feature_engineering.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "38grnsCYc3mm"
   },
   "source": [
    "# Mount google drive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 122
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 53299,
     "status": "ok",
     "timestamp": 1568519666948,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "jKlHX-60QraZ",
    "outputId": "dd268880-e36b-4cb7-8394-f160b05f743f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
      "\n",
      "Enter your authorization code:\n",
      "··········\n",
      "Mounted at /content/drive/\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive/')\n",
    "data_path = '/content/drive/My Drive/Colab Notebooks/predict_future_sales/data/prep/'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "GyfAeLzDczmO"
   },
   "source": [
    "# Install libs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 207
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 28567,
     "status": "ok",
     "timestamp": 1568519676621,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "mHBrZF9LtB_Y",
    "outputId": "b0905d6b-87fb-4ecd-8a77-c3db2f4c561d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting sparse_dot_topn\n",
      "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/7d/09/ae06509912675fec6f6f7aa5d84cf92d7e62ab215534b2cb6bfcc1d7c17b/sparse_dot_topn-0.2.5.tar.gz (54kB)\n",
      "\u001b[K     |████████████████████████████████| 61kB 2.5MB/s \n",
      "\u001b[?25hBuilding wheels for collected packages: sparse-dot-topn\n",
      "  Building wheel for sparse-dot-topn (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Created wheel for sparse-dot-topn: filename=sparse_dot_topn-0.2.5-cp36-cp36m-linux_x86_64.whl size=98256 sha256=d2bae483ca8590d05c6ce463fbc750eba6c46967144beaee75280bebb925cdeb\n",
      "  Stored in directory: /root/.cache/pip/wheels/f5/de/86/f644ccba76566c3e0dc0aafa4b4b57164a11760606f69dce22\n",
      "Successfully built sparse-dot-topn\n",
      "Installing collected packages: sparse-dot-topn\n",
      "Successfully installed sparse-dot-topn-0.2.5\n"
     ]
    }
   ],
   "source": [
    "!pip install sparse_dot_topn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bbA2nxDVc-0v"
   },
   "source": [
    "# Import libs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "x8EDotX7Q-mp"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "import datetime\n",
    "from itertools import product\n",
    "import gc\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "import lightgbm as lgb\n",
    "from statsmodels.tsa.seasonal import seasonal_decompose\n",
    "from pytz import timezone\n",
    "import time\n",
    "\n",
    "mpl.rcParams['figure.figsize'] = [8.0, 6.0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "srDJaWJwQdo4"
   },
   "source": [
    "# Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RKLpPXp7QJYH"
   },
   "outputs": [],
   "source": [
    "sales = pd.read_pickle(data_path + 'sales.pkl')\n",
    "test = pd.read_pickle(data_path + 'test.pkl')\n",
    "items = pd.read_pickle(data_path + 'items.pkl')\n",
    "categories = pd.read_pickle(data_path + 'item_categories.pkl')\n",
    "shops = pd.read_pickle(data_path + 'shops.pkl')\n",
    "\n",
    "sales = sales.merge(items[['item_id', 'item_category_id']], on = 'item_id', how = 'inner')\n",
    "\n",
    "index_cols = ['shop_id', 'item_id', 'date_block_num'] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "kj65t-q2ZK7C"
   },
   "source": [
    "# Utile Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4tgGJgDfR6hC"
   },
   "outputs": [],
   "source": [
    "def downcast_dtypes(df):\n",
    "    start_size = df.memory_usage(deep = True).sum() / 1024**2\n",
    "    print('Memory usage: {:.2f} MB'.format(start_size))\n",
    "\n",
    "    float_cols = [c for c in df if df[c].dtype == \"float64\"]\n",
    "    int_cols = [c for c in df if df[c].dtype in [\"int64\", \"int32\"]]\n",
    "    df[float_cols] = df[float_cols].astype(np.float32)\n",
    "    df[int_cols] = df[int_cols].astype(np.int16)\n",
    "    end_size = df.memory_usage(deep = True).sum() / 1024**2\n",
    "    print('New Memory usage: {:.2f} MB'.format(end_size))\n",
    "    return df\n",
    "\n",
    "def create_target_for_attribute(df, attr, target, time_col, aggfunc = np.sum, fill = 0):\n",
    "    target_for_attr = df.pivot_table(index = [attr],\n",
    "                                   values = [target],\n",
    "                                   columns = time_col, \n",
    "                                   aggfunc = aggfunc,\n",
    "                                   fill_value = fill,\n",
    "                                  ).reset_index()\n",
    "  \n",
    "    target_for_attr.columns = target_for_attr.columns.droplevel().map(str)\n",
    "    target_for_attr = target_for_attr.reset_index(drop = True).rename_axis(None, axis = 1)\n",
    "    target_for_attr.columns.values[0] = attr\n",
    "    return target_for_attr\n",
    "\n",
    "def create_target_for_attributes(df, attrs, target, time_col, aggfunc = np.sum, fill = 0):\n",
    "    target_for_attrs = df.pivot_table(index = attrs,\n",
    "                                   values = target, \n",
    "                                   columns = time_col, \n",
    "                                   aggfunc = aggfunc, \n",
    "                                   fill_value = fill,\n",
    "                                  ).reset_index()\n",
    "    target_for_attrs.columns = target_for_attrs.columns.map(str)\n",
    "    target_for_attrs = target_for_attrs.reset_index(drop = True).rename_axis(None, axis = 1)\n",
    "    return target_for_attrs\n",
    "\n",
    "def create_train_set(df, index_cols = index_cols):\n",
    "    grid = []\n",
    "  #curr_shops = df['shop_id'].unique()\n",
    "  #curr_items = df['item_id'].unique()\n",
    "    for month in df['date_block_num'].unique():\n",
    "        curr_shops = df[df['date_block_num'] == month]['shop_id'].unique()\n",
    "        curr_items = df[df['date_block_num'] == month]['item_id'].unique()\n",
    "        grid.append(np.array(list(product(*[curr_shops, curr_items, [month]])), dtype = 'int32'))\n",
    "\n",
    "    grid = pd.DataFrame(np.vstack(grid), columns = index_cols, dtype = np.int32)\n",
    "\n",
    "    gb = df.groupby(index_cols, as_index = False).agg({'item_cnt_day':'sum'})\n",
    "    gb.columns = index_cols + ['target']\n",
    "    all_data = pd.merge(grid, gb, how = 'left', on = index_cols).fillna(0)\n",
    "    all_data.sort_values(['date_block_num','shop_id','item_id'], inplace = True)\n",
    "    all_data.loc[:, 'target'] = all_data['target'].clip(0, 20).astype(np.float32)\n",
    "    print('Sales data shape:', df.shape)\n",
    "    print('Generated Train data shape:', all_data.shape)\n",
    "    del curr_shops\n",
    "    del curr_items\n",
    "    del gb\n",
    "    del grid\n",
    "    gc.collect()\n",
    "    return all_data\n",
    "\n",
    "def generate_name(my_name):\n",
    "    now = datetime.datetime.now(timezone('Europe/Berlin'))\n",
    "    name_id = now.strftime(\"%H_%M_%S__%d_%h_%Y\")\n",
    "    return my_name + '_' + name_id"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "QimFCsvHcMqz"
   },
   "source": [
    "# Construct Training data\n",
    "\n",
    "- Collect shops, items that have been sold in each month form sales dataframe.\n",
    "- Generate all possible triples for (shop_id, item_id, date_block_num)\n",
    "- Group sales dataframe according to (shop_id, item_id, date_block_num) and calculate the sum of target\n",
    "- Merge the generated triples with the grouped dataframe\n",
    "- Sort the result dataframe according to month, shop and item\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 17865,
     "status": "ok",
     "timestamp": 1567874194586,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "TFA8nleJF8so",
    "outputId": "61a57875-715b-4b86-8404-2b250dd5efb2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sales data shape: (2935842, 7)\n",
      "Generated Train data shape: (10809347, 4)\n"
     ]
    }
   ],
   "source": [
    "all_data = create_train_set(sales)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Wtrbmid8khfA"
   },
   "source": [
    "# Baseline Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "G095gJXNklkk"
   },
   "outputs": [],
   "source": [
    "all_data = all_data.merge(items, on = 'item_id', how = 'left')\n",
    "all_data = all_data.merge(shops, on = 'shop_id', how = 'left')\n",
    "all_data = all_data.merge(categories, on = 'item_category_id', how = 'left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "uhWH0px4leMj"
   },
   "outputs": [],
   "source": [
    "features = all_data.columns.values\n",
    "features = features[(features != 'target') & (features != 'date_block_num')]\n",
    "X = all_data.loc[:, features]\n",
    "y = all_data.loc[:, 'target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yfrh4I-ws_M7"
   },
   "outputs": [],
   "source": [
    "X_enc = X.copy()\n",
    "X_enc = X_enc.fillna('Unknow')\n",
    "cat_features = []\n",
    "  for col in X.columns:\n",
    "    if X_enc[col].dtypes == np.object:\n",
    "      cat_features.append(X_enc.columns.get_loc(col))\n",
    "      X_enc.loc[:, col] = LabelEncoder().fit_transform(X_enc[col].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "whSpJ_ikmuF_"
   },
   "outputs": [],
   "source": [
    "lgb_params = {\n",
    "    'feature_fraction': 0.75,\n",
    "    'metric': 'rmse',\n",
    "    'nthread':1, \n",
    "    'min_data_in_leaf': 2**7, \n",
    "    'bagging_fraction': 0.75, \n",
    "    'learning_rate': 0.03, \n",
    "    'objective': 'mse', \n",
    "    'bagging_seed': 2**7, \n",
    "    'num_leaves': 2**7,\n",
    "    'bagging_freq':1,\n",
    "    'verbose':0,\n",
    "}\n",
    "d_train = lgb.Dataset(X_enc, label = y)\n",
    "gb_model = lgb.train(lgb_params, d_train, 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 369
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1053,
     "status": "ok",
     "timestamp": 1567190420239,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "ANym-0vDntfs",
    "outputId": "f33ed4be-1614-4172-9290-fb5b636a17d8"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFgCAYAAACmDI9oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_importance = pd.DataFrame(sorted(zip(gb_model.feature_importance(importance_type = 'gain'), X_enc.columns)), columns=['Value','Feature'])\n",
    "plt.figure(figsize = (10, 5))\n",
    "sns.barplot(x = \"Value\", y = \"Feature\", data = feature_importance.sort_values(by = \"Value\", ascending = False))\n",
    "plt.title('LightGBM Features')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "xNE4J3XQah7f"
   },
   "source": [
    "# Global Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 80421,
     "status": "ok",
     "timestamp": 1568519932281,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "pqUk6ihnSQiV",
    "outputId": "298fa732-a99f-4b81-d32b-6570621f687f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "\n",
    "items_sales = create_target_for_attribute(sales, 'item_id', 'item_cnt_day', 'date_block_num', aggfunc = np.sum, fill = 0)\n",
    "items_trans = create_target_for_attribute(sales, 'item_id', 'item_cnt_day', 'date_block_num', aggfunc = np.count_nonzero, fill = 0)\n",
    "items_prices = create_target_for_attribute(sales, 'item_id', 'item_price', 'date_block_num', aggfunc = np.median, fill = np.nan)\n",
    "\n",
    "shops_sales = create_target_for_attribute(sales, 'shop_id', 'item_cnt_day', 'date_block_num', aggfunc = np.sum, fill = 0)\n",
    "shops_trans = create_target_for_attribute(sales, 'shop_id', 'item_cnt_day', 'date_block_num', aggfunc = np.count_nonzero, fill = 0)\n",
    "#shops_prices = create_target_for_attribute(sales, 'shop_id', 'item_price', 'date_block_num', aggfunc = np.median, fill  = np.nan)\n",
    "\n",
    "categories_sales = create_target_for_attribute(sales, 'item_category_id', 'item_cnt_day', 'date_block_num', aggfunc = np.sum, fill = 0)\n",
    "categories_trans = create_target_for_attribute(sales, 'item_category_id', 'item_cnt_day', 'date_block_num', aggfunc = np.count_nonzero, fill = 0)\n",
    "#categories_prices = create_target_for_attribute(sales, 'item_category_id', 'item_price', 'date_block_num', aggfunc = np.median, fill = np.nan)\n",
    "\n",
    "shops_items_sales = create_target_for_attributes(sales, ['shop_id', 'item_id'], 'item_cnt_day', 'date_block_num', aggfunc = np.sum, fill = 0)\n",
    "shops_items_trans = create_target_for_attributes(sales, ['shop_id', 'item_id'], 'item_cnt_day', 'date_block_num', aggfunc = np.count_nonzero, fill = 0)\n",
    "shops_items_sales.loc[:, '0':] = shops_items_sales.loc[:, '0':].clip(0, 20)\n",
    "\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "W38KOIR2HRvA"
   },
   "source": [
    "# Feature Engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "QGi3AfPEJd-1"
   },
   "source": [
    "Before we go crazy generating features, let us look again to the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 102
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1066,
     "status": "ok",
     "timestamp": 1568519937954,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "R-ds-FprLilQ",
    "outputId": "84c8a19f-aa0a-4237-d32c-70aa0aa0a519"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There is sales history for: 4737  items in test set out of 5100\n",
      "There is No sales history for: 363  items in test set out of 5100\n",
      "---------------------\n",
      "There is sales history for: 42  shops in test set out of 42\n",
      "There is No sales history for: 0  shops in test set out of 42\n"
     ]
    }
   ],
   "source": [
    "test_items = set(test.item_id)\n",
    "sales_items = set(sales.item_id)\n",
    "items_in_test_and_sales = test_items.intersection(sales_items)\n",
    "items_in_test_not_sales = test_items - sales_items.intersection(test_items)\n",
    "\n",
    "print('There is sales history for:', len(items_in_test_and_sales) , ' items in test set out of', test['item_id'].nunique())\n",
    "print('There is No sales history for:', len(items_in_test_not_sales), ' items in test set out of', test['item_id'].nunique())\n",
    "print('---------------------')\n",
    "shops_test = set(test.shop_id)\n",
    "shops_train = set(sales.shop_id)\n",
    "shops_in_test_and_sales = shops_test.intersection(shops_train)\n",
    "shops_in_test_not_sales = set(test.shop_id) - shops_test.intersection(shops_train)\n",
    "print('There is sales history for:', len(shops_in_test_and_sales), ' shops in test set out of', test['shop_id'].nunique())\n",
    "print('There is No sales history for:', len(shops_in_test_not_sales), ' shops in test set out of', test['shop_id'].nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 119
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2059,
     "status": "ok",
     "timestamp": 1568519954432,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "fT7ufTzVX_7O",
    "outputId": "fd6a6226-0279-432d-a18d-4fd5c0b47ed3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There is sales history for: 111543  shop/item pairs in test set out of 214200\n",
      "There is No sales history for: 15246  shop/item in test set of 214200\n",
      "There is sales history for: 87411  items ONLY from shop/item pairs in test set out of 214200\n",
      "-------------------\n",
      "Total number of shop/item pairs in train set:  416197\n",
      "Total number of Possible shop/item pairs:  1219350\n"
     ]
    }
   ],
   "source": [
    "ntest = test.reset_index()\n",
    "found_sales = ntest.merge(sales, on = ['item_id','shop_id'], how = 'left').dropna()\n",
    "pairs_with_history = ntest[ntest['ID'].isin(found_sales['ID'])]\n",
    "pairs_with_no_item_history = ntest[~(ntest['item_id'].isin(sales['item_id']))]\n",
    "pairs_with_just_item_history_num = test.shape[0] - (len(pairs_with_history) + len(pairs_with_no_item_history))\n",
    "\n",
    "print('There is sales history for:', len(pairs_with_history) , ' shop/item pairs in test set out of', len(test))\n",
    "print('There is No sales history for:',  len(pairs_with_no_item_history), ' shop/item in test set of', len(test))\n",
    "print('There is sales history for:',  pairs_with_just_item_history_num, ' items ONLY from shop/item pairs in test set out of', len(test))\n",
    "print('-------------------')\n",
    "print('Total number of shop/item pairs in train set: ', len(shops_items_sales))\n",
    "print('Total number of Possible shop/item pairs: ', len(items) * len(shops))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "eNND7iV0jRqx"
   },
   "source": [
    "## Summerize"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yx2nrPO6pvIp"
   },
   "source": [
    "We have **214200** pairs of shop/item that we want to predict their future sales. Those pairs can be divided into 3 groups:\n",
    "- **First group:** 111543 pairs with shop/item sales history. We know the previous sales of the item inside the shop\n",
    "- **Second group:** 87411 pairs with item sales history. We know the previous sales of the items in different shops\n",
    "- **Third group:** 15246 pairs have shop history. We know the previous sales of the shop only. The items have no sales history\n",
    "\n",
    "The problem is modeled as: \n",
    "\n",
    "\n",
    "target_pair1(t = 0), target_pair1(t = 1), target_pair1(t = 2), ..., target_pair1(t = k) >>--->>predict>>--->> **target_pair1(t = 34)**.\n",
    "\n",
    "target_pair2(t = 0), target_pair2(t = 1), target_pair2(t = 2), ..., target_pair2(t = k) >>--->>predict>>--->> **target_pair2(t = 34)**\n",
    "\n",
    "- \n",
    "- \n",
    "- \n",
    "\n",
    "target_pair_m(t = 0), target_pair_m(t = 1), target_pair_m(t = 2), ..., target_pair_m(t = k) >>--->>predict>>--->> **target_pair_m(t = 34)**\n",
    "\n",
    "\n",
    "**First Group:**\n",
    "- lag features for target per (item, shop, item/shop)\n",
    "- trend and seasonality\n",
    "\n",
    "**Second Group:**\n",
    "- lag features for target per(item, shop)\n",
    "- trend and seasonality\n",
    "\n",
    "**Third Group:**\n",
    "- lag features for target per similar items in the same shop\n",
    "- trend and seasonality\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "-JvnlD9XmZve"
   },
   "source": [
    "## Outdated shops/items/categories"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-9-p-JbQFGx3"
   },
   "outputs": [],
   "source": [
    "outdated_limit = '27'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1792,
     "status": "ok",
     "timestamp": 1568520137785,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "8o7QZxkSIkb4",
    "outputId": "8784a0ad-31a3-4573-cdea-2ae3c944782e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shops in train set:  55  out of 55\n",
      "Outdated shops in train set:  12.73 %\n"
     ]
    }
   ],
   "source": [
    "outdated_shops = shops_sales[shops_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('shops in train set: ', len(shops_sales), ' out of', len(shops))\n",
    "print('Outdated shops in train set: ', round(100 * len(outdated_shops) / len(shops_sales), 2), '%')\n",
    "\n",
    "shops['shop_status'] = 'unknown'\n",
    "shops.loc[shops['shop_id'].isin(outdated_shops['shop_id']), 'shop_status'] = 'deactive'\n",
    "\n",
    "active_shops =  shops_sales[shops_sales.loc[:, outdated_limit:].sum(axis = 1) > 0]\n",
    "shops.loc[shops['shop_id'].isin(active_shops['shop_id']), 'shop_status'] = 'active'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1092,
     "status": "ok",
     "timestamp": 1568520138084,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "AMtqcxMJmEny",
    "outputId": "a600389c-6f6c-4b69-9cc6-7eda08ccc57a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "items in train set:  21806  out of 22170\n",
      "Outdated items in train set:  56.84 %\n"
     ]
    }
   ],
   "source": [
    "outdated_items = items_sales[items_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('items in train set: ', len(items_sales), ' out of', len(items))\n",
    "print('Outdated items in train set: ', round(100 * len(outdated_items) /len(items_sales), 2), '%')\n",
    "\n",
    "items['item_status'] = 'unknown'\n",
    "items.loc[items['item_id'].isin(outdated_items['item_id']), 'item_status'] = 'deactive'\n",
    "\n",
    "active_items =  items_sales[items_sales.loc[:,outdated_limit:].sum(axis = 1) > 0]\n",
    "items.loc[items['item_id'].isin(active_items['item_id']), 'item_status'] = 'active'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1317,
     "status": "ok",
     "timestamp": 1568520152047,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "vpH20q6ymQ72",
    "outputId": "dc6b2ba3-6f24-4564-a103-12376ad8e78b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "categories in train set:  84  out of 84\n",
      "Outdated categories in train set:  20.24 %\n"
     ]
    }
   ],
   "source": [
    "outdated_categories = categories_sales[categories_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('categories in train set: ', len(categories_sales), ' out of', len(categories))\n",
    "print('Outdated categories in train set: ',  round(100 * len(outdated_categories)/len(categories_sales), 2), '%')\n",
    "\n",
    "categories['item_category_status'] = 'unknown'\n",
    "categories.loc[categories['item_category_id'].isin(outdated_categories['item_category_id']), 'item_category_status'] = 'deactive'\n",
    "\n",
    "active_categories = categories_sales[categories_sales.loc[:, outdated_limit:].sum(axis = 1) > 0]\n",
    "categories.loc[categories['item_category_id'].isin(active_categories['item_category_id']), 'item_category_status'] = 'active'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "2BYo1fwQZDDH"
   },
   "source": [
    "## Generate history for missing item"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "fj0p52wmZJTW"
   },
   "source": [
    "There is 363 items that we have no sales history at all. It could be that those items are outdated and nobody bought them during the 34 months. Or it could be the orgnizers of the competion took them form the train set.\n",
    "- I will assume that those items have the same sales history like items with same type and category in same shop.\n",
    "- I will give the rest the median of the sales of items in the same category and shop\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "mpBT5pUyVEdX"
   },
   "source": [
    "### Find similar items"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "DrLo1sCondKy"
   },
   "outputs": [],
   "source": [
    "def ngrams(string, n=3):\n",
    "    ngrams = zip(*[string[i:] for i in range(n)])\n",
    "    return [''.join(ngram) for ngram in ngrams]\n",
    "\n",
    "from scipy.sparse import csr_matrix\n",
    "import sparse_dot_topn.sparse_dot_topn as ct\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "def awesome_cossim_top(A, B, ntop, lower_bound=0):\n",
    "    A = A.tocsr()\n",
    "    B = B.tocsr()\n",
    "    M, _ = A.shape\n",
    "    _, N = B.shape\n",
    " \n",
    "    idx_dtype = np.int32\n",
    " \n",
    "    nnz_max = M*ntop\n",
    " \n",
    "    indptr = np.zeros(M+1, dtype=idx_dtype)\n",
    "    indices = np.zeros(nnz_max, dtype=idx_dtype)\n",
    "    data = np.zeros(nnz_max, dtype=A.dtype)\n",
    "\n",
    "    ct.sparse_dot_topn(\n",
    "        M, N, np.asarray(A.indptr, dtype=idx_dtype),\n",
    "        np.asarray(A.indices, dtype=idx_dtype),\n",
    "        A.data,\n",
    "        np.asarray(B.indptr, dtype=idx_dtype),\n",
    "        np.asarray(B.indices, dtype=idx_dtype),\n",
    "        B.data,\n",
    "        ntop,\n",
    "        lower_bound,\n",
    "        indptr, indices, data)\n",
    "\n",
    "    return csr_matrix((data,indices,indptr),shape=(M,N))\n",
    "  \n",
    "def get_matches_df(sparse_matrix, name_vector, top=100):\n",
    "    non_zeros = sparse_matrix.nonzero()\n",
    "    \n",
    "    sparserows = non_zeros[0]\n",
    "    sparsecols = non_zeros[1]\n",
    "    \n",
    "    if top:\n",
    "        nr_matches = top\n",
    "    else:\n",
    "        nr_matches = sparsecols.size\n",
    "    \n",
    "    left_side = np.empty([nr_matches], dtype=object)\n",
    "    right_side = np.empty([nr_matches], dtype=object)\n",
    "    similairity = np.zeros(nr_matches)\n",
    "    \n",
    "    for index in range(0, nr_matches):\n",
    "        left_side[index] = name_vector[sparserows[index]]\n",
    "        right_side[index] = name_vector[sparsecols[index]]\n",
    "        similairity[index] = sparse_matrix.data[index]\n",
    "        \n",
    "    df = pd.DataFrame({'left_side': left_side,\n",
    "                          'right_side': right_side,\n",
    "                           'similairity': similairity})\n",
    "    \n",
    "    return df.sort_values(['similairity'], ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 13532,
     "status": "ok",
     "timestamp": 1568520217835,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "z1uYF5tir6wg",
    "outputId": "9417e984-4e77-4052-c561-91de87b99401"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "item_name tfid shape:  (22170, 16714)\n",
      "Matches between item_names:  (266236, 3)\n"
     ]
    }
   ],
   "source": [
    "items_names = items['item_name']\n",
    "vectorizer = TfidfVectorizer(min_df = 1, analyzer=ngrams)\n",
    "tf_idf_matrix = vectorizer.fit_transform(items_names)\n",
    "print('item_name tfid shape: ', tf_idf_matrix.shape)\n",
    "matches = awesome_cossim_top(tf_idf_matrix, tf_idf_matrix.transpose(), 22170, 0.4)\n",
    "\n",
    "matches_df = get_matches_df(matches, items_names, top = 0)\n",
    "matches_df = matches_df[matches_df['similairity'] < 0.99999]\n",
    "print('Matches between item_names: ', matches_df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 18700,
     "status": "ok",
     "timestamp": 1568520261251,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "U2F3sScDtSUh",
    "outputId": "425ce6c1-1309-4537-cae2-232e2294d1e9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "72 --> 72\n",
      "20486 фигурка game of thrones jon snow legacy action 15см --> 20485 фигурка game of thrones jaime lannister legacy collection 6\n",
      "========\n",
      "20 --> 20\n",
      "6169 rock band 4 игра гитара барабаны и микрофон ps4 английская версия --> 3756 guitar hero live гитара игра ps4 английская версия\n",
      "========\n",
      "75 --> 75\n",
      "4156 kaspersky endpoint security для бизнеса стандартный russian edition 15 19 node 1 year renewal license --> 4155 kaspersky endpoint security для бизнеса стандартный russian edition 10 14 node 1 year renewal license\n",
      "========\n",
      "40 --> 40\n",
      "12409 карнавал rem --> 12406 карнавал регион\n",
      "========\n",
      "40 --> 40\n",
      "18560 свой среди чужих чужой среди своих rem --> 18561 свой среди чужих чужой среди своих регион\n",
      "========\n",
      "40 --> 40\n",
      "10372 гений регион --> 9864 вий регион\n",
      "========\n",
      "44 --> 44\n",
      "168 1с аудиокниги аудиокурсы по географии 6 класс pc цифровая версия --> 169 1с аудиокниги аудиокурсы по географии 7 класс pc цифровая версия\n",
      "========\n",
      "45 --> 45\n",
      "173 1с аудиокниги аудиоэнциклопедия для самых маленьких окружающий мир digipack --> 380 1с аудиокниги спокойной ночи сказки для самых маленьких\n",
      "========\n",
      "55 --> 55\n",
      "14521 меладзе валерий и константин полста 2cd фирм --> 14519 меладзе валерий океан\n",
      "========\n",
      "55 --> 55\n",
      "6335 stewart rod another country --> 6338 stewart rod time\n",
      "========\n",
      "44 --> 44\n",
      "204 1с аудиокниги г р хаггард клеопатра pc цифровая версия --> 208 1с аудиокниги генри хаггард дочь монтесумы pc цифровая версия\n",
      "========\n",
      "58 --> 58\n",
      "4310 led zeppelin coda original recording remastered lp --> 4312 led zeppelin i original recording remastered lp\n",
      "========\n",
      "45 --> 45\n",
      "249 1с аудиокниги жюль верн пятнадцатилетний капитан jewel --> 248 1с аудиокниги жюль верн дети капитана гранта jewel\n",
      "========\n",
      "19 --> 19\n",
      "2322 call of duty black ops iii ps3 русская версия --> 2307 call of duty black ops ii ps3 русская версия\n",
      "========\n",
      "23 --> 23\n",
      "2324 call of duty black ops iii xbox 360 русская версия --> 2309 call of duty black ops ii xbox 360 русская версия\n",
      "========\n",
      "40 --> 40\n",
      "16675 отелло rem --> 21765 чучело rem\n",
      "========\n",
      "40 --> 40\n",
      "12591 клуб самоубийц или приключения титулованной особы rem --> 12590 клуб самоубийц или приключения титулованной особы регион\n",
      "========\n",
      "55 --> 55\n",
      "4412 limp bizkit greatest hitz --> 4411 limp bizkit gold cobra\n",
      "========\n",
      "42 --> 42\n",
      "8549 артбук мир игры rise of the tomb raider --> 8550 артбук мир игры the last of us\n",
      "========\n",
      "42 --> 42\n",
      "8551 артбук мир игры total war --> 8550 артбук мир игры the last of us\n",
      "========\n",
      "31 --> 31\n",
      "2434 crusader kings 2 dlc collection pc цифровая версия --> 2433 crusader kings 2 pc цифровая версия\n",
      "========\n",
      "72 --> 72\n",
      "20880 фигурка walking dead rv walker zombie pop tv 10см --> 20864 фигурка the walking dead series 1 zombie lurker 6\n",
      "========\n",
      "44 --> 44\n",
      "402 1с аудиокниги чарская л а сказки голубой феи pc цифровая версия --> 148 1с аудиокниги английские сказки часть 1 pc цифровая версия\n",
      "========\n",
      "58 --> 58\n",
      "14760 моральный кодекс гибкий стан 2lp --> 14761 моральный кодекс зима 2lp\n",
      "========\n",
      "78 --> 78\n",
      "16811 онлайн доступ к мат по ист 5 11 кл эконом 9 11 и обществозн 10 11 1 год pc цифровая версия --> 16812 онлайн доступ к мат по ист 5 11 кл эконом 9 11 и обществозн 10 11 3 мес pc цифровая версия\n",
      "========\n",
      "58 --> 58\n",
      "14765 моральный кодекс сотрясение мозга 2lp --> 14761 моральный кодекс зима 2lp\n",
      "========\n",
      "45 --> 45\n",
      "430 1с аудиотеатр коллекция сказок а усачева digipack --> 431 1с аудиотеатр коллекция сказок в гауфа digipack\n",
      "========\n",
      "78 --> 78\n",
      "16813 онлайн доступ к матер по биологии для 5 11 кл и химии для 8 9 кл на 1 год pc цифровая версия --> 16814 онлайн доступ к матер по биологии для 5 11 кл и химии для 8 9 кл на 3 мес pc цифровая версия\n",
      "========\n",
      "40 --> 40\n",
      "18913 снайпер регион --> 18910 снайпер\n",
      "========\n",
      "40 --> 40\n",
      "8676 бармен --> 19675 титаник кармен\n",
      "========\n",
      "76 --> 76\n",
      "6645 softorbits photo retoucher домашняя обработка фото esd цифровая версия --> 6648 softorbits пакетная автоматическая обработка фотографий rus цифровая версия\n",
      "========\n",
      "61 --> 61\n",
      "20996 футболка assassin s creed syndicate bronze crest l --> 20998 футболка assassin s creed syndicate bronze crest s\n",
      "========\n",
      "72 --> 72\n",
      "20863 фигурка the simpsons series 4 marge simpson working woman marge 5 --> 20862 фигурка the simpsons series 4 homer simpson grunge homer 5\n",
      "========\n",
      "55 --> 55\n",
      "10779 ддт легенды русского рока 1 фирм --> 10780 ддт легенды русского рока 2 фирм\n",
      "========\n",
      "55 --> 55\n",
      "4642 minogue kylie kylie christmas --> 4643 minogue kylie the best of\n",
      "========\n",
      "40 --> 40\n",
      "18979 соломенная шляпка rem --> 18980 соломенная шляпка регион\n",
      "========\n",
      "77 --> 77\n",
      "566 1с образовательная коллекция веселые правила дорожного движения развивающие игры для детей jewel --> 17704 правила дорожного движения 2011 jewel\n",
      "========\n",
      "40 --> 40\n",
      "8760 белорусский вокзал rem --> 8761 белорусский вокзал регион\n",
      "========\n",
      "40 --> 40\n",
      "16976 переправа часть 2 --> 16975 переправа часть 1\n",
      "========\n",
      "55 --> 55\n",
      "12890 крематорий легенды русского рока 1 фирм --> 12891 крематорий легенды русского рока 2 фирм\n",
      "========\n",
      "47 --> 47\n",
      "14959 манга врата штейна том 3 --> 14957 манга врата штейна том 1\n",
      "========\n",
      "77 --> 77\n",
      "640 1с образовательная коллекция профессор хиггинс русский без акцента версия 6 5 jewel --> 638 1с образовательная коллекция профессор хиггинс английский без акцента версия 6 5 jewel\n",
      "========\n",
      "40 --> 40\n",
      "17030 пиковая дама rem --> 22158 яма rem\n",
      "========\n",
      "40 --> 40\n",
      "19085 старики разбойники rem --> 19086 старики разбойники регион\n",
      "========\n",
      "31 --> 31\n",
      "2722 darksiders franchise pack pc цифровая версия --> 2717 darksiders ii pc цифровая версия\n",
      "========\n",
      "55 --> 55\n",
      "8890 билан дима не молчи --> 8889 билан дима мечтатель\n",
      "========\n",
      "49 --> 49\n",
      "15034 метод материалы 1с зарплата и управление персоналом 8 редакция 3 практика применения в 2015 году --> 15035 метод материалы 1с зарплата и управление персоналом 8 редакция 3 практика применения\n",
      "========\n",
      "61 --> 61\n",
      "21196 футболка payday 2 wolf mask l --> 21197 футболка payday 2 wolf mask m\n",
      "========\n",
      "40 --> 40\n",
      "19155 стюарт литтл регион --> 8904 биттл джюс регион\n",
      "========\n",
      "40 --> 40\n",
      "17117 плохой хороший человек rem --> 21413 хороший плохой злой\n",
      "========\n",
      "61 --> 61\n",
      "21214 футболка starcraft 2 marine m --> 21225 футболка starcraft ii logo черная m\n",
      "========\n",
      "55 --> 55\n",
      "6891 turunen tarja ave maria en plein air --> 6890 turunen tarja act 1 2cd\n",
      "========\n",
      "20 --> 20\n",
      "6903 tales of zestiria ps4 русские субтитры --> 4721 mad max ps4 русские субтитры\n",
      "========\n",
      "40 --> 40\n",
      "19210 супермен 2 --> 9163 бэтмен и супермен м ф регион\n",
      "========\n",
      "36 --> 36\n",
      "13096 карта оплаты windows 750 рублей цифровая версия --> 13093 карта оплаты windows 250 рублей цифровая версия\n",
      "========\n",
      "75 --> 75\n",
      "4925 movavi screen capture 5 box --> 4922 movavi screen capture\n",
      "========\n",
      "78 --> 78\n",
      "15183 модные девчонки украшения своими руками pc цифровая версия --> 10280 вырезаем и играем театр своими руками pc цифровая версия\n",
      "========\n",
      "20 --> 20\n",
      "6996 the talos principle deluxe edition ps4 английская версия --> 2785 destiny limited edition ps4 английская версия\n",
      "========\n",
      "78 --> 78\n",
      "900 1с школа окружающий мир 2 класс pc цифровая версия --> 899 1с школа окружающий мир 1 4 кл тесты pc цифровая версия\n",
      "========\n",
      "78 --> 78\n",
      "907 1с школа русский язык литература математика окружающий мир игры и задачи 1 4 классы pc цифровая версия --> 867 1с школа игры и задачи 1 4 классы pc цифровая версия\n",
      "========\n",
      "54 --> 54\n",
      "9111 бух 1с 12 декабрь 2015 цифровая версия --> 9114 бух 1с 12 декабрь 2013 цифровая версия\n",
      "========\n",
      "47 --> 47\n",
      "13209 комикс adventure time вс наоборот --> 13210 комикс adventure time книга 1\n",
      "========\n",
      "55 --> 55\n",
      "5025 newman john revolve --> 5026 newman john tribute\n",
      "========\n",
      "47 --> 47\n",
      "13231 комикс бесобой том 4 балор --> 13229 комикс бесобой том 1 имя ему бесобой\n",
      "========\n",
      "31 --> 31\n",
      "7094 total war attila набор эпоха карла великого pc цифровая версия --> 7088 total war attila pc цифровая версия\n",
      "========\n",
      "55 --> 55\n",
      "5064 noize mc акустика фирм --> 5065 noize mc последний альбом\n",
      "========\n",
      "47 --> 47\n",
      "13257 комикс инок том 4 зверь во мне книга 2 --> 13256 комикс инок том 2 проданная реликвия книга 2\n",
      "========\n",
      "31 --> 31\n",
      "7124 total war warhammer предзаказ pc цифровая версия --> 7100 total war rome ii предзаказ pc цифровая версия\n",
      "========\n",
      "47 --> 47\n",
      "13271 комикс майор гром том 4 как в сказке --> 13268 комикс майор гром том 1 чумной доктор книга 1\n",
      "========\n",
      "47 --> 47\n",
      "13273 комикс манхэттенские проекты том 2 --> 13272 комикс манхэттенские проекты том 1\n",
      "========\n",
      "20 --> 20\n",
      "19417 снупи большое приключение ps4 английская версия --> 7172 ufc ps4 английская версия\n",
      "========\n",
      "47 --> 47\n",
      "13295 комикс сага том 1 --> 13296 комикс сага том 2\n",
      "========\n",
      "47 --> 47\n",
      "13303 комикс скотт пилигрим том 4 скотт пилигрим берется за ум --> 13302 комикс скотт пилигрим том 3 скотт пилигрим и бесконечная печаль\n",
      "========\n",
      "47 --> 47\n",
      "13309 комикс сорвиголова том 2 --> 13308 комикс сорвиголова том 1\n",
      "========\n",
      "67 --> 67\n",
      "5121 nanoblock кенгуру --> 5132 nanoblock кролик\n",
      "========\n",
      "47 --> 47\n",
      "13313 комикс супермен бэтмен том 2 супердевушка --> 13312 комикс супермен бэтмен том 1 враги общества\n",
      "========\n",
      "63 --> 63\n",
      "15418 мягкая игрушка обезьянка вика в носочках 32см --> 15417 мягкая игрушка обезьянка вика в носочках 23см\n",
      "========\n",
      "49 --> 49\n",
      "13392 компл вопр серт экзам по подс междун фин учет в 1с erp уп 2 сист пр 1с пр 8 сентябрь 2015 1с паблишинг --> 13388 компл вопр серт экзам 1с розница 8 сентябрь 2011г 1с паблишинг\n",
      "========\n",
      "15 --> 15\n",
      "13394 комплект microsoft xbox 360 250gb xbox 360 kinect сенсор s7g 00014 --> 13396 комплект microsoft xbox 360 e 250gb xbox 360 kinect сенсор kinect adventures 5cx 00012\n",
      "========\n",
      "63 --> 63\n",
      "15467 мягкая игрушка смеющаяся обезьяна интерактивная --> 15421 мягкая игрушка обезьянка вязаная интерактивная\n",
      "========\n",
      "31 --> 31\n",
      "5248 need for speed rivals ultimate edition pc цифровая версия epay --> 5239 need for speed rivals pc цифровая версия\n",
      "========\n",
      "12 --> 12\n",
      "13445 комплект sony playstation 4 1tb black cuh 1208b дополнительный контроллер черный --> 13449 комплект sony playstation 4 500 gb black cuh 1208a b01 дополнительный контроллер черный\n",
      "========\n",
      "12 --> 12\n",
      "13446 комплект sony playstation 4 1tb black cuh 1208b игра call of duty black ops iii --> 6674 sony playstation 4 1tb black cuh 1208b\n",
      "========\n",
      "12 --> 12\n",
      "13447 комплект sony playstation 4 1tb black cuh 1208b игра star wars battlefront --> 6674 sony playstation 4 1tb black cuh 1208b\n",
      "========\n",
      "24 --> 24\n",
      "5269 need for speed xbox one русская версия --> 5244 need for speed rivals xbox one русская версия\n",
      "========\n",
      "76 --> 76\n",
      "1184 abbyy lingvo x6 многоязычная профессиональная версия full версия для скачивания pc цифровая версия --> 1183 abbyy lingvo x6 многоязычная домашняя версия full версия для скачивания pc цифровая версия\n",
      "========\n",
      "40 --> 40\n",
      "15564 наследники --> 13839 ледник\n",
      "========\n",
      "55 --> 55\n",
      "21708 черный кофе легенды русского рока фирм --> 15793 ноль легенды русского рока фирм\n",
      "========\n",
      "55 --> 55\n",
      "1252 adele 25 --> 1249 adele 21\n",
      "========\n",
      "40 --> 40\n",
      "1284 angry birds коллекция мультфильмов сезон 2 том 1 регион --> 12686 коллекция мультфильмов dreamworks колл\n",
      "========\n",
      "31 --> 31\n",
      "3338 fifa 15 pc цифровая версия epay --> 3339 fifa 15 pc цифровая версия\n",
      "========\n",
      "40 --> 40\n",
      "15665 необыкновенное путешествие серафимы --> 16470 обыкновенное чудо регион\n",
      "========\n",
      "31 --> 31\n",
      "9530 в тылу врага 2 штурм полное издание pc цифровая версия --> 9527 в тылу врага 2 штурм pc цифровая версия\n",
      "========\n",
      "55 --> 55\n",
      "9537 ва банкъ легенды русского рока фирм --> 15793 ноль легенды русского рока фирм\n",
      "========\n",
      "31 --> 31\n",
      "1394 alien isolation коллекция pc цифровая версия --> 1385 alien isolation pc цифровая версия\n",
      "========\n",
      "64 --> 64\n",
      "7572 warhammer 40000 miniatures chaos cultists малый набор арт 35 34 --> 7575 warhammer 40000 miniatures chaos space marines малый набор арт 35 31\n",
      "========\n",
      "31 --> 31\n",
      "1434 anno 2205 season pass pc цифровая версия --> 1435 anno 2205 standard edition pc цифровая версия\n",
      "========\n",
      "31 --> 31\n",
      "1436 anno 2205 ultimate edition pc цифровая версия --> 1435 anno 2205 standard edition pc цифровая версия\n",
      "========\n",
      "64 --> 64\n",
      "7592 warhammer 40000 miniatures pathfinder team арт 56 09 --> 7604 warhammer 40000 miniatures tau fire warriors арт 56 06\n",
      "========\n",
      "64 --> 64\n",
      "7607 warhammer accessories citadel base paint set арт 60 22 --> 7609 warhammer accessories citadel shade paint set арт 60 23\n",
      "========\n",
      "76 --> 76\n",
      "5569 promt для ms office только для домашнего использования pc цифровая версия --> 5564 promt professional 10 домашний а р а только для домашнего использования pc цифровая версия\n",
      "========\n",
      "40 --> 40\n",
      "17858 рикки и флэш --> 17859 рикки тикки тави регион\n",
      "========\n",
      "30 --> 30\n",
      "3538 football manager 2016 pc jewel русская версия --> 3530 football manager 2012 pc jewel русская версия\n",
      "========\n",
      "61 --> 61\n",
      "21974 шарф doctor who tardis синий 1 6м --> 18105 рюкзак doctor who tardis синий 50см\n",
      "========\n",
      "40 --> 40\n",
      "11762 золотой фонд отечественного кино владимир высоцкий 4dvd --> 11774 золотой фонд отечественного кино сказки часть 1 4dvd\n",
      "========\n",
      "40 --> 40\n",
      "11764 золотой фонд отечественного кино достоевский ф м экранизации 5dvd rem --> 11768 золотой фонд отечественного кино комедии часть 1 4dvd\n",
      "========\n",
      "55 --> 55\n",
      "7669 winehouse amy amy --> 7672 winehouse amy frank\n",
      "========\n",
      "40 --> 40\n",
      "11765 золотой фонд отечественного кино зимние сказки 4dvd --> 11774 золотой фонд отечественного кино сказки часть 1 4dvd\n",
      "========\n",
      "61 --> 61\n",
      "19966 толстовка batman arkham knight с ушками m --> 19965 толстовка batman arkham knight с ушками l\n",
      "========\n",
      "61 --> 61\n",
      "19973 толстовка minecraft enderman zip up детская черная s 36 38 --> 19972 толстовка minecraft enderman zip up детская черная m 38 40\n",
      "========\n",
      "40 --> 40\n",
      "22022 экипаж rem --> 22020 экипаж\n",
      "========\n",
      "40 --> 40\n",
      "17929 розыгрыш rem --> 8664 барбоскины выпуск 4 розыгрыш регион\n",
      "========\n",
      "40 --> 40\n",
      "22035 элвин и бурундуки 2 --> 22037 элвин и бурундуки 2 р\n",
      "========\n",
      "55 --> 55\n",
      "3604 gomez selena revival --> 3605 gomez selena stars dance\n",
      "========\n",
      "40 --> 40\n",
      "13846 ледниковый период 3 эра динозавров 2dvd ххси --> 13848 ледниковый период 3 эра динозавров м ф регион\n",
      "========\n",
      "3 --> 3\n",
      "5667 ps4 контроллер игровой беспроводной золотой dualshock 4 cont cuh zct1e scee --> 5666 ps4 контроллер игровой беспроводной белый dualshock 4 cont glacier white cuh zct1e scee\n",
      "========\n",
      "31 --> 31\n",
      "1574 assassin s creed синдикат gold edition pc цифровая версия --> 1573 assassin s creed синдикат gold edition предзаказ pc цифровая версия\n",
      "========\n",
      "31 --> 31\n",
      "1575 assassin s creed синдикат season pass pc цифровая версия --> 1576 assassin s creed синдикат standard edition pc цифровая версия\n",
      "========\n",
      "40 --> 40\n",
      "9767 великая --> 9768 великая война колл\n",
      "========\n",
      "55 --> 55\n",
      "3627 gregorian masters of chant x the final chapter --> 3628 gregorian masters chant and chill\n",
      "========\n",
      "31 --> 31\n",
      "7766 warhammer 40 000 dawn of war ii retribution набор кровавых эффектов ridiculously bloody blood pack pc цифровая версия --> 7748 warhammer 40 000 dawn of war ii retribution pc цифровая версия\n",
      "========\n",
      "55 --> 55\n",
      "1683 bjork vulnicura strings --> 1682 bjork vulnicura\n",
      "========\n",
      "40 --> 40\n",
      "11946 иди и смотри rem --> 19075 станционный смотритель rem\n",
      "========\n",
      "64 --> 64\n",
      "16051 настольная игра world of tanks rush последний бой арт 1483 --> 16056 настольная игра world of tanks rush арт 1123\n",
      "========\n",
      "31 --> 31\n",
      "5812 plants vs zombies garden warfare pc цифровая версия epay --> 5818 plants vs zombies garden warfare pc цифровая версия\n",
      "========\n",
      "28 --> 28\n",
      "7862 world of warcraft warlords of draenor дополнение pc jewel русская версия --> 7863 world of warcraft warlords of draenor дополнение pc русская версия\n",
      "========\n",
      "65 --> 65\n",
      "5826 pokemon xy грохочущие небеса блистер --> 5827 pokemon xy двойной блистер\n",
      "========\n",
      "55 --> 55\n",
      "1732 bocelli andrea cinema --> 1730 bocelli andrea bocelli\n",
      "========\n",
      "64 --> 64\n",
      "16095 настольная игра диксит 2 --> 16094 настольная игра диксит 1\n",
      "========\n",
      "64 --> 64\n",
      "16096 настольная игра диксит 3 --> 16094 настольная игра диксит 1\n",
      "========\n",
      "64 --> 64\n",
      "16097 настольная игра диксит 4 --> 16094 настольная игра диксит 1\n",
      "========\n",
      "64 --> 64\n",
      "16099 настольная игра доббль --> 16057 настольная игра рш\n",
      "========\n",
      "40 --> 40\n",
      "9972 воин регион --> 9970 воин\n",
      "========\n",
      "7 --> 7\n",
      "7926 xone беспроводной геймпад для xbox one elite hm3 00005 --> 7925 xone беспроводной геймпад titanfall для xbox one j72 00016\n",
      "========\n",
      "65 --> 65\n",
      "16139 настольная игра кубики историй бэтмен --> 16138 настольная игра кубики историй\n",
      "========\n",
      "65 --> 65\n",
      "16141 настольная игра кубики историй муми тролли --> 16138 настольная игра кубики историй\n",
      "========\n",
      "76 --> 76\n",
      "3867 hetman data recovery pack домашняя версия pc цифровая версия --> 6059 rs data recovery домашняя версия recovery software pc цифровая версия\n",
      "========\n",
      "64 --> 64\n",
      "16157 настольная игра манчкин замес арт 1487 --> 16159 настольная игра манчкин зомби арт 1001\n",
      "========\n",
      "40 --> 40\n",
      "10045 восхождение юпитер регион --> 10042 восхождение юпитер\n",
      "========\n",
      "40 --> 40\n",
      "10054 враг государства 1 регион --> 12055 история государства российского том 1 регион\n",
      "========\n",
      "31 --> 31\n",
      "1867 battlefield 4 premium service pc цифровая версия epay --> 1852 battlefield 4 premium service pc цифровая версия\n",
      "========\n",
      "16 --> 16\n",
      "12117 игровая консоль microsoft xbox one 1tb гибридной памяти с геймпадом elite kg4 00062 --> 13531 комплект предзаказа на консоль microsoft xbox one 1tb гибридной памяти с геймпадом elite цифровая версия\n",
      "========\n",
      "31 --> 31\n",
      "1885 battlefield hardline standard edition pc цифровая версия epay --> 1875 battlefield hardline pc цифровая версия\n",
      "========\n",
      "55 --> 55\n",
      "3984 jarre jean michel electronika 1 the time machine --> 3987 jarre jean michel essential recollection\n",
      "========\n",
      "75 --> 75\n",
      "6066 radmin 3 522 лиц --> 6065 radmin 3 1 лиц\n",
      "========\n",
      "20 --> 20\n",
      "10203 ведьмак 3 дикая охота дополнение каменные сердца код загрузки без диска без карт ps4 русская версия --> 10202 ведьмак 3 дикая охота дополнение каменные сердца код загрузки без диска ps4 русская версия\n",
      "========\n",
      "72 --> 72\n",
      "20461 фигурка diablo diablo pop games 3 75 --> 20463 фигурка diablo iii shadow of diablo 7\n",
      "========\n",
      "58 --> 58\n",
      "8182 аквариум радио африка lp --> 8171 аквариум акустика lp\n",
      "========\n"
     ]
    }
   ],
   "source": [
    "missing_items_ids = set(items.item_id) - set(items.item_id).intersection(set(sales.item_id))\n",
    "nsales = sales.copy()\n",
    "\n",
    "for missing_item_id in missing_items_ids:\n",
    "    missing_item_name, missing_item_cat = items.loc[items['item_id'] == missing_item_id, ['item_name', 'item_category_id']].values[0]\n",
    "    similar_item_names = matches_df.loc[matches_df['left_side'] == missing_item_name, 'right_side'].values\n",
    "    if len(similar_item_names) > 0:\n",
    "        similar_item_name = similar_item_names[0]\n",
    "        similar_item_id, similar_item_cat, status = items.loc[items['item_name'] == similar_item_name, ['item_id', 'item_category_id', 'item_status']].values[0]\n",
    "        if missing_item_cat == similar_item_cat:\n",
    "            similar_item_sales = sales[sales['item_id'] == similar_item_id]\n",
    "            if len(similar_item_sales) > 0:\n",
    "                items.loc[items['item_id'] == missing_item_id, 'item_status'] = status\n",
    "                print(missing_item_cat, '-->', similar_item_cat)\n",
    "                print(missing_item_id, missing_item_name, '-->', similar_item_id, similar_item_name)\n",
    "                missing_item_sales = similar_item_sales.copy()\n",
    "                missing_item_sales.item_id = missing_item_id\n",
    "                nsales = pd.concat([nsales, missing_item_sales], ignore_index = True, sort = False, keys = nsales.columns.values)\n",
    "                print('========')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1753,
     "status": "ok",
     "timestamp": 1568520278354,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "Miw7nQLr2BQY",
    "outputId": "417b5d96-a638-4a72-e3f6-f59e86a19c01"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Remaining missing items  216\n"
     ]
    }
   ],
   "source": [
    "print('Remaining missing items ', len(set(items.item_id) - set(items.item_id).intersection(set(nsales.item_id))))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "DXYXktY-VMdl"
   },
   "source": [
    "### Give missing items median of the sales of items in the same shop and category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "P8V3f6VYR138"
   },
   "outputs": [],
   "source": [
    "aggs = {'item_cnt_day': 'median', 'item_id': 'count', 'item_price': 'median', }\n",
    "gb = sales.groupby(['item_category_id', 'date_block_num', 'shop_id', 'date']).agg(aggs).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UZbETCbnNA7Y"
   },
   "outputs": [],
   "source": [
    "missing_items_ids = set(items.item_id) - set(items.item_id).intersection(set(nsales.item_id))\n",
    "for missing_items_id in missing_items_ids:\n",
    "    missing_items_cat = items.loc[items['item_id'] == missing_items_id, 'item_category_id'].values[0]\n",
    "    missing_item_sales = gb[gb['item_category_id'] == missing_items_cat].copy()\n",
    "    missing_item_sales.loc[:, 'item_id'] = missing_items_id\n",
    "    nsales = pd.concat([nsales, missing_item_sales], ignore_index = True, sort = False, keys = nsales.columns.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2708,
     "status": "ok",
     "timestamp": 1568520364295,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "e9rpnc6WSffI",
    "outputId": "899a069b-cdae-4792-b2b4-ed128acca34b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Remaining missing items  0\n"
     ]
    }
   ],
   "source": [
    "print('Remaining missing items ', len(set(items.item_id) - set(items.item_id).intersection(set(nsales.item_id))))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "dQGPzXFk36a2"
   },
   "source": [
    "### update outdated info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2550,
     "status": "ok",
     "timestamp": 1568520392058,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "2dY1OTAV3-_3",
    "outputId": "f5c9eb06-3188-490c-c3e0-42fdc2bf22d5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "\n",
    "nitems_prices = create_target_for_attribute(nsales, 'item_id', 'item_price', 'date_block_num', aggfunc = np.median, fill = np.nan)\n",
    "\n",
    "\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 119
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1310,
     "status": "ok",
     "timestamp": 1568521242437,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "VWYblorE4Vyv",
    "outputId": "53d9f378-906d-40ce-f444-9ca5cf8c4463"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shops in train set:  55  out of 55\n",
      "Outdated shops in train set:  12.73 %\n",
      "items in train set:  21806  out of 22170\n",
      "Outdated items in train set:  56.84 %\n",
      "categories in train set:  84  out of 84\n",
      "Outdated categories in train set:  20.24 %\n"
     ]
    }
   ],
   "source": [
    "outdated_shops = shops_sales[shops_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('shops in train set: ', len(shops_sales), ' out of', len(shops))\n",
    "print('Outdated shops in train set: ',  round(100 * len(outdated_shops) / len(shops_sales), 2), '%',)\n",
    "shops['shop_status'] = 'unknown'\n",
    "shops.loc[shops['shop_id'].isin(outdated_shops['shop_id']), 'shop_status'] = 'deactive'\n",
    "active_shops =  shops_sales[shops_sales.loc[:, outdated_limit:].sum(axis = 1) > 0]\n",
    "shops.loc[shops['shop_id'].isin(active_shops['shop_id']), 'shop_status'] = 'active'\n",
    "\n",
    "outdated_items = items_sales[items_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('items in train set: ', len(items_sales), ' out of', len(items))\n",
    "print('Outdated items in train set: ',  round( 100 * len(outdated_items) / len(items_sales), 2), '%', )\n",
    "items['item_status'] = 'unknown'\n",
    "items.loc[items['item_id'].isin(outdated_items['item_id']), 'item_status'] = 'deactive'\n",
    "active_items =  items_sales[items_sales.loc[:, outdated_limit:].sum(axis = 1) > 0]\n",
    "items.loc[items['item_id'].isin(active_items['item_id']), 'item_status'] = 'active'\n",
    "\n",
    "outdated_categories = categories_sales[categories_sales.loc[:, outdated_limit:].sum(axis = 1) <= 0]\n",
    "print('categories in train set: ', len(categories_sales), ' out of', len(categories))\n",
    "print('Outdated categories in train set: ',  round( 100 * len(outdated_categories) / len(categories_sales), 2), '%', )\n",
    "\n",
    "categories['item_category_status'] = 'unknown'\n",
    "categories.loc[categories['item_category_id'].isin(outdated_categories['item_category_id']), 'item_category_status'] = 'deactive'\n",
    "\n",
    "active_categories = categories_sales[categories_sales.loc[:, outdated_limit:].sum(axis = 1) > 0]\n",
    "categories.loc[categories['item_category_id'].isin(active_categories['item_category_id']), 'item_category_status'] = 'active'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "oMHUiV44URF3"
   },
   "source": [
    "#### save results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Muy4plzXUXcW"
   },
   "outputs": [],
   "source": [
    "outdated_shops.to_pickle(data_path + 'outdated_shops.pkl')\n",
    "outdated_items.to_pickle(data_path + 'outdated_items.pkl')\n",
    "outdated_categories.to_pickle(data_path + 'outdated_categories.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WbZW8cTnaTXK"
   },
   "source": [
    "## stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1286,
     "status": "ok",
     "timestamp": 1568521289268,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "zzF9Ee-EX21l",
    "outputId": "c7d33966-b3d9-4b26-fbc8-97ed5848d16f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outdated items in test set: 164 out of 5100\n",
      "Outdated shops in test set: 0 out of 42\n"
     ]
    }
   ],
   "source": [
    "test_outdated_items = test[test['item_id'].isin(outdated_items['item_id'])]\n",
    "test_outdated_items_number = test_outdated_items['item_id'].nunique()\n",
    "print('Outdated items in test set:', test_outdated_items_number, 'out of', test.item_id.nunique())\n",
    "\n",
    "test_outdated_shops = test.loc[test['shop_id'].isin(outdated_shops['shop_id'])]['shop_id'].nunique()\n",
    "print('Outdated shops in test set:',test_outdated_shops, 'out of', test.shop_id.nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1305,
     "status": "ok",
     "timestamp": 1568521305723,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "UEzeS3rlRevG",
    "outputId": "8f303c00-e6c5-41bd-dda6-41272ecb2243"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outdated shops/items pairs in train set:  74.23 %\n"
     ]
    }
   ],
   "source": [
    "outdated_pairs = shops_items_trans[shops_items_trans.loc[:, outdated_limit:].sum(axis = 1) == 0]\n",
    "print('Outdated shops/items pairs in train set: ',  round(100 * len(outdated_pairs) / len(shops_items_trans), 2), '%',)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ANnvma1auJBS"
   },
   "source": [
    "## Generating feature from item_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RmIuGXtwh2GF"
   },
   "outputs": [],
   "source": [
    "items[['item_type',\t'item_subtype']] = items[['item_type',\t'item_subtype']].fillna('unknown')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1860,
     "status": "ok",
     "timestamp": 1568521335686,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "2bQkkJCCgHVf",
    "outputId": "92142bf1-8905-4427-fed4-c6968cf5ce84"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "features count:  26\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter\n",
    "from operator import itemgetter\n",
    "\n",
    "strings = items['item_type'].values.tolist() + items['item_subtype'].values.tolist()\n",
    "features = Counter(' '.join(strings).split(' ')).items()\n",
    "features = sorted(features, key = itemgetter(1))\n",
    "features = pd.DataFrame(features, columns=['feature', 'count'])\n",
    "features = features[(features['feature'].str.len() > 1) & (features['count'] > 150)]\n",
    "\n",
    "features = features.drop_duplicates(subset=['feature'])\n",
    "\n",
    "print('features count: ', len(features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "nbYl2EnPXyN7"
   },
   "outputs": [],
   "source": [
    "items['item_maintype'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "moCxdbzzjZbW"
   },
   "outputs": [],
   "source": [
    "for feature in features['feature']:\n",
    "    items.loc[items['item_name'].str.contains(feature), 'item_maintype'] = feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "uW7o4CXnlYyO"
   },
   "outputs": [],
   "source": [
    "items['item_maintype'] = items['item_maintype'].fillna('unknown')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 425
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 928,
     "status": "ok",
     "timestamp": 1568521368131,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "URZLGXcykdmj",
    "outputId": "91209209-e8f1-496f-f230-790541ccbab3"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "unknown       9224\n",
       "версия        3515\n",
       "bd            2279\n",
       "регион        1848\n",
       "cd            1050\n",
       "jewel          761\n",
       "dvd            669\n",
       "фирм           544\n",
       "pc             529\n",
       "русская        286\n",
       "xbox           283\n",
       "rem            254\n",
       "ps3            230\n",
       "digipack       169\n",
       "ps             157\n",
       "3d             124\n",
       "box             97\n",
       "субтитры        66\n",
       "mp3             51\n",
       "360             20\n",
       "русские          9\n",
       "цифровая         4\n",
       "essentials       1\n",
       "Name: item_maintype, dtype: int64"
      ]
     },
     "execution_count": 53,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "items['item_maintype'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3kFH6PYi3NRa"
   },
   "outputs": [],
   "source": [
    "items = items.drop(columns = ['item_type', 'item_subtype'], axis = 'columns')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qwLAspPygJuD"
   },
   "source": [
    "## Trend and seasonality"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "tMdNQnpzgNky"
   },
   "source": [
    "Any time-series signal can be expressed as:\n",
    "\n",
    "ts = tend + seanonality + noise\n",
    "\n",
    "**tend**: is the signal's amplitude inceasing or decreasing.\n",
    "\n",
    "**seasonality**: variations in the signal's amplitude that occurs at certain point at time (at holiday season, sales increase every year)\n",
    "\n",
    "\n",
    "Before feeding time-series data to the machine, we should remove the trend and seasonality for those reasons:\n",
    "- machines like to work with data that have constant mean and variance over time\n",
    "- by removing the trend and seasonality, we make the mapping task between the dependent varaibles and the target more easier for machine. In other words, there is no need for the machine to learn extra infomation(trend and seasonality)in order to predict the target. It is enough and easier to learn how the noise in the singnal influence the target\n",
    "- Trend and seasonality rely on time, by removing them we make data independent from time and that makes prediction task more easier\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 441
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1694,
     "status": "ok",
     "timestamp": 1568339748898,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "iFNFFA9HpBIy",
    "outputId": "1d596129-8aaa-41fc-f608-37bf54a65efe"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGoCAYAAABL+58oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noVAoFGWO/wMb+o8FyUYUqgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x432 with 4 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "shop_7 = shops_sales[shops_sales['shop_id'] == 7]\n",
    "shop_7 = shop_7.loc[:, '0':].T\n",
    "decomposition = seasonal_decompose(shop_7, freq = 12, two_sided = False)\n",
    "trend = decomposition.trend\n",
    "seasonal = decomposition.seasonal\n",
    "residual = decomposition.resid\n",
    "\n",
    "plt.subplot(411)\n",
    "plt.plot(shop_7, label='Shop 7 sales')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(412)\n",
    "plt.plot(trend, label='Trend')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(413)\n",
    "plt.plot(seasonal,label='Seasonality')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(414)\n",
    "plt.plot(residual, label='Residuals')\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8B4ItUeIwnx2"
   },
   "source": [
    "Any method that we use to estimate trend and seanonality in the data requires to throw away historical data. By example: to find the trend and seasonality for shop 7 sales, we needed to throw away data of the first year.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ulYtCHvHmflJ"
   },
   "source": [
    "**Notes**\n",
    "- Not all pairs(shop/item) have trend or seasonality (it is hard to analyse all pairs manually by ploting their sales)\n",
    "- lots of shops start selling late and lots of items start to be sold also late. This means first year of data is not enough to estimate accuretly those measures"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "L9FsCtbCo-BC"
   },
   "source": [
    "### shop/item sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 510
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2889,
     "status": "ok",
     "timestamp": 1568339848776,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "x-QDj1Dfwmsl",
    "outputId": "16f86cc6-dd55-496d-cc9b-451ee794f555"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Until month:  5 ,  56.85 %, Item/Shops Pairs have made No sales\n",
      "Until month:  6 ,  53.64 %, Item/Shops Pairs have made No sales\n",
      "Until month:  7 ,  50.63 %, Item/Shops Pairs have made No sales\n",
      "Until month:  8 ,  48.04 %, Item/Shops Pairs have made No sales\n",
      "Until month:  9 ,  45.2 %, Item/Shops Pairs have made No sales\n",
      "Until month:  10 ,  42.31 %, Item/Shops Pairs have made No sales\n",
      "Until month:  11 ,  38.64 %, Item/Shops Pairs have made No sales\n",
      "Until month:  12 ,  36.79 %, Item/Shops Pairs have made No sales\n",
      "Until month:  13 ,  34.89 %, Item/Shops Pairs have made No sales\n",
      "Until month:  14 ,  32.67 %, Item/Shops Pairs have made No sales\n",
      "Until month:  15 ,  30.55 %, Item/Shops Pairs have made No sales\n",
      "Until month:  16 ,  28.46 %, Item/Shops Pairs have made No sales\n",
      "Until month:  17 ,  26.6 %, Item/Shops Pairs have made No sales\n",
      "Until month:  18 ,  24.86 %, Item/Shops Pairs have made No sales\n",
      "Until month:  19 ,  23.02 %, Item/Shops Pairs have made No sales\n",
      "Until month:  20 ,  21.14 %, Item/Shops Pairs have made No sales\n",
      "Until month:  21 ,  19.02 %, Item/Shops Pairs have made No sales\n",
      "Until month:  22 ,  16.56 %, Item/Shops Pairs have made No sales\n",
      "Until month:  23 ,  13.43 %, Item/Shops Pairs have made No sales\n",
      "Until month:  24 ,  12.09 %, Item/Shops Pairs have made No sales\n",
      "Until month:  25 ,  10.72 %, Item/Shops Pairs have made No sales\n",
      "Until month:  26 ,  9.19 %, Item/Shops Pairs have made No sales\n",
      "Until month:  27 ,  7.88 %, Item/Shops Pairs have made No sales\n",
      "Until month:  28 ,  6.68 %, Item/Shops Pairs have made No sales\n",
      "Until month:  29 ,  5.49 %, Item/Shops Pairs have made No sales\n",
      "Until month:  30 ,  4.28 %, Item/Shops Pairs have made No sales\n",
      "Until month:  31 ,  3.0 %, Item/Shops Pairs have made No sales\n",
      "Until month:  32 ,  1.77 %, Item/Shops Pairs have made No sales\n",
      "Until month:  33 ,  0.0 %, Item/Shops Pairs have made No sales\n"
     ]
    }
   ],
   "source": [
    "for month in range(5, 34):\n",
    "    pairs = shops_items_trans[shops_items_trans.loc[:,'0': str(month)].sum(axis=1) == 0]\n",
    "    percent = round(100 * len(pairs) / len(shops_items_trans), 2)\n",
    "    print('Until month: ', month, ', ', percent, '%, Item/Shops Pairs have made No sales')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 441
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2055,
     "status": "ok",
     "timestamp": 1568339853742,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "w5OKEAnT1d6o",
    "outputId": "d8c00277-5a43-4596-e03d-fb7243d1f547"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGoCAYAAABL+58oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njuM0Hf8BwCJgD6ev6ggAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x432 with 4 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "shop_item = shops_items_sales[shops_items_sales['shop_id'] == 7].sample(1)\n",
    "shop_item_sale = shop_item.loc[:, '0':].T\n",
    "decomposition = seasonal_decompose(shop_item_sale, freq = 6, two_sided = False, extrapolate_trend = 6)\n",
    "trend = decomposition.trend\n",
    "seasonal = decomposition.seasonal\n",
    "residual = decomposition.resid\n",
    "\n",
    "plt.subplot(411)\n",
    "plt.plot(shop_item_sale, label='Shop = 7, item =' + str(shop_item['item_id'].values[0]))\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(412)\n",
    "plt.plot(trend, label='Trend')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(413)\n",
    "plt.plot(seasonal,label='Seasonality')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(414)\n",
    "plt.plot(residual, label='Residuals')\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9TtXgd43Zzqi"
   },
   "outputs": [],
   "source": [
    "dt_shops_items_sales = detrend_df_sales(shops_items_sales).T\n",
    "nshops_items_sales = shops_items_sales.loc[:, ['shop_id', 'item_id']]\n",
    "nshops_items_sales = pd.concat([nshops_items_sales, dt_shops_items_sales], axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ub75sfyQ2UN5"
   },
   "source": [
    "### Items sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 510
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 422,
     "status": "ok",
     "timestamp": 1568339874527,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "QPmOzlZi2ZD8",
    "outputId": "f0b0414a-8ae6-463d-8247-c5687feafec2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Until month:  5 ,  44.9 %, Items have made No sales\n",
      "Until month:  6 ,  42.87 %, Items have made No sales\n",
      "Until month:  7 ,  40.75 %, Items have made No sales\n",
      "Until month:  8 ,  39.02 %, Items have made No sales\n",
      "Until month:  9 ,  36.05 %, Items have made No sales\n",
      "Until month:  10 ,  33.78 %, Items have made No sales\n",
      "Until month:  11 ,  31.36 %, Items have made No sales\n",
      "Until month:  12 ,  30.42 %, Items have made No sales\n",
      "Until month:  13 ,  28.87 %, Items have made No sales\n",
      "Until month:  14 ,  27.47 %, Items have made No sales\n",
      "Until month:  15 ,  26.02 %, Items have made No sales\n",
      "Until month:  16 ,  24.84 %, Items have made No sales\n",
      "Until month:  17 ,  23.69 %, Items have made No sales\n",
      "Until month:  18 ,  22.24 %, Items have made No sales\n",
      "Until month:  19 ,  21.11 %, Items have made No sales\n",
      "Until month:  20 ,  19.57 %, Items have made No sales\n",
      "Until month:  21 ,  17.03 %, Items have made No sales\n",
      "Until month:  22 ,  14.94 %, Items have made No sales\n",
      "Until month:  23 ,  12.79 %, Items have made No sales\n",
      "Until month:  24 ,  11.9 %, Items have made No sales\n",
      "Until month:  25 ,  10.87 %, Items have made No sales\n",
      "Until month:  26 ,  9.28 %, Items have made No sales\n",
      "Until month:  27 ,  8.07 %, Items have made No sales\n",
      "Until month:  28 ,  7.04 %, Items have made No sales\n",
      "Until month:  29 ,  5.9 %, Items have made No sales\n",
      "Until month:  30 ,  4.94 %, Items have made No sales\n",
      "Until month:  31 ,  3.62 %, Items have made No sales\n",
      "Until month:  32 ,  2.22 %, Items have made No sales\n",
      "Until month:  33 ,  0.0 %, Items have made No sales\n"
     ]
    }
   ],
   "source": [
    "for month in range(5, 34):\n",
    "    items_temp = items_trans[items_trans.loc[:,'0': str(month)].sum(axis=1) == 0]\n",
    "    percent = round(100 * len(items_temp) / len(items_trans), 2)\n",
    "    print('Until month: ', month, ', ', percent, '%, Items have made No sales')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 441
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1904,
     "status": "ok",
     "timestamp": 1568339880571,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "qZO3gJGk2sY6",
    "outputId": "1bd2d8f6-b552-481f-bda3-bc1e53c0a7ed"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGoCAYAAABL+58oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x432 with 4 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "item_temp = items_sales.nlargest(50, columns = [str(itr) for itr in np.arange(27, 34)]).sample(1)\n",
    "#item = items_sales.sample(1)\n",
    "item_temp_sales = item_temp.loc[:, '0':].T\n",
    "decomposition = seasonal_decompose(item_temp_sales, freq = 12, two_sided = False, extrapolate_trend = 0)\n",
    "trend = decomposition.trend\n",
    "seasonal = decomposition.seasonal\n",
    "residual = decomposition.resid\n",
    "\n",
    "plt.subplot(411)\n",
    "plt.plot(item_temp_sales, label='item =' + str(item_temp['item_id'].values[0]))\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(412)\n",
    "plt.plot(trend, label='Trend')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(413)\n",
    "plt.plot(seasonal,label='Seasonality')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(414)\n",
    "plt.plot(residual, label='Residuals')\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "YKkwdJhV2YAK"
   },
   "source": [
    "Although the lake of data about item sales. But we can detrend and deseasonlize their sales. We assume that although the randomness of demand on items but there is lots of items that people buy in certain time of the year and in general all items are sold less than before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "i69f0DQ_ZyZ4"
   },
   "outputs": [],
   "source": [
    "dt_items_sales = detrend_df_sales(items_sales).T\n",
    "nitems_sales = items_sales.loc[:, 'item_id']\n",
    "nitems_sales = pd.concat([nitems_sales, dt_items_sales], axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "x7pNRPmD6ciJ"
   },
   "source": [
    "### shops sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 510
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 426,
     "status": "ok",
     "timestamp": 1568339902698,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "stncx_KU6gO1",
    "outputId": "b97dab9f-aed3-45c5-ce01-7d445645913f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Until month:  5 ,  14.55 %, Items have made No sales\n",
      "Until month:  6 ,  14.55 %, Items have made No sales\n",
      "Until month:  7 ,  14.55 %, Items have made No sales\n",
      "Until month:  8 ,  14.55 %, Items have made No sales\n",
      "Until month:  9 ,  12.73 %, Items have made No sales\n",
      "Until month:  10 ,  12.73 %, Items have made No sales\n",
      "Until month:  11 ,  10.91 %, Items have made No sales\n",
      "Until month:  12 ,  10.91 %, Items have made No sales\n",
      "Until month:  13 ,  10.91 %, Items have made No sales\n",
      "Until month:  14 ,  9.09 %, Items have made No sales\n",
      "Until month:  15 ,  7.27 %, Items have made No sales\n",
      "Until month:  16 ,  7.27 %, Items have made No sales\n",
      "Until month:  17 ,  7.27 %, Items have made No sales\n",
      "Until month:  18 ,  5.45 %, Items have made No sales\n",
      "Until month:  19 ,  3.64 %, Items have made No sales\n",
      "Until month:  20 ,  3.64 %, Items have made No sales\n",
      "Until month:  21 ,  1.82 %, Items have made No sales\n",
      "Until month:  22 ,  1.82 %, Items have made No sales\n",
      "Until month:  23 ,  1.82 %, Items have made No sales\n",
      "Until month:  24 ,  1.82 %, Items have made No sales\n",
      "Until month:  25 ,  1.82 %, Items have made No sales\n",
      "Until month:  26 ,  1.82 %, Items have made No sales\n",
      "Until month:  27 ,  1.82 %, Items have made No sales\n",
      "Until month:  28 ,  1.82 %, Items have made No sales\n",
      "Until month:  29 ,  1.82 %, Items have made No sales\n",
      "Until month:  30 ,  1.82 %, Items have made No sales\n",
      "Until month:  31 ,  1.82 %, Items have made No sales\n",
      "Until month:  32 ,  1.82 %, Items have made No sales\n",
      "Until month:  33 ,  0.0 %, Items have made No sales\n"
     ]
    }
   ],
   "source": [
    "for month in range(5, 34):\n",
    "    shops_temp = shops_trans[shops_trans.loc[:,'0': str(month)].sum(axis=1) == 0]\n",
    "    percent = round(100 * len(shops_temp) / len(shops_trans), 2)\n",
    "    print('Until month: ', month, ', ', percent, '%, Items have made No sales')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 441
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1822,
     "status": "ok",
     "timestamp": 1568339907040,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "wXRfKhaQ6sZd",
    "outputId": "71eb0631-ee5b-42c9-d143-b9414f0b5c34"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGoCAYAAABL+58oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<Figure size 576x432 with 4 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "shop_temp = shops_sales.sample(1)\n",
    "shop_temp_sales = shop_temp.loc[:, '0':].T\n",
    "decomposition = seasonal_decompose(shop_temp_sales, freq = 12, two_sided = False, extrapolate_trend = 6)\n",
    "trend = decomposition.trend\n",
    "seasonal = decomposition.seasonal\n",
    "residual = decomposition.resid\n",
    "\n",
    "plt.subplot(411)\n",
    "plt.plot(shop_temp_sales, label='shop = '+str(shop_temp['shop_id'].values[0]))\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(412)\n",
    "plt.plot(trend, label='Trend')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(413)\n",
    "plt.plot(seasonal,label='Seasonality')\n",
    "plt.legend(loc='best')\n",
    "plt.subplot(414)\n",
    "plt.plot(residual, label='Residuals')\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "zmV9U19769_3"
   },
   "source": [
    "Time series analysis is powerful and by detrending and deseasonalizing the data you throw away unneeded complexity in the data.\n",
    "\n",
    "In our case, we don't have enough sales history to perform such analysis. Indeed if you perform this analysis on the data you get nice score on the training set and validation set but your model will fail to generlize to the test set. The sales history for items/shops is relativly small and as a result the model will overfit miserably since the model will not be able to find general patterns in the noise of the sales data (no enough data)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "gKkmyOU4ZxXV"
   },
   "outputs": [],
   "source": [
    "dt_shops_sales = detrend_df_sales(shops_sales).T\n",
    "nshops_sales = shops_sales.loc[:, 'shop_id']\n",
    "nshops_sales = pd.concat([nshops_sales, dt_shops_sales], axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4PCOhLp6BzF7"
   },
   "source": [
    "## Lag features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "QZW-RDjHRjzf"
   },
   "outputs": [],
   "source": [
    "from statsmodels.tsa.stattools import acf\n",
    "time_shift = 14"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "9-68q3iUgYsw"
   },
   "source": [
    "### How much to go back in time for pairs(shop/item)\n",
    "choose 100 pairs and compute how much the previous values of the series (lags) may be helpful in predicting the current value. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "PWbL86c6oaxK"
   },
   "outputs": [],
   "source": [
    "shops_items_sales_temp = shops_items_sales.sample(1000)#.nlargest(10**4, columns = [str(itr) for itr in range(27, 33)])\n",
    "pairs_acf = np.zeros((shops_items_sales_temp.shape[0], time_shift + 1))\n",
    "for i, (ind, shop_item_sales) in enumerate(shops_items_sales_temp.iterrows()):\n",
    "    pair = shop_item_sales.loc['0': ]\n",
    "  #if np.sum(pair) != 0:\n",
    "    acf_12 = acf(pair, nlags = time_shift, fft = True)\n",
    "    pairs_acf[i, :] = acf_12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 390
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1338,
     "status": "ok",
     "timestamp": 1568521521514,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "ykJXkQ47Vjys",
    "outputId": "d3a7f6c3-f785-42c6-d839-e6d8551a132c"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAF1CAYAAADSoyIcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nI2Tl3gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "avgs = np.mean(pairs_acf, axis = 0)\n",
    "plt.bar(x = np.arange(time_shift + 1), height = avgs)\n",
    "plt.title('lag importance of shop/item pairs')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UAfptIVM6J4s"
   },
   "outputs": [],
   "source": [
    "pair_lags = [1, 2, 3, 4, 9, 10, 11, 12]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "dTdQux4Ni94E"
   },
   "source": [
    "### How much to go back in time for items"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "c5zXu4I5pGh-"
   },
   "outputs": [],
   "source": [
    "items_acf = np.zeros((items_sales.shape[0], time_shift + 1))\n",
    "for i, item_sales in items_sales.iterrows():\n",
    "    item_temp = item_sales.loc['0': ]\n",
    "    if np.sum(item_temp) != 0:\n",
    "        acf_12 = acf(item_temp, nlags = time_shift, fft = True)\n",
    "        items_acf[i, :] = acf_12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 390
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 11165,
     "status": "ok",
     "timestamp": 1568521572483,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "e6puC3Bsap8f",
    "outputId": "c8f1179a-d6cf-4864-f12f-8aca45b10efa"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAF1CAYAAADSoyIcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "avgs = np.mean(items_acf, axis = 0)\n",
    "plt.bar(x = np.arange(time_shift + 1), height = avgs)\n",
    "plt.title('lag importance of items')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Dfi8_qLg6g9d"
   },
   "outputs": [],
   "source": [
    "item_lags = [1, 2, 3, 4, 5, 6, 11, 12]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "9YbM2OvWjGdc"
   },
   "source": [
    "### How much to go back in time for shops"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9mh9bf_8oyst"
   },
   "outputs": [],
   "source": [
    "shops_acf = np.zeros((shops_sales.shape[0], time_shift + 1))\n",
    "for i, shop_sales in shops_sales.iterrows():\n",
    "    shop_temp = shop_sales.loc['0': ]\n",
    "    acf_12 = acf(shop_temp, nlags = time_shift, fft = True)\n",
    "    shops_acf[i, :] = acf_12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 390
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 893,
     "status": "ok",
     "timestamp": 1568521631473,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "6RqmD3r6es2V",
    "outputId": "98db05de-683c-41a8-a3b5-d92c94b5c6b2"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAF1CAYAAADSoyIcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "avgs = np.mean(shops_acf, axis = 0)\n",
    "plt.bar(x = np.arange(time_shift + 1), height = avgs)\n",
    "plt.title('lag importance of shops')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6BKfhpj_6x5h"
   },
   "outputs": [],
   "source": [
    "shop_lags = [1, 2, 3, 4, 5, 6, 9, 12]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "TB1px-8NSido"
   },
   "source": [
    "### How much to go back in time for categories"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SwhC5bDpSieC"
   },
   "outputs": [],
   "source": [
    "categories_acf = np.zeros((categories_sales.shape[0], time_shift + 1))\n",
    "for i, category_sales in categories_sales.iterrows():\n",
    "    category_temp = category_sales.loc['0': ]\n",
    "    acf_12 = acf(category_temp, nlags = time_shift, fft = True)\n",
    "    categories_acf[i, :] = acf_12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 390
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1348,
     "status": "ok",
     "timestamp": 1568521669734,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "0wHMdZS2SieE",
    "outputId": "6d052e7b-08d8-4327-80ca-4cd26d0af062"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAF1CAYAAADSoyIcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAGT5JREFUeJzt3X+0XWV95/H3hwQEUUFNpJoEw7KR\nmjq2agaxtiNL0QnQEteMVqhWsSi1SypW1EbbhRY7TqxVa5f4A3+MrThSStXJSFqw/qgzVRyCCkoQ\nGxFMIkr4If6qQup3/tg79XC9N/ckOZfz3Hver7XO4uy9n+z9ffa53M95nr3PuakqJElSOw4YdwGS\nJOnuDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrOaluSGJMfPwX5fleTdo97vpEhySJL/neSO\nJH877npGxZ8LtWLxuAuQxqGqXjfuGnZL8hrg56vq2eOuZS88HTgCeGBV7RrVTsd9Llr6udBkc+Qs\njVGS+foG+aHAV0cZzOM2j18LLUCGs+aNJMck+WyS7yS5Kclbkxw0sP2pSa7rp1rfluSfkjx/hn29\nJskF/fOVSSrJ85JsS3J7khcm+Y9Jru6P99aBf3takn/uj39Hkq8kefLA9ock2ZjktiRbk7xgynEv\nTnJBku8CLwReBTwzyfeTXNW3e16Sa5N8L8n1SX53YB/HJdme5OwkN/fn4nkD2w9J8sYkN/b1/d8k\nh/Tbjk3ymb5PVyU5bg/n+xFJPtW3vSbJyf36PwHOGaj59Gn+7aJ+ivhrfR+uTLKi3/aW/jx/t1//\na/36tTOci8OSvKfv544kf5pk0cBx3pjkliRfT3Jm/1ou3ofX4rTBn4vZzlf/c3B937+vJ3nWTOdS\n2mtV5cNHsw/gBuD4/vljgWPpLsesBK4FXtJvWwJ8F/gv/fazgLuA58+w39cAF/TPVwIFvAM4GHgq\n8CPgI8CDgGXAzcAT+/anAbuAPwAOBJ4J3AE8oN/+aeBt/b5+GdgJPGnguHcBT6N7c3zIYC0D9Z0E\nPAwI8ETgh8Bj+m3H9cc/tz/+if32+/fbzwM+1de9CPgV4F798q19+wOAp/TLS6c5PwcCW+nC8iDg\nScD3gKOnnr8Zzu/LgS8BR/d9+CW6KXCAZwMP7F+ns4FvAQfPtF/gw8A7gUP71+P/Ab/bb3shsAVY\nDtwf+Mf+tVy8v6/Fns5XX8t3B87Hg4FfHPf/Lz4WzsORs+aNqrqyqi6vql1VdQPdL+wn9ptPBK6p\nqg9VN9X6l3S/9PfGa6vqR1V1GfAD4INVdXNV7QD+D/DogbY3A39RVXdV1d8A1wEn9aPDJwB/2O/r\ni8C7gecM/NvPVtVHquonVfWvM/T1kqr6WnX+CbgM+LWBJncB5/bH3wR8Hzg6yQHA7wBnVdWOqvq3\nqvpMVf2YLhQ3VdWm/tgfAzb3526qY4H7ABuq6s6q+gTwUeDUIc/l84E/rqrr+j5cVVW39n27oKpu\n7V/HN9K9cTh6up0kOaKv7yVV9YOquhl4M3BK3+Q3gbdU1faquh3YMPBv9/e1mO18/QR4ZJJDquqm\nqrpmyHMjzcpw1ryR5OFJPprkW/005OvoRswADwG27W5bVQVs38tDfHvg+b9Os3yfgeUd/TF2u7Gv\n4SHAbVX1vSnblg0sb2MWSU5Icnk/HfsdukBYMtDk1rr79d4f9vUtoRslfm2a3T4UeEY/Rfudfr+/\nSjfqm+ohwLaq+ske+rEnK2aogSQv66fs7+hrOGxK36bWfCBw00DN76QbQf97nQPtB5/v72sx4/mq\nqh/QzZi8sK/tkiS/sId9SXvFcNZ88nbgK8Cqqrof3ZRr+m030U1tApAkg8tzYFl/jN2OBL7ZPx6Q\n5L5Ttu0YWJ76p+DutpzkXsDfAX8OHFFVhwOb+Glf9+QWuin5h02zbRvw/qo6fOBxaFVtmKbtN4EV\n/Uh8pn7sybbpauivL7+CbsR7/75vd/DTvk09N9uAHwNLBmq+X1X9Yr/9bq873ZuCwT7s7Wsx9dgz\nnq+qurSqnkL35uYrwLv2sC9prxjOmk/uS3ed7/v9KOX3BrZdAvyHJE/rbwZ6EfBzc1jLg4AXJzkw\nyTOAR9BNgW4DPgP89yQHJ3kUcDpwwR729W1g5UAQHkQ31bsT2JXkBLrr4LPqR7rvBd7U3wy1KMnj\n+8C/APiNJP+5X39wf3PZdG9iPkc3Gn9F38fjgN8ALhymDrrp49cmWZXOo5I8kO413NX3bXGSc4D7\nzXQuquomuin9Nya5X5IDkjwsye7LGRcBZyVZluRw4A8HzsW+vBaDZjxfSY5Isi7JoXRvHr5PN80t\njYThrPnkZcBv0d2Y9C7gb3ZvqKpbgGcAf0Z3085quuuDP56jWj4HrKIbqf434Om7r6nSXZddSTdy\n+zDw6qr6xz3sa/eXeNya5PP9NOyL6YLndro+b9yL2l5GdzPWFcBtwOuBA/qwWkc347CTbmT4cqb5\nPVBVd9KF8Ql9H98GPKeqvjJkDW/q67+M7g3Ve+huuLoU+Afgq3RTzD/i7lPLdzsX/fPn0L1h2UJ3\nPi7mp1Px7+qPcTXwBboZhl3Av/Xb9/a1+HeznK8DgJf2+72N7t6H35t+T9Ley90vm0kLQz/y2g48\nq6o+OeJ9n0Z3F/ivjnK/2n/9LMM7quqh465F2h+OnLVg9NOPh/dTuLuvR18+5rI0h9J9pvvEJIuT\nLANeTTdCluY1w1kLyePp7hC+hW5K9mkzfVRJC0aAP6Gb7v4C3WffzxlrRdIIOK0tSVJjHDlLktQY\nw1mSpMaM7a+wLFmypFauXDmuw0uSdI+68sorb6mqpcO0HVs4r1y5ks2bN4/r8JIk3aOS3DhsW6e1\nJUlqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJasys4ZzkvUluTvLl\nGbYnyV8m2Zrk6iSPGX2ZkiRNjmFGzu8D1u5h+wnAqv5xBvD2/S9LkqTJNWs4V9Wngdv20GQd8NfV\nuRw4PMmDR1WgJEmTZhR/+GIZsG1geXu/7qapDZOcQTe65sgjjxzBoX9q5fpLRrq/3W7YcNKc7FeS\npJncozeEVdX5VbWmqtYsXTrUX82SJGnijCKcdwArBpaX9+skSdI+GEU4bwSe09+1fSxwR1X9zJS2\nJEkazqzXnJN8EDgOWJJkO/Bq4ECAqnoHsAk4EdgK/BB43lwVK0nSJJg1nKvq1Fm2F/CikVUkSdKE\n8xvCJElqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxn\nSZIaYzhLktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTG\nGM6SJDXGcJYkqTGGsyRJjTGcJUlqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMk\nSY0xnCVJaozhLElSYwxnSZIaYzhLktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMM\nZ0mSGmM4S5LUGMNZkqTGDBXOSdYmuS7J1iTrp9l+ZJJPJvlCkquTnDj6UiVJmgyzhnOSRcB5wAnA\nauDUJKunNPtj4KKqejRwCvC2URcqSdKkGGbkfAywtaqur6o7gQuBdVPaFHC//vlhwDdHV6IkSZNl\n8RBtlgHbBpa3A4+b0uY1wGVJfh84FDh+JNVJkjSBRnVD2KnA+6pqOXAi8P4kP7PvJGck2Zxk886d\nO0d0aEmSFpZhwnkHsGJgeXm/btDpwEUAVfVZ4GBgydQdVdX5VbWmqtYsXbp03yqWJGmBGyacrwBW\nJTkqyUF0N3xtnNLmG8CTAZI8gi6cHRpLkrQPZg3nqtoFnAlcClxLd1f2NUnOTXJy3+xs4AVJrgI+\nCJxWVTVXRUuStJANc0MYVbUJ2DRl3TkDz7cATxhtaZIkTSa/IUySpMYYzpIkNcZwliSpMYazJEmN\nMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTGGM6SJDXGcJYkqTGGsyRJjTGcJUlqjOEsSVJjDGdJ\nkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxnSZIaYzhLktQYw1mSpMYY\nzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTGGM6SJDXGcJYkqTGGsyRJ\njTGcJUlqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSY4YK\n5yRrk1yXZGuS9TO0+c0kW5Jck+R/jrZMSZImx+LZGiRZBJwHPAXYDlyRZGNVbRloswp4JfCEqro9\nyYPmqmBJkha6YUbOxwBbq+r6qroTuBBYN6XNC4Dzqup2gKq6ebRlSpI0OYYJ52XAtoHl7f26QQ8H\nHp7kn5NcnmTtdDtKckaSzUk279y5c98qliRpgRvVDWGLgVXAccCpwLuSHD61UVWdX1VrqmrN0qVL\nR3RoSZIWlmHCeQewYmB5eb9u0HZgY1XdVVVfB75KF9aSJGkvDRPOVwCrkhyV5CDgFGDjlDYfoRs1\nk2QJ3TT39SOsU5KkiTFrOFfVLuBM4FLgWuCiqromyblJTu6bXQrcmmQL8Eng5VV161wVLUnSQjbr\nR6kAqmoTsGnKunMGnhfw0v4hSZL2g98QJklSYwxnSZIaYzhLktQYw1mSpMYYzpIkNcZwliSpMYaz\nJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTGGM6SJDXGcJYkqTFD/clI/ayV6y+Zk/3esOGk\nOdmvJGn+cOQsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxn\nSZIaYzhLktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTG\nGM6SJDXGcJYkqTGGsyRJjTGcJUlqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxQ4Vz\nkrVJrkuyNcn6PbT7r0kqyZrRlShJ0mSZNZyTLALOA04AVgOnJlk9Tbv7AmcBnxt1kZIkTZJhRs7H\nAFur6vqquhO4EFg3TbvXAq8HfjTC+iRJmjiLh2izDNg2sLwdeNxggySPAVZU1SVJXj7C+gSsXH/J\nnOz3hg0nzcl+JUn7Z79vCEtyAPAm4Owh2p6RZHOSzTt37tzfQ0uStCANE847gBUDy8v7dbvdF3gk\n8KkkNwDHAhunuymsqs6vqjVVtWbp0qX7XrUkSQvYMOF8BbAqyVFJDgJOATbu3lhVd1TVkqpaWVUr\ngcuBk6tq85xULEnSAjdrOFfVLuBM4FLgWuCiqromyblJTp7rAiVJmjTD3BBGVW0CNk1Zd84MbY/b\n/7IkSZpcfkOYJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTGGM6SJDXGcJYkqTGGsyRJjTGc\nJUlqjOEsSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxnSZIa\nYzhLktQYw1mSpMYYzpIkNcZwliSpMYvHXYDas3L9JXOy3xs2nDQn+5WkhcaRsyRJjTGcJUlqjOEs\nSVJjDGdJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxnSZIaYzhLktQY\nw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUmKHCOcnaJNcl2Zpk/TTb\nX5pkS5Krk3w8yUNHX6okSZNh1nBOsgg4DzgBWA2cmmT1lGZfANZU1aOAi4E/G3WhkiRNisVDtDkG\n2FpV1wMkuRBYB2zZ3aCqPjnQ/nLg2aMsUgvXyvWXzMl+b9hw0pzsV5LuCcNMay8Dtg0sb+/XzeR0\n4O+n25DkjCSbk2zeuXPn8FVKkjRBRnpDWJJnA2uAN0y3varOr6o1VbVm6dKlozy0JEkLxjDT2juA\nFQPLy/t1d5PkeOCPgCdW1Y9HU54kSZNnmJHzFcCqJEclOQg4Bdg42CDJo4F3AidX1c2jL1OSpMkx\nazhX1S7gTOBS4Frgoqq6Jsm5SU7um70BuA/wt0m+mGTjDLuTJEmzGGZam6raBGyasu6cgefHj7gu\nSZImlt8QJklSYwxnSZIaYzhLktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWrMUF9CIi0E/nlK\nSfOFI2dJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxnSZIaYzhLktQY\nw1mSpMYYzpIkNcY/fCHNEf/QhqR95chZkqTGGM6SJDXGcJYkqTGGsyRJjTGcJUlqjOEsSVJjDGdJ\nkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSY/yTkZI0AfwTpvOLI2dJkhrj\nyFlaABwVSQuLI2dJkhpjOEuS1BjDWZKkxhjOkiQ1xnCWJKkxhrMkSY0Z6qNUSdYCbwEWAe+uqg1T\ntt8L+GvgscCtwDOr6obRliqpFX50S5pbs46ckywCzgNOAFYDpyZZPaXZ6cDtVfXzwJuB14+6UEmS\nJsUwI+djgK1VdT1AkguBdcCWgTbrgNf0zy8G3pokVVUjrFWSFhRnIDSTYcJ5GbBtYHk78LiZ2lTV\nriR3AA8EbhlFkZImlwGmSZTZBrdJng6srarn98u/DTyuqs4caPPlvs32fvlrfZtbpuzrDOAMgCOP\nPPKxN9544yj7Ikn7zTcDo3FPnsf58polubKq1gzTdpi7tXcAKwaWl/frpm2TZDFwGN2NYXdTVedX\n1ZqqWrN06dJh6pMkaeIMM619BbAqyVF0IXwK8FtT2mwEngt8Fng68AmvN0uajyZthKs2zRrO/TXk\nM4FL6T5K9d6quibJucDmqtoIvAd4f5KtwG10AS5JkvbBUJ9zrqpNwKYp684ZeP4j4BmjLU2SpMnk\nN4RJktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGjPUl5BIkrQ3/BrU/ePI\nWZKkxhjOkiQ1xnCWJKkxhrMkSY0xnCVJaozhLElSYwxnSZIaYzhLktQYw1mSpMYYzpIkNcav75Qk\nzWsL8atCHTlLktQYw1mSpMYYzpIkNcZwliSpMYazJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZ\nkqTGGM6SJDXGcJYkqTGGsyRJjTGcJUlqTKpqPAdOdgI3juXgsAS4ZUzHnmv2bX6yb/PTQu3bQu0X\njLdvD62qpcM0HFs4j1OSzVW1Ztx1zAX7Nj/Zt/lpofZtofYL5k/fnNaWJKkxhrMkSY2Z1HA+f9wF\nzCH7Nj/Zt/lpofZtofYL5knfJvKasyRJLZvUkbMkSc2auHBOsjbJdUm2Jlk/7npGJcmKJJ9MsiXJ\nNUnOGndNo5RkUZIvJPnouGsZpSSHJ7k4yVeSXJvk8eOuaVSS/EH/s/jlJB9McvC4a9pXSd6b5OYk\nXx5Y94AkH0vyL/1/7z/OGvfVDH17Q/8zeXWSDyc5fJw17qvp+jaw7ewklWTJOGqbzUSFc5JFwHnA\nCcBq4NQkq8db1cjsAs6uqtXAscCLFlDfAM4Crh13EXPgLcA/VNUvAL/EAuljkmXAi4E1VfVIYBFw\nynir2i/vA9ZOWbce+HhVrQI+3i/PR+/jZ/v2MeCRVfUo4KvAK+/pokbkffxs30iyAngq8I17uqBh\nTVQ4A8cAW6vq+qq6E7gQWDfmmkaiqm6qqs/3z79H90t+2XirGo0ky4GTgHePu5ZRSnIY8J+A9wBU\n1Z1V9Z3xVjVSi4FDkiwG7g18c8z17LOq+jRw25TV64C/6p//FfC0e7SoEZmub1V1WVXt6hcvB5bf\n44WNwAyvG8CbgVcAzd50NWnhvAzYNrC8nQUSYIOSrAQeDXxuvJWMzF/Q/Y/0k3EXMmJHATuB/9FP\n2b87yaHjLmoUqmoH8Od0I5ObgDuq6rLxVjVyR1TVTf3zbwFHjLOYOfQ7wN+Pu4hRSbIO2FFVV427\nlj2ZtHBe8JLcB/g74CVV9d1x17O/kvw6cHNVXTnuWubAYuAxwNur6tHAD5i/U6N3019/XUf3BuQh\nwKFJnj3equZOdR97aXYUtq+S/BHdJbMPjLuWUUhyb+BVwDnjrmU2kxbOO4AVA8vL+3ULQpID6YL5\nA1X1oXHXMyJPAE5OcgPdZYgnJblgvCWNzHZge1XtnuG4mC6sF4Ljga9X1c6qugv4EPArY65p1L6d\n5MEA/X9vHnM9I5XkNODXgWfVwvnM7cPo3jBe1f9OWQ58PsnPjbWqaUxaOF8BrEpyVJKD6G5Q2Tjm\nmkYiSeiuXV5bVW8adz2jUlWvrKrlVbWS7vX6RFUtiBFYVX0L2Jbk6H7Vk4EtYyxplL4BHJvk3v3P\n5pNZIDe7DdgIPLd//lzgf42xlpFKspbuUtLJVfXDcdczKlX1pap6UFWt7H+nbAce0/+/2JSJCuf+\nBoczgUvpflFcVFXXjLeqkXkC8Nt0I8sv9o8Tx12UZvX7wAeSXA38MvC6MdczEv1swMXA54Ev0f2u\nmRffzDSdJB8EPgscnWR7ktOBDcBTkvwL3UzBhnHWuK9m6NtbgfsCH+t/l7xjrEXuoxn6Ni/4DWGS\nJDVmokbOkiTNB4azJEmNMZwlSWqM4SxJUmMMZ0mSGmM4S5LUGMNZkqTGGM6SJDXm/wMfkc+0z4kG\n4gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "avgs = np.mean(categories_acf, axis = 0)\n",
    "plt.bar(x = np.arange(time_shift + 1), height = avgs)\n",
    "plt.title('lag importance of categories')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "F9fOeQS4SieH"
   },
   "outputs": [],
   "source": [
    "category_lags = [1, 2, 3, 4, 5, 6, 12]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yB2wn7wuyU4y"
   },
   "source": [
    "## Create training grid and add lags to it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 24710,
     "status": "ok",
     "timestamp": 1568521726829,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "o5M66Xl2q4xr",
    "outputId": "77104d97-a34c-40c9-d59a-826cae25faf7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sales data shape: (2935842, 7)\n",
      "Generated Train data shape: (10809347, 4)\n",
      "23 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid = create_train_set(sales, index_cols)\n",
    "test['date_block_num'] = 34\n",
    "grid = pd.concat([grid, test], ignore_index = True, sort = False, keys = index_cols)\n",
    "grid = grid.merge(items[['item_id', 'item_category_id']], on = 'item_id', how = 'left')\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ejt9xpsAGVrI"
   },
   "outputs": [],
   "source": [
    "def lag_ex_features(grid, features, features_sales, features_lags):\n",
    "    df = grid.copy()\n",
    "    for feature_lag in features_lags:\n",
    "        max_month = df.date_block_num.max()\n",
    "        cols = [str(itr) for itr in np.arange(0, max_month)]\n",
    "        gb = features_sales.melt( id_vars = features, \n",
    "                                 var_name = 'date_block_num' , \n",
    "                                 value_vars= cols, \n",
    "                                 value_name = '_'.join(features) + '_lag_' + str(feature_lag)\n",
    "                                )\n",
    "        gb.date_block_num = gb.date_block_num.astype(np.int16)\n",
    "        gb.date_block_num = gb.date_block_num + feature_lag\n",
    "        df = pd.merge(df, gb, on = features + ['date_block_num'], how='left')\n",
    "    return df\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "11MFMpzJAhrP"
   },
   "outputs": [],
   "source": [
    "def lag_features(grid, features, features_lags):\n",
    "    df = grid.copy()\n",
    "    for month_shift in features_lags:\n",
    "        df_shift = df[index_cols + features].copy()\n",
    "        df_shift['date_block_num'] = df_shift['date_block_num'] + month_shift\n",
    "        lag_cols = lambda x: '{}_lag_{}'.format(x, month_shift) if x in features else x\n",
    "        df_shift = df_shift.rename(columns = lag_cols)\n",
    "        df = pd.merge(df, df_shift, on = index_cols, how='left')\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 74073,
     "status": "ok",
     "timestamp": 1568521812233,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "3CwCg8EvAZIR",
    "outputId": "82f32914-7534-41d3-f0fb-089656e8e432"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "73 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid = lag_features(grid, ['target'], pair_lags)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 91630,
     "status": "ok",
     "timestamp": 1568521834792,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "FxyzeSe2IOnA",
    "outputId": "446b106a-16c7-47e5-ae73-921940d2d78e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "23 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid = lag_ex_features(grid, ['item_id'], items_sales, item_lags)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 105602,
     "status": "ok",
     "timestamp": 1568521851916,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "89FfHZc4JGv4",
    "outputId": "06717837-cc17-400c-83f8-6fd9de8a3a1f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "17 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid = lag_ex_features(grid, ['shop_id'], shops_sales, shop_lags)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 119772,
     "status": "ok",
     "timestamp": 1568521869206,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "QSVi7f5udcez",
    "outputId": "161106e2-dfb6-4e0e-8a90-b4f74f019967"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "17 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid = lag_ex_features(grid, ['item_category_id'], categories_sales, category_lags)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 646
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 125693,
     "status": "ok",
     "timestamp": 1568521878302,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "-ogxRwusvVzE",
    "outputId": "96aa4a19-9837-4820-ef96-64b71db4ba70"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "shop_id                          0\n",
       "item_id                          0\n",
       "date_block_num                   0\n",
       "target                      214200\n",
       "item_category_id                 0\n",
       "target_lag_1               2466770\n",
       "target_lag_2               3091897\n",
       "target_lag_3               3688208\n",
       "target_lag_4               4253668\n",
       "target_lag_9               6724591\n",
       "target_lag_10              7129579\n",
       "target_lag_11              7524904\n",
       "target_lag_12              7884005\n",
       "item_id_lag_1               372306\n",
       "item_id_lag_2               739866\n",
       "item_id_lag_3              1113456\n",
       "item_id_lag_4              1479981\n",
       "item_id_lag_5              1855236\n",
       "item_id_lag_6              2245224\n",
       "item_id_lag_11             4081468\n",
       "item_id_lag_12             4471226\n",
       "shop_id_lag_1               357060\n",
       "shop_id_lag_2               724620\n",
       "shop_id_lag_3              1098210\n",
       "shop_id_lag_4              1464735\n",
       "shop_id_lag_5              1839990\n",
       "shop_id_lag_6              2229978\n",
       "shop_id_lag_9              3332423\n",
       "shop_id_lag_12             4455980\n",
       "item_category_id_lag_1      357060\n",
       "item_category_id_lag_2      724620\n",
       "item_category_id_lag_3     1098210\n",
       "item_category_id_lag_4     1464735\n",
       "item_category_id_lag_5     1839990\n",
       "item_category_id_lag_6     2229978\n",
       "item_category_id_lag_12    4455980\n",
       "dtype: int64"
      ]
     },
     "execution_count": 93,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "IBhhM13REqY0"
   },
   "outputs": [],
   "source": [
    "grid = grid[grid['date_block_num'] > 11]\n",
    "grid = grid.fillna(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "umz5GWIvhWWo"
   },
   "source": [
    "## Item Price"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 461
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1758,
     "status": "ok",
     "timestamp": 1568521948991,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "E5mjJRethVYP",
    "outputId": "2168a371-9ff2-4706-eabf-05cc65719a4d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFpCAYAAABTSWtMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xt4VPW59//3nRAIZyHEikQFJSog\nECRSFE9gFU9bbH/uZ2PVVq2Xj1p3wbYqttVdeepPbK1WRbS2HitWrO5HKGo9YquWrQUJCCImKEoA\nFRIBOQRIcj9/zEocyJxymKxJ8nld11yZWeu71rrXfDNzz1rrO/eYuyMiIiKZKSvsAERERCQ+JWoR\nEZEMpkQtIiKSwZSoRUREMpgStYiISAZTohYREclgStQiIiIZTIlaREQkgylRi4iIZDAlahERkQzW\nKewAAPr16+cDBw4MOwwREZFWs3jx4k3unp+sXUYk6oEDB7Jo0aKwwxAREWk1ZvZJKu106ltERCSD\nKVGLiIhkMCVqERGRDJYR16hFRDqiPXv2UF5eTlVVVdihSBrl5uZSUFBATk5Ok5ZXohYRCUl5eTk9\ne/Zk4MCBmFnY4UgauDsVFRWUl5czaNCgJq1Dp75FREJSVVVFXl6eknQ7Zmbk5eU166yJErWISIiU\npNu/5vZxyonazLLNbImZzQ8eP2JmH5tZSXArCqabmd1tZmVmtszMjm5WhCIikjZ33303Q4YM4YIL\nLmDevHnMmDEDgF/+8pfcfvvtAFx88cU8/fTTKa9zzZo1HHXUUQ2mr1+/nvPOOw+AkpISnn/++RbY\ng6bbunUrBQUFXH311fXTTj/9dEaOHMmwYcO44oorqKmpAeDGG29kxIgRFBUVcdppp7F+/XoAZs+e\nzYgRIxg+fDjHHXccS5cubfE4G3NEPQVYuc+0a929KLiVBNPOAAqD2+XAfc0PU0RE0mHWrFm8/PLL\nzJ49m3POOYdp06albVsHHnhgfcLPhER94403cuKJJ+417amnnmLp0qUsX76cjRs38pe//AWAa6+9\nlmXLllFSUsLZZ5/N9OnTARg0aBB///vfee+997jxxhu5/PLLWzzOlBK1mRUAZwF/TKH5JOAxj/gf\nYD8z69+MGEVEJA2uuOIKPvroI8444wzuvPNOHnnkkb2OLmNZvHgxJ510EqNHj2bixIls2LChfvrI\nkSMZOXIk9957b8xl6460d+/ezU033cScOXMoKipizpw5bN++nUsvvZQxY8YwatQo5s6dC8AjjzzC\nueeey6mnnsrAgQOZOXMmd9xxB6NGjWLs2LFUVlY2ad8XL17M559/zmmnnbbX9F69egFQXV3N7t27\n609b100H2L59e/304447jj59+gAwduxYysvLmxRPIqmO+v4dcB3Qc5/pt5jZTcCrwDR33wUMANZG\ntSkPpm1oZqwiIu3X1KlQUpK8XWMUFcHvfhd39v3338/f/vY3FixYQL9+/XjkkUcSrm7Pnj3853/+\nJ3PnziU/P585c+bw85//nIceeohLLrmEmTNncuKJJ3LttdcmXE/nzp2ZPn06ixYtYubMmQD87Gc/\nY8KECTz00ENs3ryZMWPG8K1vfQuA5cuXs2TJEqqqqhg8eDC33XYbS5Ys4ZprruGxxx5j6tSpe61/\n9uzZ/OY3v2mw3cGDB/P0009TW1vLT37yEx5//HFeeeWVBu0mTpzIO++8wxlnnFF/qh7g5z//OY89\n9hi9e/dmwYIFDZZ78MEHOeOMMxLue1MkTdRmdjbwhbsvNrOTo2bdAHwGdAYeAK4Hpqe6YTO7nMip\ncQ4++OBGhJzYzi1f8T9/faPF1iciEk/P3t0ZfdYJWFbHGJe7atUqli9fzqmnngpATU0N/fv3Z/Pm\nzWzevLn+NPJFF13ECy+80Kh1v/TSS8ybN6/+unhVVRWffvopAOPHj6dnz5707NmT3r1782//9m8A\nDB8+nGXLljVY1wUXXMAFF1wQd1uzZs3izDPPpKCgIOb8F198kaqqKi644AJee+21+v295ZZbuOWW\nW7j11luZOXMmN998c/0yCxYs4MEHH+TNN99s1H6nIpUj6nHAOWZ2JpAL9DKzx939wmD+LjN7GPhp\n8HgdcFDU8gXBtL24+wNEEjzFxcXexPgb2PjxOi5Z3mKrExFJYBvzurzNiNOObf6qEhz5Zgp3Z9iw\nYSxcuHCv6Zs3b26RdT/zzDMcccQRe01/++236dKlS/3jrKys+sdZWVlUV1c3WFeyI+qFCxfyxhtv\nMGvWLLZt28bu3bvp0aNH/UA6iBQpmTRpEnPnzq1P1HUuuOACzjzzzPpEvWzZMi677DJeeOEF8vLy\nmv4kxJE0Ubv7DUSOngmOqH/q7heaWX9332CRE/XnAsuDReYBV5vZk8A3gS3u3mqnvfc/7CD+70mb\nWmtzItJBbfyskstXZbNy+cctk6jbgCOOOIKNGzeycOFCjj32WPbs2cOHH37IsGHD2G+//XjzzTc5\n/vjjmT17dtJ19ezZk6+++qr+8cSJE7nnnnu45557MDOWLFnCqFGjmhRnsiPq6PgeeeQRFi1axIwZ\nM9i2bRtfffUV/fv3p7q6mueee44TTjgBgNLSUgoLCwGYO3cuRx55JACffvop3/nOd/jTn/7E4Ycf\n3qR4k2lOZbLZZpYPGFACXBFMfx44EygDdgCXNCvCRsrt2Z1RZxzfmpsUkQ6oZk81XW6YR+mGLWGH\n0mo6d+7M008/zY9+9CO2bNlCdXU1U6dOZdiwYTz88MNceumlmFmDAVqxjB8/nhkzZlBUVMQNN9zA\njTfeyNSpUxkxYgS1tbUMGjSI+fPnt8JefW379u2cc8457Nq1i9raWsaPH88VV0RS27Rp01i1ahVZ\nWVkccsgh3H///QBMnz6diooKrrrqKgA6derU4j/bbO7hnyYuLi52/R61iLQ1Z171B/LZw6OzrmrS\n8itXrmTIkCEtHJVkolh9bWaL3b042bIdYwSEiEgaFGbvpiy7R9hhSDunRC0i0kSFfbqwrnse2ys7\nzulvaX1K1CIiTTT4kH4ArF70fsiRSHumRC0i0kSDhx0KQOkHn4QcibRnStQiIk10SNER5NTsoXR9\n879HLBJPc76eJSLSoeXkdmHQ9grKrCbsUKQd0xG1iEgzFGZVUWrdww6jyY477jgg8oMZTzzxRKix\n1NTUMGrUKM4+++z6aT/4wQ8YOXIkI0aM4LzzzmPbtm1ApE758OHDKSoq4vjjj+f99yPjBF5++WVG\njx7N8OHDGT16NK+99loo+9KSlKhFRJphcO8c1nbrS9XWbWGH0iT//Oc/gcxI1HfddVeD7xrfeeed\nLF26lGXLlnHwwQfX/4jHd7/7Xd577z1KSkq47rrr+PGPfwxAv379+Otf/8p7773Ho48+ykUXXdTq\n+9HSlKhFRJqh8OA8arOy+ejdlWGH0iQ9ekS+Bz5t2jTeeOMNioqKuPPOO6mpqeHaa6/lmGOOYcSI\nEfz+978H4PXXX+ekk05i0qRJHHrooUybNo3Zs2czZswYhg8fzurVq5sUR3l5Oc899xyXXXbZXtPr\nfl7S3dm5c2fSn50cNWoUBx54IADDhg1j586d7Nq1q0kxZQpdoxYRaYbCYYPg4w2Uvr+GoScf0+T1\n3PzXFby/fmsLRgZDD+zFf/3bsJTazpgxg9tvv72+bOcDDzxA7969+de//sWuXbsYN25cfWnQpUuX\nsnLlSvr27cuhhx7KZZddxjvvvMNdd93FPffcw+/2+YGRBQsWcM011zTYZrdu3eqP6KdOncqvf/3r\nvep/17nkkkt4/vnnGTp0KL/97W/rp997773ccccd7N69O+Yp7meeeYajjz56rx/1aIt0RC0i0gwD\nRw0hq7aGsvLKsENpUS+99BKPPfYYRUVFfPOb36SiooLS0lIAjjnmGPr370+XLl047LDD6hP48OHD\nWbNmTYN1jR8/npKSkga3uiQ9f/589t9/f0aPHh0zlocffpj169czZMgQ5syZUz/9hz/8IatXr+a2\n227jV7/61V7LrFixguuvv77+TEBbpiNqEZFm6NK9KwO3V1BqDX9usTFSPfJtLe7OPffcw8SJE/ea\n/vrrrzf6ZyeTHVG/9dZbzJs3j+eff56qqiq2bt3KhRdeyOOPP17fNjs7m8mTJ/PrX/+aSy7Z+7ee\nJk+ezJVXXln/uLy8nG9/+9s89thjHHbYYU17AjKIErWISDMNZgel3jXsMJol1s9O3nfffUyYMIGc\nnBw+/PBDBgwY0KR11x1Rx3Prrbdy6623ApEPArfffjuPP/447s7q1asZPHgw7s68efPqf14y+mcn\nn3vuufr7mzdv5qyzzmLGjBmMGzeuSfFmGiVqEZFmKuzViVdr+rF7RxWdu+WGHU6TjBgxguzsbEaO\nHMnFF1/MlClTWLNmDUcffTTuTn5+Ps8++2yrxuTufP/732fr1q24OyNHjuS+++4DYObMmbzyyivk\n5OTQp08fHn300frpZWVlTJ8+nenTpwOR0/j7779/q8bekvQzlyIizfTszKeYWt6dl88ZQOFxRSkv\np5+57Dj0M5ciIiEaPHQgAKXLPwo3EGmXlKhFRJrpsNFDMK+ldG1F2KFIO6Rr1CIizdS1d08KtldS\numN32KFIO6RELSLSAgprt1FG2y6sIZlJp75FRFpAYc9sPurWj+pdOqqWlqVELSLSAgb3783uTjms\nfa807FCknVGiFhFpAYVHHgxA6XtN+1GKMGzevJlZs2albf11P/ghzaNELSLSAgYXDwWgdM0XIUeS\nuniJOlYZUAmPBpOJiLSAHv360H97JWU72s5PKk6bNo3Vq1dTVFRETk4Oubm59OnThw8++IAPP/yQ\nxx9/nLvvvpvdu3fzzW9+k1mzZpGdnU2PHj2YMmUK8+fPp2vXrsydO5dvfOMbfPzxx3z3u99l27Zt\nTJo0KezdazeUqEVEWsjgmq8opXPTFn5hGnz2XssGdMBwOGNG3NkzZsxg+fLllJSU8Prrr3PWWWex\nfPlyBg0axMqVK5kzZw5vvfUWOTk5XHXVVcyePZvvfe97bN++nbFjx3LLLbdw3XXX8Yc//IFf/OIX\nTJkyhSuvvJLvfe973HvvvS27Lx2YTn2LiLSQwm5GWdc8aqtrwg6lScaMGcOgQYMAePXVV1m8eDHH\nHHMMRUVFvPrqq3z0UaTyWufOnTn77LMBGD16dP1PW7711lucf/75AFx00UWtvwPtlI6oRURaSGH/\nXlRt7MK6FWUcNPKIxi2c4Mi3tXTv3r3+ft0PYtT9qlW0nJwczAyI/Pxk9DXtuunScnRELSLSQgoP\nLwCg7L2ykCNJzb4/bRntlFNO4emnn+aLLyKD4yorK/nkk08Srm/cuHE8+eSTAMyePbtlg+3AUk7U\nZpZtZkvMbH7weLaZrTKz5Wb2kJnlBNNPNrMtZlYS3G5KV/AiIplk8NHBbyV/9HnIkaQmLy+PcePG\ncdRRR3HttdfuNW/o0KH86le/4rTTTmPEiBGceuqpbNiwIeH67rrrLu69916GDx/OunXr0hl6h5Ly\nz1ya2Y+BYqCXu59tZmcCLwSznwD+4e73mdnJwE/d/exUg9DPXIpIe1E85QnG8yW/ueuHSdvqZy47\njrT/zKWZFQBnAX+sm+buz3sAeAcoaFTUIiLtUOGeLZTu0fAfaTmpnvr+HXAdULvvjOCU90XA36Im\nH2tmS83sBTMb1vwwRUTahsJcpyy3L17b4O1SpEmSJmozOxv4wt0Xx2kyi8hp7zeCx+8Ch7j7SOAe\n4Nk4673czBaZ2aKNGzc2IXQRkcxT+I0ebOvcjc8+XBN2KNJOpHJEPQ44x8zWAE8CE8zscQAz+y8g\nH/hxXWN33+ru24L7zwM5ZtZv35W6+wPuXuzuxfn5+c3fExGRDDB48AAAykpS+3GOVMcJSdvV3D5O\nmqjd/QZ3L3D3gcBk4DV3v9DMLgMmAue7e/05HjM7wIIv0pnZmGAbFc2KUkSkjSgcXTfyO/EIaYDc\n3FwqKiqUrNsxd6eiooLc3Nwmr6M5Ix7uBz4BFgZ5+b/dfTpwHnClmVUDO4HJrv9CEekg8g7uT5+q\nNyjdtT1p24KCAsrLy9Hlv/YtNzeXgoKmj7duVKJ299eB14P7MZd195nAzCZHJCLShllWFoN3baaM\n7KRtc3Jy6kt2isSjymQiIi1scG4NH3bpo5Hf0iKUqEVEWlhhfne25PZg05r1YYci7YAStYhICys8\ntD8Ape9+EHIk0h4oUYuItLDCUYcDsHq1jqil+ZSoRURa2DcKD6Hnrh2Ufr4t7FAkltrayK2NUKIW\nEWlhlpXFYbsqKa3SbzNnpH/eDX+eDDsqw44kJUrUIiJpUJhTTWlO77DDkFi69ICPFsD9J8Dad8KO\nJiklahGRNCjM68qmrr35svyzsEORfR1zGVz6ImRlw8NnwD/vgQyuy6VELSKSBoWHfgOAMo38zkwD\njob//Q84/HR46Rfw5/Mz9lS4ErWISBoMHlEIQFnpupAjkbi67gf/8TicPgPKXoHfnwTli8KOqgEl\nahGRNBgw9DC67qmidMPWsEORRMxg7JWRU+EAD50OC2dl1KlwJWoRkTTI6pTN4J0VlO7InDd8SaBg\nNFzxDyg8FV68AeZcCDs3hx0VoEQtIpI2gzvtoSy7Z9hhSKq69oHJT8Bpt8CHf4Pfnwjr3g07KiVq\nEZF0GdynCxu69+WrjZk5SEliMIPjroZLXoDaGnhoIrz9QKinwpWoRUTSpHDg/gCULXo/5Eik0Q4a\nA1e8AYeOhxeuhb9cDFVbQglFiVpEJE0Khx8GQOmqtSFHIk3SrS+c/yScOh1KX4YvPwklDCVqEZE0\nOWh4IZ2rd7N6QzhHYtICsrJg3BSY+h70HxFKCJ1C2aqISAfQqUtnDt1RQSk1YYcizdU9L7RN64ha\nRCSNBmfvojSrR9hhSBumRC0ikkaF+3WmvHtfdnypwifSNErUIiJpVHhQHm5ZfPTuyrBDkTZKiVpE\nJI0KjzoUgNKV4YwYlrZPiVpEJI0OKTqSTjXVlK37MuxQpI3SqG8RkTTq3C2XgTsqKKU67FCkjdIR\ntYhImhXaTsroFnYY0kYpUYuIpNngXp1Y0z2PXdt3hh2KtEFK1CIiaTa4oC+1Wdl8vFg1v6XxlKhF\nRNKscOhAAErfXxNqHNI2pZyozSzbzJaY2fzg8SAze9vMysxsjpl1DqZ3CR6XBfMHpid0EZG24dCj\nh5BVW0Pp2oqwQ5E2qDFH1FOA6G/s3wbc6e6DgS+BHwTTfwB8GUy/M2gnItJh5fbqwcE7Klm9ZU/Y\noUgblFKiNrMC4Czgj8FjAyYATwdNHgXODe5PCh4TzD8laC8i0mEN9u2U1uaGHYa0QakeUf8OuA6o\nDR7nAZvdve6LgeXAgOD+AGAtQDB/S9BeRKTDGtwzm4+757GnalfYoUgbkzRRm9nZwBfuvrglN2xm\nl5vZIjNbtHHjxpZctYhIxik8cD/2ZOfwScmqsEORNiaVymTjgHPM7EwgF+gF3AXsZ2adgqPmAmBd\n0H4dcBBQbmadgN5AgxEU7v4A8ABAcXGxN3dHREQyWeGRh8D6SlYsXsUBgw8KO5x62Tmd6Nq7Z9hh\nSAJJE7W73wDcAGBmJwM/dfcLzOwvwHnAk8D3gbnBIvOCxwuD+a+5uxKxiHRohxUPJeuVvzNlbTe4\n/Z9hh1PPvJaHhmcx/sKzwg5F4mhOre/rgSfN7FfAEuDBYPqDwJ/MrAyoBCY3L0QRkbave9/e3D8E\nPin/IuxQ6lXX1nLbzgP4YPVnjA87GImrUYna3V8HXg/ufwSMidGmCvj3FohNRKRdOe2Sc8IOoYG7\nf/IMFTW7ww5DEtCvZ4mIdGB9d2+n0mqTN5TQKFGLiHRgeTVVVKBSF5lMtb5FRDqwPKqpICfsMCQB\nJWoRkQ6sb45Tma2KaZlMiVpEpAPL65JFRZfueK2uU2cqJWoRkQ4sr0cXdnXqwvbKLWGHInEoUYuI\ndGB9e3UFoLL885AjkXiUqEVEOrC8vF4AVGzYFHIkEo8StYhIB5a3fx8AKr74MuRIJB4lahGRDqxv\n/34AVFZ+FXIkEo8StYhIB5Z30AEAVGzdGXIkEo8qk4mIdGDd+vSi654q1fvOYErUIiIdnOp9ZzYl\nahGRDk71vjObrlGLiHRwqved2ZSoRUQ6ONX7zmxK1CIiHZzqfWc2JWoRkQ5O9b4zmxK1iEgHp3rf\nmU2JWkSkg6ur971J9b4zkhK1iEgHV1fvu1L1vjOSErWISAenet+ZTYlaRKSDq6v3vWmL6n1nIlUm\nExHp4OrqfVeq3ndGUqIWERHV+85gStQiIkK/mio2qd53RtI1ahERoS/VVKred0ZSohYREdX7zmBJ\nE7WZ5ZrZO2a21MxWmNnNwfQ3zKwkuK03s2eD6Seb2ZaoeTeleydERKR5+uVms6lLD9X7zkCpXKPe\nBUxw921mlgO8aWYvuPsJdQ3M7BlgbtQyb7j72S0cq4iIpEnf7p3ZvaMz2yu30KNfn7DDkShJj6g9\nYlvwMCe4ed18M+sFTACeTUuEIiKSdqr3nblSukZtZtlmVgJ8Abzs7m9HzT4XeNXdt0ZNOzY4Vf6C\nmQ1rwXhFRCQN+vUL6n2v3xhyJLKvlBK1u9e4exFQAIwxs6OiZp8P/Dnq8bvAIe4+EriHOEfaZna5\nmS0ys0UbN+ofQ0QkTH3zg3rfGzeHHInsq1Gjvt19M7AAOB3AzPoBY4DnotpsrTtV7u7PAzlBu33X\n9YC7F7t7cX5+fjN2QUREmkv1vjNXKqO+881sv+B+V+BU4INg9nnAfHevimp/gJlZcH9MsI2Klg5c\nRERajup9Z65URn33Bx41s2wiSfcpd58fzJsMzNin/XnAlWZWDewEJru7IyIiGUv1vjNX0kTt7suA\nUXHmnRxj2kxgZrMjExGRVqV635lJtb5FRARQve9MpRKiIiICqN53plKiFhERQPW+M5UStYiIAKr3\nnamUqEVEBAjqfXeK1PuWzKFELSIiwNf1vitU7zujKFGLiAjwdb3vCtX7zihK1CIiAqjed6ZSohYR\nEeDret8VqvedUZSoRUQE+Lred4XqfWcUVSYTERFA9b4zlRK1iIjU67t7OxWq951RlKhFRKRev5oq\nKlTvO6PoGrWIiNRTve/Mo0QtIiL18nKcCtX7zihK1CIiUi8vN5sK1fvOKErUIiJST/W+M48StYiI\n1Mvr3Q2AirWfhRyJ1FGiFhGRenl5PQGo2LAp5EikjhK1iIjUU73vzKNELSIi9fIG5ANQUaF635lC\niVpEROrV1/veqnrfmUKVyUREpF7X3j1V7zvDKFGLiMhe8nZtoyJL36POFErUIiKyl7zaXVTUqt53\nptA1ahER2YvqfWcWJWoREdmL6n1nFiVqERHZi+p9Z5akidrMcs3sHTNbamYrzOzmYPojZvaxmZUE\nt6JgupnZ3WZWZmbLzOzodO+EiIi0HNX7ziypDCbbBUxw921mlgO8aWYvBPOudfen92l/BlAY3L4J\n3Bf8FRGRNiCvdzfYEan33aNfn7DD6fCSHlF7xLbgYU5w8wSLTAIeC5b7H2A/M+vf/FBFRKQ1qN53\nZknpGrWZZZtZCfAF8LK7vx3MuiU4vX2nmXUJpg0A1kYtXh5MExGRNqCu3nfFF6r3nQlSStTuXuPu\nRUABMMbMjgJuAI4EjgH6Atc3ZsNmdrmZLTKzRRs3bmxk2CIiki519b4rK1XvOxM0atS3u28GFgCn\nu/uG4PT2LuBhYEzQbB1wUNRiBcG0fdf1gLsXu3txfn5+06IXEZEWp3rfmSWVUd/5ZrZfcL8rcCrw\nQd11ZzMz4FxgebDIPOB7wejvscAWd9+QluhFRKTF1dX7rtiuet+ZIJVR3/2BR80sm0hif8rd55vZ\na2aWDxhQAlwRtH8eOBMoA3YAl7R82CIikk55u7ZRqXrfGSFponb3ZcCoGNMnxGnvwA+bH5qIiIRF\n9b4zhyqTiYhIA32ppkL1vjOCErWIiDSQl+NUqt53RlCiFhGRBlTvO3MoUYuISAN19b63qd536JSo\nRUSkgbze3QCoXPtZyJGIErWIiDSget+ZQ4laREQayNu/LwAVX3wZciSiRC0iIg30PbAfAJWV25K0\nlHRTohYRkQZU7ztzpFJCVEREOpiuvXvSbXcVFTWq9x02JWoREYmp727V+84EStQiIhKT6n1nBl2j\nFhGRmPLYo3rfGUCJWkREYuqbg+p9ZwAlahERiUn1vjODErWIiMSUV1fvu2Jz2KF0aErUIiISU9+6\net/ln4ccScemRC0iIjHV1fvepHrfoVKiFhGRmOrqfVeq3neolKhFRCQm1fvODErUIiISU129702q\n9x0qVSYTEZGY6up9V6red6iUqEVEJC7V+w6fErWIiMSVV7uLTar3HSpdoxYRkbjy2EOl6n2HSola\nRETiUr3v8ClRi4hIXJF63z1V7ztESRO1meWa2TtmttTMVpjZzcH02Wa2ysyWm9lDZpYTTD/ZzLaY\nWUlwuyndOyEiIukRqfedo3rfIUrliHoXMMHdRwJFwOlmNhaYDRwJDAe6ApdFLfOGuxcFt+ktHbSI\niLQO1fsOX9JE7RF1ZWlygpu7+/PBPAfeAQrSGKeIiIRA9b7Dl9I1ajPLNrMS4AvgZXd/O2peDnAR\n8LeoRY4NTpW/YGbDWjRiERFpNar3Hb6UErW717h7EZGj5jFmdlTU7FnAP9z9jeDxu8Ahwanye4Bn\nY63TzC43s0Vmtmjjxo1N3wMREUkb1fsOX6NGfbv7ZmABcDqAmf0XkA/8OKrN1rpT5e7+PJBjZv1i\nrOsBdy929+L8/Pxm7IKIiKSL6n2HL5VR3/lmtl9wvytwKvCBmV0GTATOd/faqPYHmJkF98cE26hI\nR/AiIpJe9fW+t6ved1hSKSHaH3jUzLKJJN2n3H2+mVUDnwALg7z838EI7/OAK4P5O4HJwYAzERFp\ng1TvO1xJE7W7LwNGxZgec1l3nwnMbH5oIiKSCVTvO1yqTCYiIgmp3ne4lKhFRCQh1fsOlxK1iIgk\npHrf4VKiFhGRhFTvO1xK1CIikpDqfYdLiVpERBLKy+sFwKb1qiIZBiVqERFJKG//PgBUbtSp7zAo\nUYuISEJ19b4rVO87FErUIiJgRQamAAATSklEQVSSUF2974otO0KOpGNKpYSoiIh0YPX1vmv2hB1K\nh6RELSIiSfXdvY0K1fsOhRK1iIgklVe7iwrV+w6FrlGLiEhSqvcdHiVqERFJqm8OVHTqGnYYHZIS\ntYiIJJWXm01l5x6q9x0CJWoREUlK9b7Do0QtIiJJ1dX7rliret+tTYlaRESSqqv3XbFB9b5bmxK1\niIgkpXrf4VGiFhGRpFTvOzxK1CIikpTqfYdHlclERCQp1fsOjxK1iIikRPW+w6FELSIiKVG973Do\nGrWIiKRE9b7DoUQtIiIpUb3vcChRi4hISlTvOxxK1CIikhLV+w5H0kRtZrlm9o6ZLTWzFWZ2czB9\nkJm9bWZlZjbHzDoH07sEj8uC+QPTuwsiItIa8vYL6n1/+lnIkXQsqRxR7wImuPtIoAg43czGArcB\nd7r7YOBL4AdB+x8AXwbT7wzaiYhIG9e3b1Dv+7NNIUfSsSRN1B5RVzMuJ7g5MAF4Opj+KHBucH9S\n8Jhg/ilmpvH8IiJtXF297wrV+25VKV2jNrNsMysBvgBeBlYDm929OmhSDgwI7g8A1gIE87cAeS0Z\ntIiItL68AfkAVFZ8FXIkHUtKidrda9y9CCgAxgBHNnfDZna5mS0ys0UbN+pn00REMl19ve+tO0OO\npGNp1Khvd98MLACOBfYzs7rKZgXAuuD+OuAggGB+b6AixroecPdidy/Oz89vYvgiItJacnv1oNvu\nKip2qN53a0paQtTM8oE97r7ZzLoCpxIZILYAOA94Evg+MDdYZF7weGEw/zV39zTELiIirSxv9zYq\nt1TA9denvtARR8Cll6YvqHYulVrf/YFHzSybyBH4U+4+38zeB540s18BS4AHg/YPAn8yszKgEpic\nhrhFRCQE/TobVTt3wd13p7ZAdXXkdt550KtXeoNrpywTDnaLi4t90aJFYYchIiJJ1NY6WVmN+CLP\nvHkwaRIsXAhjx6YvsDbIzBa7e3GydqpMJiIiKWtUkgYYNizyd8WKlg+mg1CiFhGR9Bk4EHJz4f33\nw46kzVKiFhGR9MnOhiFDdETdDErUIiKSXsOGKVE3gxK1iIik19ChUF4OW7eGHUmbpEQtIiLpVTeg\nTNepm0SJWkRE0ksjv5tFiVpERNJLI7+bRYlaRETSSyO/m0WJWkRE0k8jv5tMiVpERNJPI7+bTIla\nRETSTyO/m0yJWkRE0k8jv5tMiVpERNJPI7+bTIlaRETSTyO/m0yJWkREWodGfjeJErWIiLQOjfxu\nEiVqERFpHRr53SRK1CIi0jo08rtJlKhFRKR1DBwIXbvqiLqRlKhFRKR1ZGfDkUfqiLqRlKhFRKT1\naOR3oylRi4hI6xk2TCO/G0mJWkREWs/QoZG/uk6dMiVqERFpPRr53WhK1CIi0no08rvRlKhFRKT1\naOR3oylRi4hI69LI70ZJmqjN7CAzW2Bm75vZCjObEkyfY2YlwW2NmZUE0wea2c6oefeneydERKQN\nqRv5vWVL2JG0CZ1SaFMN/MTd3zWznsBiM3vZ3f+jroGZ/RaIfsZXu3tRC8cqIiLtQd3I75UrYezY\ncGNpA5IeUbv7Bnd/N7j/FbASGFA338wM+F/An9MVpIiItCMa+d0ojbpGbWYDgVHA21GTTwA+d/fS\nqGmDzGyJmf3dzE5odpQiItJ+1I38VqJOSSqnvgEwsx7AM8BUd48uKXM+ex9NbwAOdvcKMxsNPGtm\nw/ZZBjO7HLgc4OCDD25q/CIi0tbUjfzWV7RSktIRtZnlEEnSs939v6OmdwK+A8ypm+buu9y9Iri/\nGFgNHL7vOt39AXcvdvfi/Pz85u2FiIi0LRr5nbJURn0b8CCw0t3v2Gf2t4AP3L08qn2+mWUH9w8F\nCoGPWi5kERFp8zTyO2WpHFGPAy4CJkR95erMYN5kGg4iOxFYFnxd62ngCnevbLGIRUSk7Yse+S0J\nJb1G7e5vAhZn3sUxpj1D5DS5iIhIbNEjv/UVrYRUmUxERFqfRn6nTIlaRERan0Z+p0yJWkREwqGR\n3ylRohYRkXBo5HdKlKhFRCQcGvmdEiVqEREJh2p+p0SJWkREwqGR3ylRohYRkXBo5HdKlKhFRCQ8\nGvmdlBK1iIiERyO/k1KiFhGR8Gjkd1JK1CIiEh6N/E5KiVpERMKjkd9JKVGLiEh4NPI7KSVqEREJ\nl0Z+J6RELSIi4dLI74SUqEVEJFwa+Z2QErWIiIRLI78TUqIWEZFwDRqkkd8JKFGLiEi4srJgyBAl\n6jiUqEVEJHxDh+orWnEoUYuISPg08jsuJWoREQlf3YAyHVU3oEQtIiLhq/uKlhJ1A0rUIiISPo38\njkuJWkREwqeR33EpUYuISGbQyO+YlKhFRCQzaOR3TEkTtZkdZGYLzOx9M1thZlOC6b80s3VmVhLc\nzoxa5gYzKzOzVWY2MZ07ICIi7YRGfsfUKYU21cBP3P1dM+sJLDazl4N5d7r77dGNzWwoMBkYBhwI\nvGJmh7t7TUsGLiIi7Uz0yO9jjw03lgyS9Ija3Te4+7vB/a+AlcCABItMAp50913u/jFQBoxpiWBF\nRKQd08jvmBp1jdrMBgKjgLeDSVeb2TIze8jM+gTTBgBroxYrJ3FiFxER0cjvOFJO1GbWA3gGmOru\nW4H7gMOAImAD8NvGbNjMLjezRWa2aOPGjY1ZVERE2iuN/G4gpURtZjlEkvRsd/9vAHf/3N1r3L0W\n+ANfn95eBxwUtXhBMG0v7v6Auxe7e3F+fn5z9kFERNoLjfxuIOlgMjMz4EFgpbvfETW9v7tvCB5+\nG1ge3J8HPGFmdxAZTFYIvNOiUYuISLtUOmgY1114O1z/RORUeIomblrFFWsXpjEy4LHHoLAwvduI\nIZVR3+OAi4D3zKwkmPYz4HwzKwIcWAP8bwB3X2FmTwHvExkx/kON+BYRkVRkjTmGHm9tgOpqIPXU\n0aVzJ+jVK32BAWRnp3f9cZi7h7LhaMXFxb5o0aKwwxAREWk1ZrbY3YuTtVNlMhERkQymRC0iIpLB\nlKhFREQymBK1iIhIBlOiFhERyWBK1CIiIhlMiVpERCSDKVGLiIhkMCVqERGRDKZELSIiksGUqEVE\nRDKYErWIiEgGU6IWERHJYBnx61lmthH4pIVX2w/YlEHtW2MbHXEfWmMbHXEfWmMb7SGm9rAPrbGN\n9hBTU/YhmUPcPT9pK3dvlzdgUSa1z8SY2sM+ZGJM7WEfFFPH2QfFlL59aKmbTn2LiIhkMCVqERGR\nDNaeE/UDGda+NbbREfehNbbREfehNbbRHmJqD/vQGttoDzE1ZR9aREYMJhMREZHY2vMRtYiISNsX\n1ii2dN2A04FVQBkwLYX2DwFfAMtTXP9BwALgfWAFMCVJ+1zgHWBp0P7mFLeTDSwB5qfYfg3wHlBC\nCqMTgf2Ap4EPgJXAsQnaHhGst+62FZiaZP3XBPu7HPgzkJtCTFOC9itirT9WXwF9gZeB0uBvnxSW\n+fdgG7VAcQrtfxM8T8uA/wvsl6T9/wnalgAvAQem+j8H/ARwoF+SbfwSWBfVJ2cmWz/wn8F+rAB+\nnWT9c6LWvQYoSeF5KgL+p+5/EBiTpP1IYGHwf/tXoFey11m8/k7QPlFfx1smZn8naB+zv+O1j9fX\nCdafqK/jbiNWfyfYRsz+TtA+UV/HWyZmfxPnPRIYBLxN5L18DtA5Sfurg7Z7vX6SLDObSL5YTuR/\nNCdJ+weDacuIvH/2SOX9ubm3tG+gNW9Ekttq4FCgc/CEDk2yzInA0aSeqPsDRwf3ewIfJtoGYHWd\nCeQE/3hjU9jOj4EnaFyi7pdK26D9o8Blwf3ORCWfFJ7jz4h8/y9emwHAx0DX4PFTwMVJ1ntU8GLp\nBnQCXgEGJ+sr4NcEH8iAacBtKSwzhMiHj9dp+OYdq/1pQKfg/m3R24jTPjrh/Ai4P5X/OSJvcC8S\nqSnQL8k2fgn8NNX/aWB88Jx2CR7vn+prAPgtcFMK23gJOCO4fybwepL2/wJOCu5fCvyfZK+zeP2d\noH2ivo63TMz+TtA+Zn/Hax+vrxOsP1Ffx1smZn8niilWfydYf6K+jrdMzP4mznskkfeNycH0+4Er\nk7QfBQwkxnthgmXODOYZkQOKZNuI7us7SOFgsCVu7e3U9xigzN0/cvfdwJPApEQLuPs/gMpUN+Du\nG9z93eD+V0SORgckaO/uvi14mBPcPNE2zKwAOAv4Y6pxNYaZ9SbyxvlgEONud9+c4uKnAKvdPVmB\nmk5AVzPrRCT5rk/SfgjwtrvvcPdq4O/Ad6IbxOmrSUQ+dBD8PTfZMu6+0t1XxQoiTvuXgpggchRR\nkKT91qiH3dmnvxP8z90JXNeI9jHFaX8lMMPddwVtvkhl/WZmwP8i8iaWbBsO9Aru9yaqz+O0Pxz4\nR3D/ZeD/i2of73UWs7/jtU/S1/GWidnfCdrH7O8k7xUN+rqx7y1JlonZ38m2sW9/J2ifqK/jLROz\nvxO8R04gctQKe/d1zPbuvsTd18R5nuIt83wwz4kcQRckab816nnqSpL38pbS3hL1AGBt1ONykvyj\nN4eZDSTyKe7tJO2yzayEyKm/l909YXvgd0RexLWNCMeBl8xssZldnqTtIGAj8LCZLTGzP5pZ9xS3\nM5l93rQbBOK+Drgd+BTYAGxx95eSrHc5cIKZ5ZlZNyKfdA9KIZ5vuPuG4P5nwDdSWKY5LgVeSNbI\nzG4xs7XABcBNKbSfBKxz96WNiOVqM1tmZg+ZWZ8kbQ8n8vy+bWZ/N7NjUtzGCcDn7l6aQtupwG+C\n/b4duCFJ+xV8/UH634nT3/u8zpL2d6qvyxSXidnf+7ZP1t/R7VPp6xjxJO3rfZZJ2t9x9jluf+/T\nPqW+3meZuP2973skkTOjm6M+MO31Xt6E99SEy5hZDnAR8Ldk7c3sYSL/e0cC9yTbbktob4m61ZhZ\nD+AZItdStyZq6+417l5E5NPaGDM7KsF6zwa+cPfFjQzpeHc/GjgD+KGZnZigbScipyHvc/dRwHYi\npxETMrPOwDnAX5K060PkBTkIOBDobmYXJlrG3VcSOc34EpEXSwlQkyymfdbhpPETrpn9HKgmcl0r\nWSw/d/eDgrZXJ1lvN+BnpJDQo9wHHEbkWuEGIqcrE+lE5PruWOBa4KngqCCZ80nywSzKlcA1wX5f\nQ3DGJoFLgavMbDGRU6S7922Q6HUWq78b87pMtky8/o7VPlF/R7cP1pewr2OsP2lfx1gmYX8neJ5i\n9neM9kn7OsYycft73/dIIkkwrsa8p6a4zCzgH+7+RrL27n4Jkfe1lcB/JNtui/BWOL/eWjfgWODF\nqMc3ADeksNxAUrxG7V9fs3gR+HETYryJONebgvm3Evn0uIbIp7YdwOON3MYvk2zjAGBN1OMTgOdS\nWO8k4KUU2v078GDU4+8Bsxq5D/8/cFWyviIyEKR/cL8/sCrV/iXGdct47YGLiQyE6daY/x/g4Djb\nrl8GGE7kU/ua4FZN5GzEASluI1a8+z5PfwPGRz1eDeQnWUcn4HOgIJXtAlv4+iufBmxtxPN0OPDO\nPtMavM4S9Xes9in0dcxl4vV3om3E6u992yfr6xTWH6ufYj1Pcfs7wT7H7O8460/W18n2o0F/R827\niciHi018PVZgr/f2GO1/GvV4DUnG60QvA/wX8CyQlUr7qGknkuIYoube2tsR9b+AQjMbFBz9TQbm\nteQGgk+lDwIr3f2OFNrnm9l+wf2uwKlERmLG5O43uHuBuw8kEv9r7p7waNTMuptZz7r7RAbDLE+w\njc+AtWZ2RDDpFCIjNJNJ9ejqU2CsmXULnq9TiHz6TMjM9g/+Hkzk+vQTKWxrHvD94P73gbkpLNMo\nZnY6kUsR57j7jhTaF0Y9nESC/gZw9/fcfX93Hxj0ezmRwTifJdhG/6iH3yZBfweeJTLACDM7nMgA\nwmQ/MPAt4AN3L0/Srs564KTg/gQiI7PjiurvLOAXRAYM1c2L9zqL2d+NfV0mWiZefydoH7O/Y7VP\n1NcJ1h+3rxPsd8z+TvI8NejvBO3j9nWC/YjZ33HeI1cSGTl+XrB4dF836j010TJmdhkwETjf3WuT\ntF9lZoOj9vGcZNttMa3xaaA1b0SubX5I5BPkz1No/2cip5P2EHnR/CBJ++OJnG6r+zrGXl+XiNF+\nBJGvWS0j8gK7qRH7cjIpfGIjMsp9KV9/lSCV/S4i8rWKZURe1H2StO8OVAC9U4z9ZiL/xMuBPxGM\nPk2yzBtEPjAsBU5Jpa+APOBVIm8UrwB9U1jm28H9XUSOIF5M0r6MyNiHuv6+P0n7Z4L9XkbkaygD\nGvM/xz5HBHG28SciX3NZRiR59U/SvjPweBDXu8CEZPEAjwBXpPq6IfLaWBz039vA6CTtpxB5rX4I\nzCA4Qkv0OovX3wnaJ+rreMvE7O8E7WP2d7z28fo6wfoT9XW8ZWL2d6KYYvV3gvUn6ut4y8Tsb+K8\nRxJ5X3sn6I+/8PUI9njtfxT0dTWRDxJ/TPY+HLRdHRXnTfHaE7lU/FbQF8uJXOboFf18peumymQi\nIiIZrL2d+hYREWlXlKhFREQymBK1iIhIBlOiFhERyWBK1CIiIhlMiVpERCSDKVGLiIhkMCVqERGR\nDPb/ACm35mhaHbVJAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4332</th>\n",
       "      <td>4332</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>449.0</td>\n",
       "      <td>398.0</td>\n",
       "      <td>398.0</td>\n",
       "      <td>398.0</td>\n",
       "      <td>398.0</td>\n",
       "      <td>298.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>248.0</td>\n",
       "      <td>248.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      item_id      0      1      2      3  ...  29     30     31  32  33\n",
       "4332     4332  449.0  449.0  449.0  449.0  ... NaN  248.0  248.0 NaN NaN\n",
       "\n",
       "[1 rows x 35 columns]"
      ]
     },
     "execution_count": 95,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "item_temp = nitems_prices.sample(1)#[items_prices['item_id'] == 17523]\n",
    "item_temp_prices = item_temp.loc[:, '0':].T\n",
    "filled_item_temp_prices = item_temp_prices.fillna(method = 'ffill')\n",
    "filled_item_temp_prices = filled_item_temp_prices.fillna(method = 'bfill')\n",
    "\n",
    "decomposition = seasonal_decompose(filled_item_temp_prices, freq = 32, two_sided = False)\n",
    "trend = decomposition.trend\n",
    "\n",
    "plt.plot(filled_item_temp_prices, label='filled item =' + str(item_temp['item_id'].values[0]), color = 'red')\n",
    "plt.plot(item_temp_prices, label='item =' + str(item_temp['item_id'].values[0]))\n",
    "plt.plot(trend, label='trend')\n",
    "plt.legend(loc='best')\n",
    "plt.show()\n",
    "item_temp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "VC3xjVQRhkWY"
   },
   "outputs": [],
   "source": [
    "filled_items_prices = nitems_prices.copy()\n",
    "filled_items_prices.loc[:, '0':] =  filled_items_prices.loc[:, '0':].fillna(method = 'ffill', axis = 1)\n",
    "filled_items_prices.loc[:, '0':] =  filled_items_prices.loc[:, '0':].fillna(method = 'bfill', axis = 1)\n",
    "filled_items_prices_temp = filled_items_prices.loc[:, '0':'33'].T\n",
    "decomposition = seasonal_decompose(filled_items_prices_temp, freq = 32, two_sided = False)\n",
    "trend = decomposition.trend\n",
    "trend = trend.T\n",
    "trend = trend.dropna(axis = 1)\n",
    "trend.columns = ['trend32', 'trend33']\n",
    "trend.trend32 = round(trend.trend32, 4)\n",
    "trend.trend33 = round(trend.trend33, 4)\n",
    "filled_items_prices = pd.concat([filled_items_prices, trend], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jnApJn9q4Avv"
   },
   "outputs": [],
   "source": [
    "filled_items_prices['item_price_status'] ='constant'\n",
    "filled_items_prices.loc[filled_items_prices['trend32'] < filled_items_prices['trend33'], 'item_price_status'] = 'increasing'\n",
    "filled_items_prices.loc[filled_items_prices['trend32'] > filled_items_prices['trend33'], 'item_price_status'] = 'decreasing'\n",
    "filled_items_prices = filled_items_prices.drop(columns = ['trend32', 'trend33'], axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dCDj0iBC9A5V"
   },
   "outputs": [],
   "source": [
    "filled_items_prices['item_price_avg'] = filled_items_prices.loc[:, '0':].median(axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "JFZ-9XY8F1Ab"
   },
   "outputs": [],
   "source": [
    "items = items.merge(filled_items_prices[['item_id', 'item_price_status', 'item_price_avg']], on = 'item_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "iu9ouud6Iqab"
   },
   "source": [
    "## In how many shops each item is getting sold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-GKneK_VIpqC"
   },
   "outputs": [],
   "source": [
    "gb = sales.groupby('item_id')[['shop_id']].nunique()\n",
    "gb.columns = ['item_sold_in_shops']\n",
    "gb = gb.reset_index()\n",
    "grid = grid.merge(gb, on = ['item_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bdeNQvY9Jpzm"
   },
   "source": [
    "## How many items does each shop contains(has sold)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "XnKIWHvEJ2YI"
   },
   "outputs": [],
   "source": [
    "gb = sales.groupby('shop_id')[['item_id']].nunique()\n",
    "gb.columns = ['shop_sold_items']\n",
    "gb = gb.reset_index()\n",
    "grid = grid.merge(gb, on = ['shop_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "M5jeHQksMb9I"
   },
   "source": [
    "## How many items per category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ouBHJFjOoBKd"
   },
   "outputs": [],
   "source": [
    "gb = sales.groupby('item_category_id')[['item_id']].nunique()\n",
    "gb.columns = ['items_per_category']\n",
    "gb = gb.reset_index()\n",
    "grid = grid.merge(gb, on = ['item_category_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Ypyqb1wpoW-y"
   },
   "source": [
    "## How many categories each shop has"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "g6t71nMCoWUi"
   },
   "outputs": [],
   "source": [
    "gb = sales.groupby('shop_id')[['item_category_id']].nunique()\n",
    "gb.columns = ['shop_sold_categories']\n",
    "gb = gb.reset_index()\n",
    "grid = grid.merge(gb, on = ['shop_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ULMi2BsQpHF-"
   },
   "source": [
    "## Statistics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ba6-1OGzpJzw"
   },
   "outputs": [],
   "source": [
    "def create_df_stats(df, stat_name, attr):\n",
    "    gb = df.copy()\n",
    "    max_month =  df.columns.values[-1]\n",
    "    \n",
    "    gb[attr + '_sum'] = gb.loc[:, '0': str(max_month)].sum(axis = 1)\n",
    "    gb[attr + '_max'] = gb.loc[:, '0': str(max_month)].max(axis = 1)\n",
    "    gb[attr + '_std'] = gb.loc[:, '0': str(max_month)].std(axis = 1)\n",
    "    gb[attr + '_median'] = gb.loc[:, '0': str(max_month)].median(axis = 1)\n",
    "    gb = gb.drop(columns = [str(itr) for itr in np.arange(34)], axis = 0)\n",
    "    return gb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_15tlmCKsfsn"
   },
   "source": [
    "### Pairs stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5Gd6A60pslcs"
   },
   "outputs": [],
   "source": [
    "gb = create_df_stats(shops_items_sales, 'pairs', 'pair')\n",
    "gb = gb.drop(columns = ['pair_median'], axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "xn2KbrN2uUb8"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(gb, on = ['shop_id', 'item_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5-fsm-63wb1c"
   },
   "outputs": [],
   "source": [
    "grid = grid.fillna(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "QgOgxaMRq5iF"
   },
   "source": [
    "### item stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_aJ6xHoEpfP7"
   },
   "outputs": [],
   "source": [
    "gb = create_df_stats(items_sales, 'items', 'item')\n",
    "gb = gb.drop(columns = ['item_median'], axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yU42aUBlqpM8"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(gb, on = ['item_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "APEoD7YXrUym"
   },
   "source": [
    "### Shop stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oLeIfll7rSwK"
   },
   "outputs": [],
   "source": [
    "gb = create_df_stats(shops_sales, 'shops', 'shop')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7KbRnNTB-foS"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(gb, on = ['shop_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "b1coDUdiATfE"
   },
   "source": [
    "### Category stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YGIRJh4KAbPI"
   },
   "outputs": [],
   "source": [
    "gb = create_df_stats(categories_sales, 'categories', 'category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KfRQvSrQAuqi"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(gb, on = ['item_category_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZeRATDNmCEb2"
   },
   "source": [
    "## Transactions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "gg2307yqNYMc"
   },
   "source": [
    "### shop/item transactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-wrl6qRiNdwC"
   },
   "outputs": [],
   "source": [
    "shops_items_trans['shop_item_trans'] = shops_items_trans.loc[:, '0':'33'].sum(axis = 1)\n",
    "grid = grid.merge(shops_items_trans[['shop_id', 'item_id', 'shop_item_trans']], on = ['shop_id', 'item_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Mhcf-7LAYtE_"
   },
   "source": [
    "### item transactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "TjhsOWYHYsGu"
   },
   "outputs": [],
   "source": [
    "items_trans['item_trans'] = items_trans.loc[:, '0':'33'].sum(axis = 1)\n",
    "grid = grid.merge(items_trans[['item_id', 'item_trans']], on = 'item_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4plr1Z-rCIHF"
   },
   "source": [
    "### shop transactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oy7ujvZ0aYwe"
   },
   "outputs": [],
   "source": [
    "shops_trans['shop_trans'] = shops_trans.loc[:, '0':'33'].sum(axis = 1)\n",
    "grid = grid.merge(shops_trans[['shop_id', 'shop_trans']], on = 'shop_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Pym71XatbKly"
   },
   "source": [
    "### category transactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yPiw91JdbJvL"
   },
   "outputs": [],
   "source": [
    "categories_trans['category_trans'] = categories_trans.loc[:, '0':'33'].sum(axis = 1)\n",
    "grid = grid.merge(categories_trans[['item_category_id', 'category_trans']], on = 'item_category_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "tWRuwCEtcGY2"
   },
   "source": [
    "## Date Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3wK8PAgYcJLS"
   },
   "outputs": [],
   "source": [
    "dates_train = sales.loc[:, ['date', 'date_block_num']].drop_duplicates()\n",
    "dates_train = dates_train.reset_index(drop = True)\n",
    "dates_test = dates_train.loc[dates_train.loc[:, 'date_block_num'] == 34 - 12]\n",
    "dates_test = dates_test.reset_index(drop=True)\n",
    "dates_test.loc[:,'date_block_num'] = 34\n",
    "dates_test.loc[:, 'date'] = dates_test.loc[:, 'date'] + pd.DateOffset(years = 1)\n",
    "dates_all = pd.concat([dates_train, dates_test])\n",
    "dates_all.loc[:, 'dow'] = dates_all.loc[:, 'date'].dt.dayofweek\n",
    "dates_all.loc[:, 'year'] = dates_all.loc[:, 'date'].dt.year\n",
    "\n",
    "dates_all.loc[:, 'month'] = dates_all.loc[:, 'date'].dt.month\n",
    "dates_all = pd.get_dummies(dates_all, columns = ['dow'])\n",
    "\n",
    "dow_col = ['dow_' + str(x) for x in range(7)]\n",
    "\n",
    "date_features = dates_all.groupby(['year', 'month', 'date_block_num'])[dow_col].agg('sum').reset_index()\n",
    "\n",
    "date_features.loc[:, 'days_of_month'] = date_features.loc[:, dow_col].sum(axis=1)\n",
    "\n",
    "date_features.loc[:, 'year'] = date_features.loc[:, 'year'] - 2013\n",
    "\n",
    "date_features = date_features.loc[:, ['month', 'year', 'days_of_month', 'date_block_num']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "thDCA12kcS1J"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(date_features, on = 'date_block_num', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "1SYT49o9e9PO"
   },
   "source": [
    "## Item selling duration/First sale/last sale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7O9DCHCWfuCd"
   },
   "outputs": [],
   "source": [
    "def get_sale_info(row, name):\n",
    "    duration = 0\n",
    "    sale_start = 34\n",
    "    first_sale = 0\n",
    "    sale_end = 34\n",
    "    last_sale = 0\n",
    "    for month in np.arange(34):\n",
    "        if row[str(month)] > 0:\n",
    "            duration = duration + 1\n",
    "    for month in np.arange(34):\n",
    "        if row[str(month)] > 0:\n",
    "            sale_start = month\n",
    "            first_sale = row[str(month)]\n",
    "            break\n",
    "    for month in np.arange(34):\n",
    "        if row[str( 33 - month)] > 0:\n",
    "            sale_end = 33 - month\n",
    "            last_sale = row[str(33 - month)]\n",
    "            break\n",
    "    row[name + '_sale_start'] = sale_start\n",
    "    row[name + '_first_sale'] = first_sale\n",
    "    row[name + '_sale_end'] = sale_end\n",
    "    row[name + '_last_sale'] = last_sale\n",
    "    row[name + '_selling_duration'] = duration\n",
    "    return row\n",
    "   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 135273,
     "status": "ok",
     "timestamp": 1568522512172,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "fVnbNbEYfA29",
    "outputId": "99c3fc39-b406-462c-85e1-344359c676a0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "134 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "items_sales_info = items_sales.apply(get_sale_info, axis = 1, name = 'item')\n",
    "items_sales_info = items_sales_info.drop(columns = [str(itr) for itr in np.arange(34)], axis = 0)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SKexqfANpErr"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(items_sales_info, on = 'item_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "P_ckCr0VpblT"
   },
   "source": [
    "## shop selling duration/First sale/last sale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1347,
     "status": "ok",
     "timestamp": 1568522521875,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "ZQ8KPlOApgr6",
    "outputId": "6b24237d-8273-4cfa-caf1-9ca56d8a9251"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "shops_sales_info = shops_sales.apply(get_sale_info, axis = 1, name = 'shop')\n",
    "shops_sales_info = shops_sales_info.drop(columns = [str(itr) for itr in np.arange(34)], axis = 0)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dqgWDlXhpsI_"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(shops_sales_info, on = 'shop_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bh3CS6CIqTEn"
   },
   "source": [
    "## category selling duration/First sale/last sale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1564,
     "status": "ok",
     "timestamp": 1568522538312,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "R1C5dqqvqXWd",
    "outputId": "28e5f426-da87-4e8c-f71d-1457a57ead7a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "categories_sales_info = categories_sales.apply(get_sale_info, axis = 1, name = 'category')\n",
    "categories_sales_info = categories_sales_info.drop(columns = [str(itr) for itr in np.arange(34)], axis = 0)\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UvONUXQjqieu"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(categories_sales_info, on = 'item_category_id', how = 'left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 11578,
     "status": "ok",
     "timestamp": 1568522966311,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "VkPFMdbGOsJF",
    "outputId": "9f7282d0-b5fe-4c85-97f7-1198010b1cc1"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "shop_id                            0\n",
       "item_id                            0\n",
       "date_block_num                     0\n",
       "target                             0\n",
       "item_category_id                   0\n",
       "target_lag_1                       0\n",
       "target_lag_2                       0\n",
       "target_lag_3                       0\n",
       "target_lag_4                       0\n",
       "target_lag_9                       0\n",
       "target_lag_10                      0\n",
       "target_lag_11                      0\n",
       "target_lag_12                      0\n",
       "item_id_lag_1                      0\n",
       "item_id_lag_2                      0\n",
       "item_id_lag_3                      0\n",
       "item_id_lag_4                      0\n",
       "item_id_lag_5                      0\n",
       "item_id_lag_6                      0\n",
       "item_id_lag_11                     0\n",
       "item_id_lag_12                     0\n",
       "shop_id_lag_1                      0\n",
       "shop_id_lag_2                      0\n",
       "shop_id_lag_3                      0\n",
       "shop_id_lag_4                      0\n",
       "shop_id_lag_5                      0\n",
       "shop_id_lag_6                      0\n",
       "shop_id_lag_9                      0\n",
       "shop_id_lag_12                     0\n",
       "item_category_id_lag_1             0\n",
       "                              ...   \n",
       "shop_sum                           0\n",
       "shop_max                           0\n",
       "shop_std                           0\n",
       "shop_median                        0\n",
       "category_sum                       0\n",
       "category_max                       0\n",
       "category_std                       0\n",
       "category_median                    0\n",
       "item_trans                     15246\n",
       "shop_trans                         0\n",
       "category_trans                     0\n",
       "shop_item_trans              2561079\n",
       "month                              0\n",
       "year                               0\n",
       "days_of_month                      0\n",
       "item_sale_start                15246\n",
       "item_first_sale                15246\n",
       "item_sale_end                  15246\n",
       "item_last_sale                 15246\n",
       "item_selling_duration          15246\n",
       "shop_sale_start                    0\n",
       "shop_first_sale                    0\n",
       "shop_sale_end                      0\n",
       "shop_last_sale                     0\n",
       "shop_selling_duration              0\n",
       "category_sale_start                0\n",
       "category_first_sale                0\n",
       "category_sale_end                  0\n",
       "category_last_sale                 0\n",
       "category_selling_duration          0\n",
       "Length: 76, dtype: int64"
      ]
     },
     "execution_count": 131,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qKx22yXbPQkD"
   },
   "outputs": [],
   "source": [
    "grid.shop_item_trans = grid.shop_item_trans.fillna(0)\n",
    "grid.item_trans = grid.item_trans.fillna(0)\n",
    "\n",
    "grid.item_sale_start = grid.item_sale_start.fillna(33)\n",
    "grid.item_first_sale = grid.item_sale_start.fillna(0)\n",
    "\n",
    "grid.item_sale_end = grid.item_sale_end.fillna(33)\n",
    "grid.item_last_sale = grid.item_last_sale.fillna(0)\n",
    "\n",
    "grid.item_selling_duration = grid.item_selling_duration.fillna(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "JWHmGrcbrDL5"
   },
   "source": [
    "## Feature Interaction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SJ_qKMxLrE_r"
   },
   "source": [
    "### Item/category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ebkxJ1KwrLR-"
   },
   "outputs": [],
   "source": [
    "grid['item_category_inter'] = grid['item_id'].astype(str) + '_' + grid['item_category_id'].astype(str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "zZeTQ0SmrVPi"
   },
   "outputs": [],
   "source": [
    "grid.loc[:, 'item_category_inter'] = LabelEncoder().fit_transform(grid.loc[:, 'item_category_inter'].values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WEaGRHnvrtdw"
   },
   "source": [
    "### item/shop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KEML_dCrr2UG"
   },
   "outputs": [],
   "source": [
    "grid['item_shop_inter'] = grid['item_id'].astype(str) + '_' + grid['shop_id'].astype(str)\n",
    "grid.loc[:, 'item_shop_inter'] = LabelEncoder().fit_transform(grid.loc[:, 'item_shop_inter'].values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "jjnZbMw-sG_Q"
   },
   "source": [
    "### shop/category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "E9RcaTwQsKY1"
   },
   "outputs": [],
   "source": [
    "grid['category_shop_inter'] = grid['item_category_id'].astype(str) + '_' + grid['shop_id'].astype(str)\n",
    "grid.loc[:, 'category_shop_inter'] = LabelEncoder().fit_transform(grid.loc[:, 'category_shop_inter'].values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8kATU6m0vMq8"
   },
   "source": [
    "## Add categorical features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZwRyFR5ARsd6"
   },
   "source": [
    "### Item categorical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZR--33OcKYNb"
   },
   "outputs": [],
   "source": [
    "for col in items.columns:\n",
    "    if items[col].dtypes == np.object:\n",
    "        items.loc[:, col] = LabelEncoder().fit_transform(items[col].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HDR0wRKsKFMo"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(items, on = ['item_id', 'item_category_id'], how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "NVjsWlZ2Rz4p"
   },
   "source": [
    "### Categories categorical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "BUJLDmXGLjrK"
   },
   "outputs": [],
   "source": [
    "for col in categories.columns:\n",
    "    if categories[col].dtypes == np.object:\n",
    "        categories.loc[:, col] = LabelEncoder().fit_transform(categories[col].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Pw5BcfLwLuZZ"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(categories, on = 'item_category_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ofN-PgxVR4QD"
   },
   "source": [
    "### shop categorical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Nv_sEYwWL6Oh"
   },
   "outputs": [],
   "source": [
    "for col in shops.columns:\n",
    "    if shops[col].dtypes == np.object:\n",
    "        shops.loc[:, col] = LabelEncoder().fit_transform(shops[col].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "h7NC_ieRNuyu"
   },
   "outputs": [],
   "source": [
    "grid = grid.merge(shops, on = 'shop_id', how = 'left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "dkKvuOYOHYMB"
   },
   "source": [
    "# Save grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1327,
     "status": "ok",
     "timestamp": 1568523640427,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "KnCDh35Vs3RU",
    "outputId": "eb00de0f-aa22-441e-e558-58302296d879"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6567567, 93)\n"
     ]
    }
   ],
   "source": [
    "print(grid.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 15748,
     "status": "ok",
     "timestamp": 1568446697305,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "S9FuHN0Os_Px",
    "outputId": "d7cfe6ef-7ba7-4c4d-e7ee-aa2df1126d38"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Memory usage: 3251.97 MB\n",
      "New Memory usage: 1977.20 MB\n"
     ]
    }
   ],
   "source": [
    "grid = downcast_dtypes(grid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1348,
     "status": "ok",
     "timestamp": 1568523477617,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "nO1SjzoGSDki",
    "outputId": "7907f62d-fe41-4848-ec47-cbe49653e437"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'grid_No_missing_06_57_56__15_Sep_2019'"
      ]
     },
     "execution_count": 147,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "save_name = generate_name('grid_No_missing')\n",
    "save_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 129901,
     "status": "ok",
     "timestamp": 1568523631040,
     "user": {
      "displayName": "BA22",
      "photoUrl": "",
      "userId": "02603670721502132211"
     },
     "user_tz": -120
    },
    "id": "OPUsAyQzs0di",
    "outputId": "70cf5c81-34fa-47f6-fe9a-13daa63caa68"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "129 seconds\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "grid.to_pickle(data_path + save_name + '.pkl')\n",
    "print(round(time.time() - start), 'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3IzdiZ_7uuzW"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [
    "QimFCsvHcMqz",
    "Wtrbmid8khfA",
    "eNND7iV0jRqx",
    "-JvnlD9XmZve",
    "qwLAspPygJuD",
    "yB2wn7wuyU4y",
    "iu9ouud6Iqab",
    "bdeNQvY9Jpzm",
    "M5jeHQksMb9I",
    "Ypyqb1wpoW-y",
    "ULMi2BsQpHF-",
    "ZeRATDNmCEb2",
    "tWRuwCEtcGY2",
    "1SYT49o9e9PO",
    "P_ckCr0VpblT",
    "bh3CS6CIqTEn",
    "JWHmGrcbrDL5",
    "8kATU6m0vMq8",
    "dkKvuOYOHYMB"
   ],
   "machine_shape": "hm",
   "name": "feature_engineering.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}