2020-Jan-33.csv\n", "/content/data/silver/2020-Jan-34.csv\n", "/content/data/silver/2020-Jan-35.csv\n", "/content/data/silver/2020-Jan-36.csv\n", "/content/data/silver/2020-Jan-37.csv\n", "/content/data/silver/2020-Jan-38.csv\n", "/content/data/silver/2020-Jan-39.csv\n", "/content/data/silver/2020-Jan-40.csv\n", "/content/data/silver/2020-Jan-41.csv\n", "/content/data/silver/2020-Jan-42.csv\n", "/content/data/silver/2020-Jan-43.csv\n", "/content/data/silver/2020-Jan-44.csv\n", "/content/data/silver/2020-Jan-45.csv\n", "/content/data/silver/2020-Jan-46.csv\n", "/content/data/silver/2020-Jan-47.csv\n", "/content/data/silver/2020-Jan-48.csv\n", "/content/data/silver/2020-Jan-49.csv\n", "/content/data/silver/2020-Jan-50.csv\n", "/content/data/silver/2020-Jan-51.csv\n", "/content/data/silver/2020-Jan-52.csv\n", "/content/data/silver/2020-Jan-53.csv\n", "/content/data/silver/2020-Jan-54.csv\n", "/content/data/silver/2020-Jan-55.csv\n", "/content/2019-Dec.csv\n", "/content/data/silver/2019-Dec-0.csv\n", "/content/data/silver/2019-Dec-1.csv\n", "/content/data/silver/2019-Dec-2.csv\n", "/content/data/silver/2019-Dec-3.csv\n", "/content/data/silver/2019-Dec-4.csv\n", "/content/data/silver/2019-Dec-5.csv\n", "/content/data/silver/2019-Dec-6.csv\n", "/content/data/silver/2019-Dec-7.csv\n", "/content/data/silver/2019-Dec-8.csv\n", "/content/data/silver/2019-Dec-9.csv\n", "/content/data/silver/2019-Dec-10.csv\n", "/content/data/silver/2019-Dec-11.csv\n", "/content/data/silver/2019-Dec-12.csv\n", "/content/data/silver/2019-Dec-13.csv\n", "/content/data/silver/2019-Dec-14.csv\n", "/content/data/silver/2019-Dec-15.csv\n", "/content/data/silver/2019-Dec-16.csv\n", "/content/data/silver/2019-Dec-17.csv\n", "/content/data/silver/2019-Dec-18.csv\n", "/content/data/silver/2019-Dec-19.csv\n", "/content/data/silver/2019-Dec-20.csv\n", "/content/data/silver/2019-Dec-21.csv\n", "/content/data/silver/2019-Dec-22.csv\n", "/content/data/silver/2019-Dec-23.csv\n", "/content/data/silver/2019-Dec-24.csv\n", "/content/data/silver/2019-Dec-25.csv\n", "/content/data/silver/2019-Dec-26.csv\n", "/content/data/silver/2019-Dec-27.csv\n", "/content/data/silver/2019-Dec-28.csv\n", "/content/data/silver/2019-Dec-29.csv\n", "/content/data/silver/2019-Dec-30.csv\n", "/content/data/silver/2019-Dec-31.csv\n", "/content/data/silver/2019-Dec-32.csv\n", "/content/data/silver/2019-Dec-33.csv\n", "/content/data/silver/2019-Dec-34.csv\n", "/content/data/silver/2019-Dec-35.csv\n", "/content/data/silver/2019-Dec-36.csv\n", "/content/data/silver/2019-Dec-37.csv\n", "/content/data/silver/2019-Dec-38.csv\n", "/content/data/silver/2019-Dec-39.csv\n", "/content/data/silver/2019-Dec-40.csv\n", "/content/data/silver/2019-Dec-41.csv\n", "/content/data/silver/2019-Dec-42.csv\n", "/content/data/silver/2019-Dec-43.csv\n", "/content/data/silver/2019-Dec-44.csv\n", "/content/data/silver/2019-Dec-45.csv\n", "/content/data/silver/2019-Dec-46.csv\n", "/content/data/silver/2019-Dec-47.csv\n", "/content/data/silver/2019-Dec-48.csv\n", "/content/data/silver/2019-Dec-49.csv\n", "/content/data/silver/2019-Dec-50.csv\n", "/content/data/silver/2019-Dec-51.csv\n", "/content/data/silver/2019-Dec-52.csv\n", "/content/data/silver/2019-Dec-53.csv\n", "/content/data/silver/2019-Dec-54.csv\n", "/content/data/silver/2019-Dec-55.csv\n", "/content/data/silver/2019-Dec-56.csv\n", "/content/data/silver/2019-Dec-57.csv\n", "/content/data/silver/2019-Dec-58.csv\n", "/content/data/silver/2019-Dec-59.csv\n", "/content/data/silver/2019-Dec-60.csv\n", "/content/data/silver/2019-Dec-61.csv\n", "/content/data/silver/2019-Dec-62.csv\n", "/content/data/silver/2019-Dec-63.csv\n", "/content/data/silver/2019-Dec-64.csv\n", "/content/data/silver/2019-Dec-65.csv\n", "/content/data/silver/2019-Dec-66.csv\n", "/content/data/silver/2019-Dec-67.csv\n", "/content/2020-Mar.csv\n", "/content/data/silver/2020-Mar-0.csv\n", "/content/data/silver/2020-Mar-1.csv\n", "/content/data/silver/2020-Mar-2.csv\n", "/content/data/silver/2020-Mar-3.csv\n", "/content/data/silver/2020-Mar-4.csv\n", "/content/data/silver/2020-Mar-5.csv\n", "/content/data/silver/2020-Mar-6.csv\n", "/content/data/silver/2020-Mar-7.csv\n", "/content/data/silver/2020-Mar-8.csv\n", "/content/data/silver/2020-Mar-9.csv\n", "/content/data/silver/2020-Mar-10.csv\n", "/content/data/silver/2020-Mar-11.csv\n", "/content/data/silver/2020-Mar-12.csv\n", "/content/data/silver/2020-Mar-13.csv\n", "/content/data/silver/2020-Mar-14.csv\n", "/content/data/silver/2020-Mar-15.csv\n", "/content/data/silver/2020-Mar-16.csv\n", "/content/data/silver/2020-Mar-17.csv\n", "/content/data/silver/2020-Mar-18.csv\n", "/content/data/silver/2020-Mar-19.csv\n", "/content/data/silver/2020-Mar-20.csv\n", "/content/data/silver/2020-Mar-21.csv\n", "/content/data/silver/2020-Mar-22.csv\n", "/content/data/silver/2020-Mar-23.csv\n", "/content/data/silver/2020-Mar-24.csv\n", "/content/data/silver/2020-Mar-25.csv\n", "/content/data/silver/2020-Mar-26.csv\n", "/content/data/silver/2020-Mar-27.csv\n", "/content/data/silver/2020-Mar-28.csv\n", "/content/data/silver/2020-Mar-29.csv\n", "/content/data/silver/2020-Mar-30.csv\n", "/content/data/silver/2020-Mar-31.csv\n", "/content/data/silver/2020-Mar-32.csv\n", "/content/data/silver/2020-Mar-33.csv\n", "/content/data/silver/2020-Mar-34.csv\n", "/content/data/silver/2020-Mar-35.csv\n", "/content/data/silver/2020-Mar-36.csv\n", "/content/data/silver/2020-Mar-37.csv\n", "/content/data/silver/2020-Mar-38.csv\n", "/content/data/silver/2020-Mar-39.csv\n", "/content/data/silver/2020-Mar-40.csv\n", "/content/data/silver/2020-Mar-41.csv\n", "/content/data/silver/2020-Mar-42.csv\n", "/content/data/silver/2020-Mar-43.csv\n", "/content/data/silver/2020-Mar-44.csv\n", "/content/data/silver/2020-Mar-45.csv\n", "/content/data/silver/2020-Mar-46.csv\n", "/content/data/silver/2020-Mar-47.csv\n", "/content/data/silver/2020-Mar-48.csv\n", "/content/data/silver/2020-Mar-49.csv\n", "/content/data/silver/2020-Mar-50.csv\n", "/content/data/silver/2020-Mar-51.csv\n", "/content/data/silver/2020-Mar-52.csv\n", "/content/data/silver/2020-Mar-53.csv\n", "/content/data/silver/2020-Mar-54.csv\n", "/content/data/silver/2020-Mar-55.csv\n", "/content/data/silver/2020-Mar-56.csv\n", "/content/2019-Oct.csv\n", "/content/data/silver/2019-Oct-0.csv\n", "/content/data/silver/2019-Oct-1.csv\n", "/content/data/silver/2019-Oct-2.csv\n", "/content/data/silver/2019-Oct-3.csv\n", "/content/data/silver/2019-Oct-4.csv\n", "/content/data/silver/2019-Oct-5.csv\n", "/content/data/silver/2019-Oct-6.csv\n", "/content/data/silver/2019-Oct-7.csv\n", "/content/data/silver/2019-Oct-8.csv\n", "/content/data/silver/2019-Oct-9.csv\n", "/content/data/silver/2019-Oct-10.csv\n", "/content/data/silver/2019-Oct-11.csv\n", "/content/data/silver/2019-Oct-12.csv\n", "/content/data/silver/2019-Oct-13.csv\n", "/content/data/silver/2019-Oct-14.csv\n", "/content/data/silver/2019-Oct-15.csv\n", "/content/data/silver/2019-Oct-16.csv\n", "/content/data/silver/2019-Oct-17.csv\n", "/content/data/silver/2019-Oct-18.csv\n", "/content/data/silver/2019-Oct-19.csv\n", "/content/data/silver/2019-Oct-20.csv\n", "/content/data/silver/2019-Oct-21.csv\n", "/content/data/silver/2019-Oct-22.csv\n", "/content/data/silver/2019-Oct-23.csv\n", "/content/data/silver/2019-Oct-24.csv\n", "/content/data/silver/2019-Oct-25.csv\n", "/content/data/silver/2019-Oct-26.csv\n", "/content/data/silver/2019-Oct-27.csv\n", "/content/data/silver/2019-Oct-28.csv\n", "/content/data/silver/2019-Oct-29.csv\n", "/content/data/silver/2019-Oct-30.csv\n", "/content/data/silver/2019-Oct-31.csv\n", "/content/data/silver/2019-Oct-32.csv\n", "/content/data/silver/2019-Oct-33.csv\n", "/content/data/silver/2019-Oct-34.csv\n", "/content/data/silver/2019-Oct-35.csv\n", "/content/data/silver/2019-Oct-36.csv\n", "/content/data/silver/2019-Oct-37.csv\n", "/content/data/silver/2019-Oct-38.csv\n", "/content/data/silver/2019-Oct-39.csv\n", "/content/data/silver/2019-Oct-40.csv\n", "/content/data/silver/2019-Oct-41.csv\n", "/content/data/silver/2019-Oct-42.csv\n", "/content/2020-Feb.csv\n", "/content/data/silver/2020-Feb-0.csv\n", "/content/data/silver/2020-Feb-1.csv\n", "/content/data/silver/2020-Feb-2.csv\n", "/content/data/silver/2020-Feb-3.csv\n", "/content/data/silver/2020-Feb-4.csv\n", "/content/data/silver/2020-Feb-5.csv\n", "/content/data/silver/2020-Feb-6.csv\n", "/content/data/silver/2020-Feb-7.csv\n", "/content/data/silver/2020-Feb-8.csv\n", "/content/data/silver/2020-Feb-9.csv\n", "/content/data/silver/2020-Feb-10.csv\n", "/content/data/silver/2020-Feb-11.csv\n", "/content/data/silver/2020-Feb-12.csv\n", "/content/data/silver/2020-Feb-13.csv\n", "/content/data/silver/2020-Feb-14.csv\n", "/content/data/silver/2020-Feb-15.csv\n", "/content/data/silver/2020-Feb-16.csv\n", "/content/data/silver/2020-Feb-17.csv\n", "/content/data/silver/2020-Feb-18.csv\n", "/content/data/silver/2020-Feb-19.csv\n", "/content/data/silver/2020-Feb-20.csv\n", "/content/data/silver/2020-Feb-21.csv\n", "/content/data/silver/2020-Feb-22.csv\n", "/content/data/silver/2020-Feb-23.csv\n", "/content/data/silver/2020-Feb-24.csv\n", "/content/data/silver/2020-Feb-25.csv\n", "/content/data/silver/2020-Feb-26.csv\n", "/content/data/silver/2020-Feb-27.csv\n", "/content/data/silver/2020-Feb-28.csv\n", "/content/data/silver/2020-Feb-29.csv\n", "/content/data/silver/2020-Feb-30.csv\n", "/content/data/silver/2020-Feb-31.csv\n", "/content/data/silver/2020-Feb-32.csv\n", "/content/data/silver/2020-Feb-33.csv\n", "/content/data/silver/2020-Feb-34.csv\n", "/content/data/silver/2020-Feb-35.csv\n", "/content/data/silver/2020-Feb-36.csv\n", "/content/data/silver/2020-Feb-37.csv\n", "/content/data/silver/2020-Feb-38.csv\n", "/content/data/silver/2020-Feb-39.csv\n", "/content/data/silver/2020-Feb-40.csv\n", "/content/data/silver/2020-Feb-41.csv\n", "/content/data/silver/2020-Feb-42.csv\n", "/content/data/silver/2020-Feb-43.csv\n", "/content/data/silver/2020-Feb-44.csv\n", "/content/data/silver/2020-Feb-45.csv\n", "/content/data/silver/2020-Feb-46.csv\n", "/content/data/silver/2020-Feb-47.csv\n", "/content/data/silver/2020-Feb-48.csv\n", "/content/data/silver/2020-Feb-49.csv\n", "/content/data/silver/2020-Feb-50.csv\n", "/content/data/silver/2020-Feb-51.csv\n", "/content/data/silver/2020-Feb-52.csv\n", "/content/data/silver/2020-Feb-53.csv\n", "/content/data/silver/2020-Feb-54.csv\n", "/content/data/silver/2020-Feb-55.csv\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "id": "fluni49HqMrd" }, "source": [ "for file in list_files:\n", " !rm $file" ], "execution_count": 27, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "-f6cVkEiqtzh", "outputId": "f7d12876-69c3-4184-ae7d-a5c5936d5ea0" }, "source": [ "list_chunks = glob.glob(os.path.join(base_path_silver,'*.csv'))\n", "list_chunks[:10]" ], "execution_count": 3, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "['/content/data/silver/2020-Jan-40.csv',\n", " '/content/data/silver/2019-Dec-54.csv',\n", " '/content/data/silver/2019-Oct-27.csv',\n", " '/content/data/silver/2019-Nov-12.csv',\n", " '/content/data/silver/2020-Mar-36.csv',\n", " '/content/data/silver/2020-Apr-14.csv',\n", " '/content/data/silver/2020-Feb-40.csv',\n", " '/content/data/silver/2019-Nov-3.csv',\n", " '/content/data/silver/2020-Mar-3.csv',\n", " '/content/data/silver/2020-Apr-39.csv']" ] }, "metadata": { "tags": [] }, "execution_count": 3 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "N5VRhVZsq0-Z", "outputId": "e2a638bd-8a64-4a39-abbc-4e0596289b82" }, "source": [ "!cd $base_path_silver && zip /content/data_silver.zip ./*.csv" ], "execution_count": 31, "outputs": [ { "output_type": "stream", "text": [ " adding: 2019-Dec-0.csv (deflated 76%)\n", " adding: 2019-Dec-10.csv (deflated 76%)\n", " adding: 2019-Dec-11.csv (deflated 76%)\n", " adding: 2019-Dec-12.csv (deflated 76%)\n", " adding: 2019-Dec-13.csv (deflated 76%)\n", " adding: 2019-Dec-14.csv (deflated 76%)\n", " adding: 2019-Dec-15.csv (deflated 76%)\n", " adding: 2019-Dec-16.csv (deflated 77%)\n", " adding: 2019-Dec-17.csv (deflated 76%)\n", " adding: 2019-Dec-18.csv (deflated 76%)\n", " adding: 2019-Dec-19.csv (deflated 76%)\n", " adding: 2019-Dec-1.csv (deflated 76%)\n", " adding: 2019-Dec-20.csv (deflated 76%)\n", " adding: 2019-Dec-21.csv (deflated 76%)\n", " adding: 2019-Dec-22.csv (deflated 76%)\n", " adding: 2019-Dec-23.csv (deflated 77%)\n", " adding: 2019-Dec-24.csv (deflated 76%)\n", " adding: 2019-Dec-25.csv (deflated 77%)\n", " adding: 2019-Dec-26.csv (deflated 76%)\n", " adding: 2019-Dec-27.csv (deflated 77%)\n", " adding: 2019-Dec-28.csv (deflated 77%)\n", " adding: 2019-Dec-29.csv (deflated 76%)\n", " adding: 2019-Dec-2.csv (deflated 76%)\n", " adding: 2019-Dec-30.csv (deflated 76%)\n", " adding: 2019-Dec-31.csv (deflated 77%)\n", " adding: 2019-Dec-32.csv (deflated 76%)\n", " adding: 2019-Dec-33.csv (deflated 76%)\n", " adding: 2019-Dec-34.csv (deflated 77%)\n", " adding: 2019-Dec-35.csv (deflated 76%)\n", " adding: 2019-Dec-36.csv (deflated 76%)\n", " adding: 2019-Dec-37.csv (deflated 76%)\n", " adding: 2019-Dec-38.csv (deflated 76%)\n", " adding: 2019-Dec-39.csv (deflated 76%)\n", " adding: 2019-Dec-3.csv (deflated 76%)\n", " adding: 2019-Dec-40.csv (deflated 76%)\n", " adding: 2019-Dec-41.csv (deflated 76%)\n", " adding: 2019-Dec-42.csv (deflated 76%)\n", " adding: 2019-Dec-43.csv (deflated 76%)\n", " adding: 2019-Dec-44.csv (deflated 76%)\n", " adding: 2019-Dec-45.csv (deflated 76%)\n", " adding: 2019-Dec-46.csv (deflated 76%)\n", " adding: 2019-Dec-47.csv (deflated 76%)\n", " adding: 2019-Dec-48.csv (deflated 76%)\n", " adding: 2019-Dec-49.csv (deflated 76%)\n", " adding: 2019-Dec-4.csv (deflated 76%)\n", " adding: 2019-Dec-50.csv (deflated 76%)\n", " adding: 2019-Dec-51.csv (deflated 76%)\n", " adding: 2019-Dec-52.csv (deflated 76%)\n", " adding: 2019-Dec-53.csv (deflated 76%)\n", " adding: 2019-Dec-54.csv (deflated 76%)\n", " adding: 2019-Dec-55.csv (deflated 76%)\n", " adding: 2019-Dec-56.csv (deflated 76%)\n", " adding: 2019-Dec-57.csv (deflated 77%)\n", " adding: 2019-Dec-58.csv (deflated 76%)\n", " adding: 2019-Dec-59.csv (deflated 77%)\n", " adding: 2019-Dec-5.csv (deflated 76%)\n", " adding: 2019-Dec-60.csv (deflated 76%)\n", " adding: 2019-Dec-61.csv (deflated 77%)\n", " adding: 2019-Dec-62.csv (deflated 77%)\n", " adding: 2019-Dec-63.csv (deflated 76%)\n", " adding: 2019-Dec-64.csv (deflated 77%)\n", " adding: 2019-Dec-65.csv (deflated 77%)\n", " adding: 2019-Dec-66.csv (deflated 77%)\n", " adding: 2019-Dec-67.csv (deflated 78%)\n", " adding: 2019-Dec-6.csv (deflated 76%)\n", " adding: 2019-Dec-7.csv (deflated 76%)\n", " adding: 2019-Dec-8.csv (deflated 76%)\n", " adding: 2019-Dec-9.csv (deflated 76%)\n", " adding: 2019-Nov-0.csv (deflated 77%)\n", " adding: 2019-Nov-10.csv (deflated 77%)\n", " adding: 2019-Nov-11.csv (deflated 76%)\n", " adding: 2019-Nov-12.csv (deflated 76%)\n", " adding: 2019-Nov-13.csv (deflated 75%)\n", " adding: 2019-Nov-14.csv (deflated 76%)\n", " adding: 2019-Nov-15.csv (deflated 75%)\n", " adding: 2019-Nov-16.csv (deflated 76%)\n", " adding: 2019-Nov-17.csv (deflated 75%)\n", " adding: 2019-Nov-18.csv (deflated 76%)\n", " adding: 2019-Nov-19.csv (deflated 75%)\n", " adding: 2019-Nov-1.csv (deflated 77%)\n", " adding: 2019-Nov-20.csv (deflated 76%)\n", " adding: 2019-Nov-21.csv (deflated 75%)\n", " adding: 2019-Nov-22.csv (deflated 76%)\n", " adding: 2019-Nov-23.csv (deflated 75%)\n", " adding: 2019-Nov-24.csv (deflated 76%)\n", " adding: 2019-Nov-25.csv (deflated 78%)\n", " adding: 2019-Nov-26.csv (deflated 76%)\n", " adding: 2019-Nov-27.csv (deflated 76%)\n", " adding: 2019-Nov-28.csv (deflated 76%)\n", " adding: 2019-Nov-29.csv (deflated 77%)\n", " adding: 2019-Nov-2.csv (deflated 78%)\n", " adding: 2019-Nov-30.csv (deflated 77%)\n", " adding: 2019-Nov-31.csv (deflated 76%)\n", " adding: 2019-Nov-32.csv (deflated 77%)\n", " adding: 2019-Nov-33.csv (deflated 77%)\n", " adding: 2019-Nov-34.csv (deflated 76%)\n", " adding: 2019-Nov-35.csv (deflated 75%)\n", " adding: 2019-Nov-36.csv (deflated 75%)\n", " adding: 2019-Nov-37.csv (deflated 77%)\n", " adding: 2019-Nov-38.csv (deflated 76%)\n", " adding: 2019-Nov-39.csv (deflated 76%)\n", " adding: 2019-Nov-3.csv (deflated 77%)\n", " adding: 2019-Nov-40.csv (deflated 75%)\n", " adding: 2019-Nov-41.csv (deflated 75%)\n", " adding: 2019-Nov-42.csv (deflated 75%)\n", " adding: 2019-Nov-43.csv (deflated 75%)\n", " adding: 2019-Nov-44.csv (deflated 75%)\n", " adding: 2019-Nov-45.csv (deflated 76%)\n", " adding: 2019-Nov-46.csv (deflated 76%)\n", " adding: 2019-Nov-47.csv (deflated 75%)\n", " adding: 2019-Nov-48.csv (deflated 76%)\n", " adding: 2019-Nov-49.csv (deflated 75%)\n", " adding: 2019-Nov-4.csv (deflated 77%)\n", " adding: 2019-Nov-50.csv (deflated 76%)\n", " adding: 2019-Nov-51.csv (deflated 76%)\n", " adding: 2019-Nov-52.csv (deflated 76%)\n", " adding: 2019-Nov-53.csv (deflated 76%)\n", " adding: 2019-Nov-54.csv (deflated 76%)\n", " adding: 2019-Nov-55.csv (deflated 76%)\n", " adding: 2019-Nov-56.csv (deflated 76%)\n", " adding: 2019-Nov-57.csv (deflated 75%)\n", " adding: 2019-Nov-58.csv (deflated 76%)\n", " adding: 2019-Nov-59.csv (deflated 76%)\n", " adding: 2019-Nov-5.csv (deflated 78%)\n", " adding: 2019-Nov-60.csv (deflated 76%)\n", " adding: 2019-Nov-61.csv (deflated 76%)\n", " adding: 2019-Nov-62.csv (deflated 76%)\n", " adding: 2019-Nov-63.csv (deflated 76%)\n", " adding: 2019-Nov-64.csv (deflated 76%)\n", " adding: 2019-Nov-65.csv (deflated 76%)\n", " adding: 2019-Nov-66.csv (deflated 76%)\n", " adding: 2019-Nov-67.csv (deflated 76%)\n", " adding: 2019-Nov-6.csv (deflated 77%)\n", " adding: 2019-Nov-7.csv (deflated 77%)\n", " adding: 2019-Nov-8.csv (deflated 77%)\n", " adding: 2019-Nov-9.csv (deflated 77%)\n", " adding: 2019-Oct-0.csv (deflated 78%)\n", " adding: 2019-Oct-10.csv (deflated 78%)\n", " adding: 2019-Oct-11.csv (deflated 78%)\n", " adding: 2019-Oct-12.csv (deflated 78%)\n", " adding: 2019-Oct-13.csv (deflated 78%)\n", " adding: 2019-Oct-14.csv (deflated 78%)\n", " adding: 2019-Oct-15.csv (deflated 78%)\n", " adding: 2019-Oct-16.csv (deflated 78%)\n", " adding: 2019-Oct-17.csv (deflated 77%)\n", " adding: 2019-Oct-18.csv (deflated 77%)\n", " adding: 2019-Oct-19.csv (deflated 78%)\n", " adding: 2019-Oct-1.csv (deflated 78%)\n", " adding: 2019-Oct-20.csv (deflated 77%)\n", " adding: 2019-Oct-21.csv (deflated 77%)\n", " adding: 2019-Oct-22.csv (deflated 77%)\n", " adding: 2019-Oct-23.csv (deflated 77%)\n", " adding: 2019-Oct-24.csv (deflated 77%)\n", " adding: 2019-Oct-25.csv (deflated 77%)\n", " adding: 2019-Oct-26.csv (deflated 77%)\n", " adding: 2019-Oct-27.csv (deflated 77%)\n", " adding: 2019-Oct-28.csv (deflated 77%)\n", " adding: 2019-Oct-29.csv (deflated 77%)\n", " adding: 2019-Oct-2.csv (deflated 78%)\n", " adding: 2019-Oct-30.csv (deflated 77%)\n", " adding: 2019-Oct-31.csv (deflated 77%)\n", " adding: 2019-Oct-32.csv (deflated 77%)\n", " adding: 2019-Oct-33.csv (deflated 77%)\n", " adding: 2019-Oct-34.csv (deflated 77%)\n", " adding: 2019-Oct-35.csv (deflated 77%)\n", " adding: 2019-Oct-36.csv (deflated 77%)\n", " adding: 2019-Oct-37.csv (deflated 77%)\n", " adding: 2019-Oct-38.csv (deflated 78%)\n", " adding: 2019-Oct-39.csv (deflated 78%)\n", " adding: 2019-Oct-3.csv (deflated 78%)\n", " adding: 2019-Oct-40.csv (deflated 77%)\n", " adding: 2019-Oct-41.csv (deflated 77%)\n", " adding: 2019-Oct-42.csv (deflated 77%)\n", " adding: 2019-Oct-4.csv (deflated 78%)\n", " adding: 2019-Oct-5.csv (deflated 78%)\n", " adding: 2019-Oct-6.csv (deflated 78%)\n", " adding: 2019-Oct-7.csv (deflated 78%)\n", " adding: 2019-Oct-8.csv (deflated 78%)\n", " adding: 2019-Oct-9.csv (deflated 78%)\n", " adding: 2020-Apr-0.csv (deflated 77%)\n", " adding: 2020-Apr-10.csv (deflated 78%)\n", " adding: 2020-Apr-11.csv (deflated 77%)\n", " adding: 2020-Apr-12.csv (deflated 77%)\n", " adding: 2020-Apr-13.csv (deflated 77%)\n", " adding: 2020-Apr-14.csv (deflated 77%)\n", " adding: 2020-Apr-15.csv (deflated 77%)\n", " adding: 2020-Apr-16.csv (deflated 77%)\n", " adding: 2020-Apr-17.csv (deflated 77%)\n", " adding: 2020-Apr-18.csv (deflated 77%)\n", " adding: 2020-Apr-19.csv (deflated 77%)\n", " adding: 2020-Apr-1.csv (deflated 77%)\n", " adding: 2020-Apr-20.csv (deflated 77%)\n", " adding: 2020-Apr-21.csv (deflated 77%)\n", " adding: 2020-Apr-22.csv (deflated 78%)\n", " adding: 2020-Apr-23.csv (deflated 77%)\n", " adding: 2020-Apr-24.csv (deflated 77%)\n", " adding: 2020-Apr-25.csv (deflated 77%)\n", " adding: 2020-Apr-26.csv (deflated 77%)\n", " adding: 2020-Apr-27.csv (deflated 76%)\n", " adding: 2020-Apr-28.csv (deflated 77%)\n", " adding: 2020-Apr-29.csv (deflated 77%)\n", " adding: 2020-Apr-2.csv (deflated 77%)\n", " adding: 2020-Apr-30.csv (deflated 76%)\n", " adding: 2020-Apr-31.csv (deflated 76%)\n", " adding: 2020-Apr-32.csv (deflated 77%)\n", " adding: 2020-Apr-33.csv (deflated 77%)\n", " adding: 2020-Apr-34.csv (deflated 77%)\n", " adding: 2020-Apr-35.csv (deflated 77%)\n", " adding: 2020-Apr-36.csv (deflated 77%)\n", " adding: 2020-Apr-37.csv (deflated 77%)\n", " adding: 2020-Apr-38.csv (deflated 76%)\n", " adding: 2020-Apr-39.csv (deflated 76%)\n", " adding: 2020-Apr-3.csv (deflated 77%)\n", " adding: 2020-Apr-40.csv (deflated 76%)\n", " adding: 2020-Apr-41.csv (deflated 76%)\n", " adding: 2020-Apr-42.csv (deflated 76%)\n", " adding: 2020-Apr-43.csv (deflated 75%)\n", " adding: 2020-Apr-44.csv (deflated 75%)\n", " adding: 2020-Apr-45.csv (deflated 75%)\n", " adding: 2020-Apr-46.csv (deflated 75%)\n", " adding: 2020-Apr-47.csv (deflated 75%)\n", " adding: 2020-Apr-48.csv (deflated 75%)\n", " adding: 2020-Apr-49.csv (deflated 75%)\n", " adding: 2020-Apr-4.csv (deflated 77%)\n", " adding: 2020-Apr-50.csv (deflated 75%)\n", " adding: 2020-Apr-51.csv (deflated 75%)\n", " adding: 2020-Apr-52.csv (deflated 75%)\n", " adding: 2020-Apr-53.csv (deflated 75%)\n", " adding: 2020-Apr-54.csv (deflated 75%)\n", " adding: 2020-Apr-55.csv (deflated 75%)\n", " adding: 2020-Apr-56.csv (deflated 75%)\n", " adding: 2020-Apr-57.csv (deflated 75%)\n", " adding: 2020-Apr-58.csv (deflated 75%)\n", " adding: 2020-Apr-59.csv (deflated 75%)\n", " adding: 2020-Apr-5.csv (deflated 77%)\n", " adding: 2020-Apr-60.csv (deflated 75%)\n", " adding: 2020-Apr-61.csv (deflated 75%)\n", " adding: 2020-Apr-62.csv (deflated 75%)\n", " adding: 2020-Apr-63.csv (deflated 75%)\n", " adding: 2020-Apr-64.csv (deflated 75%)\n", " adding: 2020-Apr-65.csv (deflated 75%)\n", " adding: 2020-Apr-66.csv (deflated 75%)\n", " adding: 2020-Apr-6.csv (deflated 77%)\n", " adding: 2020-Apr-7.csv (deflated 77%)\n", " adding: 2020-Apr-8.csv (deflated 77%)\n", " adding: 2020-Apr-9.csv (deflated 77%)\n", " adding: 2020-Feb-0.csv (deflated 78%)\n", " adding: 2020-Feb-10.csv (deflated 77%)\n", " adding: 2020-Feb-11.csv (deflated 77%)\n", " adding: 2020-Feb-12.csv (deflated 78%)\n", " adding: 2020-Feb-13.csv (deflated 77%)\n", " adding: 2020-Feb-14.csv (deflated 78%)\n", " adding: 2020-Feb-15.csv (deflated 77%)\n", " adding: 2020-Feb-16.csv (deflated 78%)\n", " adding: 2020-Feb-17.csv (deflated 80%)\n", " adding: 2020-Feb-18.csv (deflated 82%)\n", " adding: 2020-Feb-19.csv (deflated 82%)\n", " adding: 2020-Feb-1.csv (deflated 79%)\n", " adding: 2020-Feb-20.csv (deflated 82%)\n", " adding: 2020-Feb-21.csv (deflated 82%)\n", " adding: 2020-Feb-22.csv (deflated 82%)\n", " adding: 2020-Feb-23.csv (deflated 82%)\n", " adding: 2020-Feb-24.csv (deflated 82%)\n", " adding: 2020-Feb-25.csv (deflated 82%)\n", " adding: 2020-Feb-26.csv (deflated 82%)\n", " adding: 2020-Feb-27.csv (deflated 82%)\n", " adding: 2020-Feb-28.csv (deflated 82%)\n", " adding: 2020-Feb-29.csv (deflated 84%)\n", " adding: 2020-Feb-2.csv (deflated 78%)\n", " adding: 2020-Feb-30.csv (deflated 82%)\n", " adding: 2020-Feb-31.csv (deflated 82%)\n", " adding: 2020-Feb-32.csv (deflated 82%)\n", " adding: 2020-Feb-33.csv (deflated 82%)\n", " adding: 2020-Feb-34.csv (deflated 82%)\n", " adding: 2020-Feb-35.csv (deflated 79%)\n", " adding: 2020-Feb-36.csv (deflated 77%)\n", " adding: 2020-Feb-37.csv (deflated 77%)\n", " adding: 2020-Feb-38.csv (deflated 77%)\n", " adding: 2020-Feb-39.csv (deflated 77%)\n", " adding: 2020-Feb-3.csv (deflated 78%)\n", " adding: 2020-Feb-40.csv (deflated 77%)\n", " adding: 2020-Feb-41.csv (deflated 77%)\n", " adding: 2020-Feb-42.csv (deflated 77%)\n", " adding: 2020-Feb-43.csv (deflated 77%)\n", " adding: 2020-Feb-44.csv (deflated 77%)\n", " adding: 2020-Feb-45.csv (deflated 77%)\n", " adding: 2020-Feb-46.csv (deflated 77%)\n", " adding: 2020-Feb-47.csv (deflated 77%)\n", " adding: 2020-Feb-48.csv (deflated 77%)\n", " adding: 2020-Feb-49.csv (deflated 77%)\n", " adding: 2020-Feb-4.csv (deflated 78%)\n", " adding: 2020-Feb-50.csv (deflated 77%)\n", " adding: 2020-Feb-51.csv (deflated 76%)\n", " adding: 2020-Feb-52.csv (deflated 77%)\n", " adding: 2020-Feb-53.csv (deflated 77%)\n", " adding: 2020-Feb-54.csv (deflated 77%)\n", " adding: 2020-Feb-55.csv (deflated 77%)\n", " adding: 2020-Feb-5.csv (deflated 78%)\n", " adding: 2020-Feb-6.csv (deflated 78%)\n", " adding: 2020-Feb-7.csv (deflated 78%)\n", " adding: 2020-Feb-8.csv (deflated 78%)\n", " adding: 2020-Feb-9.csv (deflated 78%)\n", " adding: 2020-Jan-0.csv (deflated 78%)\n", " adding: 2020-Jan-10.csv (deflated 76%)\n", " adding: 2020-Jan-11.csv (deflated 76%)\n", " adding: 2020-Jan-12.csv (deflated 76%)\n", " adding: 2020-Jan-13.csv (deflated 76%)\n", " adding: 2020-Jan-14.csv (deflated 76%)\n", " adding: 2020-Jan-15.csv (deflated 76%)\n", " adding: 2020-Jan-16.csv (deflated 76%)\n", " adding: 2020-Jan-17.csv (deflated 76%)\n", " adding: 2020-Jan-18.csv (deflated 77%)\n", " adding: 2020-Jan-19.csv (deflated 76%)\n", " adding: 2020-Jan-1.csv (deflated 78%)\n", " adding: 2020-Jan-20.csv (deflated 76%)\n", " adding: 2020-Jan-21.csv (deflated 76%)\n", " adding: 2020-Jan-22.csv (deflated 76%)\n", " adding: 2020-Jan-23.csv (deflated 76%)\n", " adding: 2020-Jan-24.csv (deflated 76%)\n", " adding: 2020-Jan-25.csv (deflated 76%)\n", " adding: 2020-Jan-26.csv (deflated 76%)\n", " adding: 2020-Jan-27.csv (deflated 77%)\n", " adding: 2020-Jan-28.csv (deflated 77%)\n", " adding: 2020-Jan-29.csv (deflated 77%)\n", " adding: 2020-Jan-2.csv (deflated 79%)\n", " adding: 2020-Jan-30.csv (deflated 77%)\n", " adding: 2020-Jan-31.csv (deflated 77%)\n", " adding: 2020-Jan-32.csv (deflated 77%)\n", " adding: 2020-Jan-33.csv (deflated 77%)\n", " adding: 2020-Jan-34.csv (deflated 77%)\n", " adding: 2020-Jan-35.csv (deflated 77%)\n", " adding: 2020-Jan-36.csv (deflated 77%)\n", " adding: 2020-Jan-37.csv (deflated 77%)\n", " adding: 2020-Jan-38.csv (deflated 77%)\n", " adding: 2020-Jan-39.csv (deflated 77%)\n", " adding: 2020-Jan-3.csv (deflated 78%)\n", " adding: 2020-Jan-40.csv (deflated 77%)\n", " adding: 2020-Jan-41.csv (deflated 76%)\n", " adding: 2020-Jan-42.csv (deflated 76%)\n", " adding: 2020-Jan-43.csv (deflated 76%)\n", " adding: 2020-Jan-44.csv (deflated 76%)\n", " adding: 2020-Jan-45.csv (deflated 76%)\n", " adding: 2020-Jan-46.csv (deflated 76%)\n", " adding: 2020-Jan-47.csv (deflated 76%)\n", " adding: 2020-Jan-48.csv (deflated 76%)\n", " adding: 2020-Jan-49.csv (deflated 77%)\n", " adding: 2020-Jan-4.csv (deflated 77%)\n", " adding: 2020-Jan-50.csv (deflated 76%)\n", " adding: 2020-Jan-51.csv (deflated 76%)\n", " adding: 2020-Jan-52.csv (deflated 78%)\n", " adding: 2020-Jan-53.csv (deflated 80%)\n", " adding: 2020-Jan-54.csv (deflated 79%)\n", " adding: 2020-Jan-55.csv (deflated 79%)\n", " adding: 2020-Jan-5.csv (deflated 76%)\n", " adding: 2020-Jan-6.csv (deflated 76%)\n", " adding: 2020-Jan-7.csv (deflated 76%)\n", " adding: 2020-Jan-8.csv (deflated 76%)\n", " adding: 2020-Jan-9.csv (deflated 76%)\n", " adding: 2020-Mar-0.csv (deflated 77%)\n", " adding: 2020-Mar-10.csv (deflated 76%)\n", " adding: 2020-Mar-11.csv (deflated 76%)\n", " adding: 2020-Mar-12.csv (deflated 76%)\n", " adding: 2020-Mar-13.csv (deflated 77%)\n", " adding: 2020-Mar-14.csv (deflated 77%)\n", " adding: 2020-Mar-15.csv (deflated 77%)\n", " adding: 2020-Mar-16.csv (deflated 77%)\n", " adding: 2020-Mar-17.csv (deflated 77%)\n", " adding: 2020-Mar-18.csv (deflated 77%)\n", " adding: 2020-Mar-19.csv (deflated 77%)\n", " adding: 2020-Mar-1.csv (deflated 77%)\n", " adding: 2020-Mar-20.csv (deflated 76%)\n", " adding: 2020-Mar-21.csv (deflated 77%)\n", " adding: 2020-Mar-22.csv (deflated 76%)\n", " adding: 2020-Mar-23.csv (deflated 77%)\n", " adding: 2020-Mar-24.csv (deflated 76%)\n", " adding: 2020-Mar-25.csv (deflated 76%)\n", " adding: 2020-Mar-26.csv (deflated 76%)\n", " adding: 2020-Mar-27.csv (deflated 77%)\n", " adding: 2020-Mar-28.csv (deflated 76%)\n", " adding: 2020-Mar-29.csv (deflated 76%)\n", " adding: 2020-Mar-2.csv (deflated 77%)\n", " adding: 2020-Mar-30.csv (deflated 76%)\n", " adding: 2020-Mar-31.csv (deflated 76%)\n", " adding: 2020-Mar-32.csv (deflated 76%)\n", " adding: 2020-Mar-33.csv (deflated 76%)\n", " adding: 2020-Mar-34.csv (deflated 77%)\n", " adding: 2020-Mar-35.csv (deflated 76%)\n", " adding: 2020-Mar-36.csv (deflated 76%)\n", " adding: 2020-Mar-37.csv (deflated 76%)\n", " adding: 2020-Mar-38.csv (deflated 77%)\n", " adding: 2020-Mar-39.csv (deflated 76%)\n", " adding: 2020-Mar-3.csv (deflated 77%)\n", " adding: 2020-Mar-40.csv (deflated 76%)\n", " adding: 2020-Mar-41.csv (deflated 77%)\n", " adding: 2020-Mar-42.csv (deflated 77%)\n", " adding: 2020-Mar-43.csv (deflated 76%)\n", " adding: 2020-Mar-44.csv (deflated 77%)\n", " adding: 2020-Mar-45.csv (deflated 77%)\n", " adding: 2020-Mar-46.csv (deflated 77%)\n", " adding: 2020-Mar-47.csv (deflated 77%)\n", " adding: 2020-Mar-48.csv (deflated 77%)\n", " adding: 2020-Mar-49.csv (deflated 77%)\n", " adding: 2020-Mar-4.csv (deflated 76%)\n", " adding: 2020-Mar-50.csv (deflated 77%)\n", " adding: 2020-Mar-51.csv (deflated 77%)\n", " adding: 2020-Mar-52.csv (deflated 77%)\n", " adding: 2020-Mar-53.csv (deflated 77%)\n", " adding: 2020-Mar-54.csv (deflated 77%)\n", " adding: 2020-Mar-55.csv (deflated 77%)\n", " adding: 2020-Mar-56.csv (deflated 77%)\n", " adding: 2020-Mar-5.csv (deflated 76%)\n", " adding: 2020-Mar-6.csv (deflated 77%)\n", " adding: 2020-Mar-7.csv (deflated 76%)\n", " adding: 2020-Mar-8.csv (deflated 76%)\n", " adding: 2020-Mar-9.csv (deflated 76%)\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "id": "vYCsnvmgrpAW" }, "source": [ "!cp data_silver.zip /content/drive/MyDrive/Recommendation" ], "execution_count": 32, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "S4FtC2JHOzpY" }, "source": [ "lp = []\n", "for file in list_chunks:\n", " lp.append(pd.read_csv(file))" ], "execution_count": 4, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "7Bc_EWETuEcP", "outputId": "91a58438-9285-4773-d175-bad2a70d7b72" }, "source": [ "df = pd.concat(lp)\n", "df.shape" ], "execution_count": 5, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(16742775, 16)" ] }, "metadata": { "tags": [] }, "execution_count": 5 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 462 }, "id": "mDU1BYyisccz", "outputId": "a4d2c045-1b99-47d9-ac46-d93e732e8525" }, "source": [ "df.head()" ], "execution_count": 35, "outputs": [ { "output_type": "execute_result", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
event_timeevent_typeproduct_idcategory_codebrandpriceuser_iduser_sessiontargettimestampts_hourts_minutets_weekdayts_dayts_monthts_year
02020-01-23 04:32:33 UTCcart100039953appliances.personal.massagerkivi267.445129131399bebab16-bce8-43cd-ac97-f763ee4664dd02020-01-23 04:32:3343232312020
12020-01-23 04:32:36 UTCcart1004767construction.tools.lightsamsung226.27549154456659387e4-032e-4e71-8b0b-3b90f0538f1402020-01-23 04:32:3643232312020
22020-01-23 04:32:38 UTCcart1003306construction.tools.lightapple614.7958562479841a24f4f-afb8-4342-ae6f-003aa968f82f02020-01-23 04:32:3843232312020
32020-01-23 04:32:38 UTCcart4804055sport.bicycleapple192.0752586080670b5ed9a-4395-481e-9f91-d811ced8a4c302020-01-23 04:32:3843232312020
42020-01-23 04:32:43 UTCcart1005100construction.tools.lightsamsung141.06599724791bc35693d-9c73-457f-a7ef-8ced2f0bb72902020-01-23 04:32:4343232312020
\n", "
" ], "text/plain": [ " event_time event_type product_id ... ts_day ts_month ts_year\n", "0 2020-01-23 04:32:33 UTC cart 100039953 ... 23 1 2020\n", "1 2020-01-23 04:32:36 UTC cart 1004767 ... 23 1 2020\n", "2 2020-01-23 04:32:38 UTC cart 1003306 ... 23 1 2020\n", "3 2020-01-23 04:32:38 UTC cart 4804055 ... 23 1 2020\n", "4 2020-01-23 04:32:43 UTC cart 1005100 ... 23 1 2020\n", "\n", "[5 rows x 16 columns]" ] }, "metadata": { "tags": [] }, "execution_count": 35 } ] }, { "cell_type": "code", "metadata": { "id": "lMDUiiz7sarl" }, "source": [ "# df2 = df['category_code'].str.split(\"\\.\", n=3, expand=True).fillna('NA')\n", "# df2.columns = ['cat_{}'.format(x+1) for x in df2.columns]\n", "# df2.to_parquet('/content/data/silver_l2/df_cat.parquet', index=False)" ], "execution_count": 2, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "YFkf_K_jPEGS" }, "source": [ "df_test = df[df['ts_month']==4]\n", "df_valid = df[df['ts_month']==3]\n", "df_train = df[(df['ts_month']!=3)&(df['ts_month']!=4)]" ], "execution_count": 6, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "OXVudXP-yIJG", "outputId": "14323bda-293f-4d13-e202-5b9f0b4779e4" }, "source": [ "df_train.shape, df_valid.shape, df_test.shape" ], "execution_count": 7, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "((11495245, 16), (2466048, 16), (2781482, 16))" ] }, "metadata": { "tags": [] }, "execution_count": 7 } ] }, { "cell_type": "code", "metadata": { "id": "KBEAURWbxtBa" }, "source": [ "!mkdir -p /content/data/silver_l2\n", "df_train.to_parquet('/content/data/silver_l2/train.parquet', index=False)\n", "df_valid.to_parquet('/content/data/silver_l2/valid.parquet', index=False)\n", "df_test.to_parquet('/content/data/silver_l2/test.parquet', index=False)" ], "execution_count": 8, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "Q8E5moaL3pb6" }, "source": [ "### Data load" ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "pN-Foz-x2-uO", "outputId": "b2e3eed1-23fa-4b3c-ce21-09c3337e0d21" }, "source": [ "!cd /content/data/silver_l2 && zip /content/data_silver_l2.zip ./*.parquet\n", "!cp /content/data_silver_l2.zip /content/drive/MyDrive/Recommendation" ], "execution_count": 9, "outputs": [ { "output_type": "stream", "text": [ " adding: test.parquet (deflated 24%)\n", " adding: train.parquet (deflated 25%)\n", " adding: valid.parquet (deflated 24%)\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "t5g5KwR66gTi", "outputId": "784bda5b-5cb6-49fd-9afa-ded102c66b9e" }, "source": [ "import pandas as pd\n", "\n", "df_train = pd.read_parquet('/content/data/silver_l2/train.parquet')\n", "df_valid = pd.read_parquet('/content/data/silver_l2/valid.parquet')\n", "df_test = pd.read_parquet('/content/data/silver_l2/test.parquet')\n", "\n", "df = pd.concat([df_train, df_valid, df_test],ignore_index=True)\n", "\n", "df.shape" ], "execution_count": 1, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(16742775, 16)" ] }, "metadata": { "tags": [] }, "execution_count": 1 } ] }, { "cell_type": "markdown", "metadata": { "id": "tcmIcf3s3sZo" }, "source": [ "### EDA" ] }, { "cell_type": "code", "metadata": { "id": "L_qjZpHn5izl" }, "source": [ "import IPython\n", "\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "\n", "import warnings\n", "warnings.filterwarnings(\"ignore\")\n", "\n", "%matplotlib inline" ], "execution_count": 2, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "H16gdtbZ3aMb", "outputId": "6c4c7561-2041-43c2-876d-3e152f0a880d" }, "source": [ "df.dtypes" ], "execution_count": 3, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "event_time object\n", "event_type object\n", "product_id int64\n", "category_code object\n", "brand object\n", "price float64\n", "user_id int64\n", "user_session object\n", "target int64\n", "timestamp object\n", "ts_hour int64\n", "ts_minute int64\n", "ts_weekday int64\n", "ts_day int64\n", "ts_month int64\n", "ts_year int64\n", "dtype: object" ] }, "metadata": { "tags": [] }, "execution_count": 3 } ] }, { "cell_type": "code", "metadata": { "id": "LIdlCF-l4wvD" }, "source": [ "df['timestamp'] = pd.to_datetime(df['timestamp'])" ], "execution_count": 4, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "zXcy3O_q40cX", "outputId": "5a67dad2-b1d7-4e34-9aba-5c3662474169" }, "source": [ "df.target.mean()" ], "execution_count": 5, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "0.36775719676099095" ] }, "metadata": { "tags": [] }, "execution_count": 5 } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "v9P5YniL5Bh9", "outputId": "3ad07e88-1560-4303-cafa-756c8a0effb2" }, "source": [ "df['event_type'].value_counts(normalize=True)" ], "execution_count": 6, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "cart 0.632243\n", "purchase 0.367757\n", "Name: event_type, dtype: float64" ] }, "metadata": { "tags": [] }, "execution_count": 6 } ] }, { "cell_type": "markdown", "metadata": { "id": "XxVZV8Vn5HdU" }, "source": [ "Around 37.0% of datapoints are purchases. Lets take a look on the sparsity of the data." ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "uukjSMRT5FQG", "outputId": "8b8ab191-dedf-4b8f-c6f1-d9ad81da7f68" }, "source": [ "print('# of datapoints:' + str(df.shape))\n", "print('# of unique users:' + str(df['user_id'].drop_duplicates().shape))\n", "print('# of unique products:' + str(df['product_id'].drop_duplicates().shape))\n", "print('# of unique sessions:' + str(df['user_session'].drop_duplicates().shape))" ], "execution_count": 7, "outputs": [ { "output_type": "stream", "text": [ "# of datapoints:(16742775, 16)\n", "# of unique users:(3584809,)\n", "# of unique products:(214907,)\n", "# of unique sessions:(10715053,)\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "id": "gIWDc9iF5Ktq" }, "source": [ "def plot_sparse(df, col):\n", " stats = df[[col, 'target']].groupby(col).agg(['count', 'mean', 'sum'])\n", " stats = stats.reset_index()\n", " stats.columns = [col, 'count', 'mean', 'sum']\n", " stats_sort = stats['count'].value_counts().reset_index()\n", " stats_sort = stats_sort.sort_values('index')\n", " plt.figure(figsize=(15,4))\n", " plt.bar(stats_sort['index'].astype(str).values[0:20], stats_sort['count'].values[0:20])\n", " plt.title('Frequency of ' + str(col))\n", " plt.xlabel('Number frequency')\n", " plt.ylabel('Frequency')" ], "execution_count": 8, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 295 }, "id": "ThKVAx3K5bqH", "outputId": "d174827b-6d6a-49be-acd4-e057872a4fba" }, "source": [ "plot_sparse(df, 'product_id')" ], "execution_count": 9, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 295 }, "id": "zlYjAUzq5fD9", "outputId": "765f7fa1-5560-4f40-bbaf-ad0854c08020" }, "source": [ "plot_sparse(df, 'user_id')" ], "execution_count": 10, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 295 }, "id": "8powyDiy50ef", "outputId": "7d6155a4-b4f8-459b-b1c1-6a63b634f28c" }, "source": [ "plot_sparse(df, 'brand')" ], "execution_count": 11, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "markdown", "metadata": { "id": "NgUmBmmS597f" }, "source": [ "We can observe following pattern:\n", "- There are ~45000 products which appear only once in the dataset\n", "- There are 1200000 users which appear only once in the dataset\n", "- There are 350 brands which appears only once in the dataset" ] }, { "cell_type": "markdown", "metadata": { "id": "YN9112LL8mHy" }, "source": [ "The sparsity is important for understanding which features can be better used in a model. Product_id and User_id have many values which appear only once and the model is less able to learn a good patterns from them.\n", "On the other hand, brands has many observations and can be leveraged for prediction." ] }, { "cell_type": "code", "metadata": { "id": "7Y2MwjKF55ES" }, "source": [ "def plot_top20(df, col):\n", " stats = df[[col, 'target']].groupby(col).agg(['count', 'mean', 'sum'])\n", " stats = stats.reset_index()\n", " stats.columns = [col, 'count', 'mean', 'sum']\n", " stats = stats.sort_values('count', ascending=False)\n", " fig, ax1 = plt.subplots(figsize=(15,4))\n", " ax2 = ax1.twinx()\n", " ax1.bar(stats[col].astype(str).values[0:20], stats['count'].values[0:20])\n", " ax1.set_xticklabels(stats[col].astype(str).values[0:20], rotation='vertical')\n", " ax2.plot(stats['mean'].values[0:20], color='red')\n", " ax2.set_ylim(0,1)\n", " ax2.set_ylabel('Mean Target')\n", " ax1.set_ylabel('Frequency')\n", " ax1.set_xlabel(col)\n", " ax1.set_title('Top20 ' + col + 's based on frequency')" ], "execution_count": 12, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 342 }, "id": "QrCM9B2b8tLV", "outputId": "27b7658e-8cfe-4a20-c2d9-b39a827c7741" }, "source": [ "plot_top20(df, 'product_id')" ], "execution_count": 13, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 343 }, "id": "-Daz2Etw8ywW", "outputId": "197a2029-58ce-46b9-bcb3-c336b47d7cb3" }, "source": [ "plot_top20(df, 'user_id')" ], "execution_count": 14, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 331 }, "id": "IfVtFT4V80yK", "outputId": "3a38366e-6cc6-4d40-d751-14538abd4997" }, "source": [ "plot_top20(df, 'brand')" ], "execution_count": 15, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "id": "TeNKLjJU81_Y" }, "source": [ "df['date'] = pd.to_datetime(df['timestamp']).dt.date" ], "execution_count": 3, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 344 }, "id": "yyUDEaxC9fKc", "outputId": "61e2cb92-fc44-4e3c-9711-e029d0a4682d" }, "source": [ "plt.figure(figsize=(15,4))\n", "plt.plot(df[['date', 'target']].groupby('date').target.mean())\n", "plt.ylabel('average mean')\n", "plt.xlabel('date')\n", "plt.xticks(df[['date', 'target']].groupby('date').target.mean().index[::3], rotation='vertical')\n", "print('')" ], "execution_count": 4, "outputs": [ { "output_type": "stream", "text": [ "\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "p2xBxTn2-3KX", "outputId": "03cdc5af-6190-48c9-80b7-c2a5609f0884" }, "source": [ "df[['date', 'target']].groupby('date').target.mean().sort_values().head(20)" ], "execution_count": 5, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "date\n", "2019-11-15 0.000000\n", "2020-01-02 0.000000\n", "2020-04-20 0.000096\n", "2020-04-21 0.000147\n", "2020-01-01 0.063013\n", "2020-02-27 0.114378\n", "2019-11-14 0.135596\n", "2019-11-16 0.147902\n", "2020-01-31 0.161678\n", "2020-02-01 0.203322\n", "2020-01-30 0.223492\n", "2020-02-02 0.258615\n", "2020-02-03 0.270665\n", "2020-01-03 0.282014\n", "2020-02-04 0.295918\n", "2020-02-29 0.300146\n", "2020-02-28 0.300432\n", "2020-02-05 0.310188\n", "2020-02-07 0.317174\n", "2020-02-08 0.318376\n", "Name: target, dtype: float64" ] }, "metadata": { "tags": [] }, "execution_count": 5 } ] }, { "cell_type": "markdown", "metadata": { "id": "LJ7POegh-7rk" }, "source": [ "- We explored the data and saw the different raw features available in the dataset.\n", "- We analzyed basic statistics of the raw features and saw long-tail distribution for categorical features (user, item, brand)\n", "- Some categorical features (categories) have high occurances\n", "- In general, we see that categorical features have variance in the target, which we can leverage to engineer more powerful features" ] } ] }