{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Untitled",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyNH5Y3J3nEfsbpgiRYvSzX8"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VGKjc2Bda-Qn"
      },
      "source": [
        "# Session-based travel recommendations\n",
        "> Experimenting with 8 ML models on a simple Trivago-inspired session-based travel dataset\n",
        "\n",
        "- toc: true\n",
        "- badges: true\n",
        "- comments: true\n",
        "- categories: [Logistic, LightGBM, KNN, Session, Sequence, Trivago, Travel]\n",
        "- image:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FN5eEPeRlt4r"
      },
      "source": [
        "## Problem statement"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ENvgsFF9l7UV"
      },
      "source": [
        "### Context"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RlHe9UDwl9G3"
      },
      "source": [
        "Recommending hotels and other travel-related items is still a difficult task, as travel and tourism is a very complex domain. Planning a trip usually involves searching for a set or package of products that are interconnected (e.g., means of transportation, lodging, attractions), with rather limited availability, and where contextual aspects may have a major impact (e.g., time, location, social context). Users book much fewer hotels than, for example, listen to music tracks, and, given the financial obligation of booking a stay at a hotel, users usually exhibit a strong price sensitivity and a bigger need to be convinced by any given offer. Besides, travelers are often emotionally connected to the products and the experience they provide. Therefore, decision making is not only based on rational and objective criteria. As such, providing the right information to visitors to a travel site, such as a hotel booking service, at the right time is challenging. Information about items such as hotels is often available as item metadata. However, usually, in this domain information about users and their goals and preferences is harder to obtain. Systems need to analyze session-based data of anonymous or first-time users to adapt the search results and anticipate the hotels the users may be interested in.\n",
        "\n",
        "Trivago is a global hotel search platform focused on reshaping the way travelers search for and compare hotels, while enabling advertisers of hotels to grow their businesses by providing access to a broad audience of travelers via their websites and apps. Trivago provide aggregated information about the characteristics of each accommodation to help travelers to make an informed decision and find their ideal place to stay. Once a choice is made, the users get redirected to the selected booking site to complete the booking. Trivago has established 55 localized platforms in over 190 countries and provides access to over two million hotels, including alternative accommodations, with prices and availability from over 400+ booking sites and hotel chains. \n",
        "\n",
        "Our users can narrow down their search by selecting filters and specifying the desired characteristics of their preferred accommodation. They can interact with the different offers presented to them and consume the aggregated information for each listing to make an informed decision and find their ideal place to stay. Once a choice is made, the users get redirected to the selected booking site to complete the booking. It is in the interest of all participants (traveler, advertising booking site, and trivago) to suggest suitable accommodations that fit the needs of the traveler.\n",
        "\n",
        "> *We partnered with researchers from TU Wien, Politecnico di Milano, and Karlsruhe Institute of Technology to launch the RecSys Challenge 2019, the annual data science challenge of the ACM Recommender Systems conference. In this challenge, we invite participants to dig deep into our data and come up with creative ideas to detect the intent of our users and build a click-prediction model that can be used to update the recommendation of accommodations. To this end, we have released a data set of user interactions on our website.*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JKH-ASs5mFdI"
      },
      "source": [
        "### Challenges\n",
        "\n",
        "Trivago face a few challenges when it comes to recommending the best options for their visitors, so it’s important to effectively make use of the explicit and implicit user signals within a session (clicks, search refinement, filter usage) to detect the users’ intent as quickly as possible and to update the recommendations to tailor the result list to these needs.\n",
        "\n",
        "Due to the nature of our domain, we face specific challenges that make it difficult to build predictive models and recommendation systems that are tailored to the needs of our visitors. Here are a few examples of the problems that trivago data scientists have to address:\n",
        "\n",
        "- Users search for accommodations comparatively infrequently with sometimes long time intervals between their trips. Furthermore, user intent and preferences change over time and depend on the purpose of the trip (e.g. a business traveler who books accommodation for a weekend trip with her family).\n",
        "- Booking accommodation is an expensive transaction. Visitors are price sensitive and careful when they make a decision. As the availability of the accommodations, the search criteria, and the actual pricing of the deals from the advertisers vary over time, the context­ of each search has to be taken into consideration.\n",
        "- Information about the personal preferences of travelers is sparse. The service provided to our users is free of charge ─ users do not have to provide personal data or make an account in order to use the website.\n",
        "\n",
        "### Goal\n",
        "\n",
        "The task of the challenge is to use all the information about the behavioral, time-dependent patterns of the users and the content of the displayed accommodations to develop models that allow to predict which accommodations a user is most likely to click on when presented with a list of potential options."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rPJD7CrAmJC7"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fp_zJZsWl_P5"
      },
      "source": [
        "### Illustration"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RDta8BgvlwWs"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SohhKa_Jl0xO"
      },
      "source": [
        "Schematic illustration of the dataset and how it is split between train and test set. Each icon in the schematic represents a different type of website interaction, such as clicking on item content (image, info, rating, deals), refining the search parameters via filtering, or triggering new searches for accommodations (items) or destinations. All interactions can be performed by the users and are indicated by the icons in the schematic. Gaps between consecutive interactions indicate the start of a new user session. The train set contains sessions before November 7, 2018, while the test set contains sessions after said date. The item_id of the final click out (shown as the box with the question marks) has been withheld. Note that the question mark refers only to the accommodation identifier that needs to be predicted and not the action type and that every event for which a prediction needs to be made is a click-out. For the evaluation of the leaderboard, the test set has been split into confirmation and a validation set on a user basis."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0ZVdi9l5mRMV"
      },
      "source": [
        "## Data description"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lX3-ZRc5mTdt"
      },
      "source": [
        "Trivago provided two sets of data, the user-item interaction and the item metadata. The interaction dataset consists of the sequential website interactions of users visiting the trivago website between November 1, 2018, and November 8, 2018. Each website interaction corresponds to a specific timestamp in the dataset. Multiple website interactions can appear in a user session. A session is defined as all interactions of a user on a specific trivago country platform with no gaps between the interaction timestamps of >60 minutes. If a user stops interacting with the website and returns after a couple of minutes to continue the search, then the continued interactions will still be counted to belong to the same session. Because of the grouping of website interactions into sessions, the interactions are in the following often referred to as session actions. For each session interaction, data about the context of the interaction are provided, e.g., the country platform on which the interaction took place or the list of items that were shown at the moment of a click and the prices of the accommodations."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vxHgnnLBmgO8"
      },
      "source": [
        "### Description of Action Types and Reference Values for All Possible Session Interactions"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RZOJX7pkmjie"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Sqp40_X7mmNf"
      },
      "source": [
        "Each website interaction corresponds to a specific timestamp in the dataset. Multiple website interactions can appear in a user session. A session is defined as all interactions of a user on a specific trivago country platform with no gaps between the interaction timestamps of >60 minutes. If a user stops interacting with the website and returns after a couple of minutes to continue the search, then the continued interactions will still be counted to belong to the same session. Because of the grouping of website interactions into sessions, the interactions are in the following often referred to as session actions. For each session interaction, data about the context of the interaction are provided, e.g., the country platform on which the interaction took place or the list of items that were shown at the moment of a click and the prices of the accommodations. Metadata for each of the accommodations are provided in a separate file."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lP1ApcRgmrzf"
      },
      "source": [
        "### General Statistics of the RecSys Challenge 2019 Dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CcIq5gzamt38"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qRvA-jW9m1uB"
      },
      "source": [
        "### Session Actions Files"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KRNvto-1m4AQ"
      },
      "source": [
        "Each row in the session action files (train.csv and test.csv) corresponds to a particular user action in a given session. The schema of these files is shown in the below table. The split between the train and test sets was done at a particular split date. That is, sessions that occurred before November 7, 2018, were put into the train set, while those that occurred after were put into the test set. The target of the test set is items clicked out at the end of the sessions in the test set."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xVxRCkxqm51p"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bMAQ3a_3m8TP"
      },
      "source": [
        "In addition to the user actions, the train.csv and test.csv files also contain information about the accommodations that were displayed to the user at the time a clickout was made. An accommodation that is displayed is referred to as being “impressed” and all displayed accommodations are stored in the “impressions” column. Each row in that column is a list of accommodations (items) in the order in which they were displayed on the website. In case the user action was not a clickout, the impressions column is left empty."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l2Tn8P3LWLII"
      },
      "source": [
        "## Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "69ynq4AbT8sc",
        "outputId": "a86ad68c-12e8-447b-b5bf-042a9eee874e"
      },
      "source": [
        "!pip install -q git+https://github.com/sparsh-ai/recochef.git"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
            "    Preparing wheel metadata ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for recochef (PEP 517) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SNLFGrR2jDPK"
      },
      "source": [
        "%load_ext autoreload\n",
        "%autoreload 2"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yfD8l7n3VqUS"
      },
      "source": [
        "import sys\n",
        "import time\n",
        "import math\n",
        "import random\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from scipy import sparse\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "import lightgbm as lgb\n",
        "\n",
        "from recochef.datasets.trivago import Trivago\n",
        "from recochef.datasets.synthetic import Session"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u0CbvZnWWMo4"
      },
      "source": [
        "## Data loading"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4NQRh7CSivjy"
      },
      "source": [
        "### Full load"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tdC6SWFyWHg3"
      },
      "source": [
        "trivago = Trivago()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        },
        "id": "gMSx11nvV0ua",
        "outputId": "e369a779-3a89-482e-a3c9-12e4ed6af22d"
      },
      "source": [
        "df_train = trivago.load_train()\n",
        "df_train.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>USERID</th>\n",
              "      <th>SESSIONID</th>\n",
              "      <th>TIMESTAMP</th>\n",
              "      <th>STEP</th>\n",
              "      <th>EVENTTYPE</th>\n",
              "      <th>REFERENCE</th>\n",
              "      <th>PLATFORM</th>\n",
              "      <th>CITY</th>\n",
              "      <th>DEVICE</th>\n",
              "      <th>FILTERS</th>\n",
              "      <th>IMPRESSIONS</th>\n",
              "      <th>PRICES</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>00RL8Z82B2Z1</td>\n",
              "      <td>aff3928535f48</td>\n",
              "      <td>1541037460</td>\n",
              "      <td>1</td>\n",
              "      <td>search for poi</td>\n",
              "      <td>Newtown</td>\n",
              "      <td>AU</td>\n",
              "      <td>Sydney, Australia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>00RL8Z82B2Z1</td>\n",
              "      <td>aff3928535f48</td>\n",
              "      <td>1541037522</td>\n",
              "      <td>2</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>666856</td>\n",
              "      <td>AU</td>\n",
              "      <td>Sydney, Australia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>00RL8Z82B2Z1</td>\n",
              "      <td>aff3928535f48</td>\n",
              "      <td>1541037522</td>\n",
              "      <td>3</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>666856</td>\n",
              "      <td>AU</td>\n",
              "      <td>Sydney, Australia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>00RL8Z82B2Z1</td>\n",
              "      <td>aff3928535f48</td>\n",
              "      <td>1541037532</td>\n",
              "      <td>4</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>666856</td>\n",
              "      <td>AU</td>\n",
              "      <td>Sydney, Australia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>00RL8Z82B2Z1</td>\n",
              "      <td>aff3928535f48</td>\n",
              "      <td>1541037532</td>\n",
              "      <td>5</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>109038</td>\n",
              "      <td>AU</td>\n",
              "      <td>Sydney, Australia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         USERID      SESSIONID   TIMESTAMP  ...  FILTERS IMPRESSIONS PRICES\n",
              "0  00RL8Z82B2Z1  aff3928535f48  1541037460  ...     None        None   None\n",
              "1  00RL8Z82B2Z1  aff3928535f48  1541037522  ...     None        None   None\n",
              "2  00RL8Z82B2Z1  aff3928535f48  1541037522  ...     None        None   None\n",
              "3  00RL8Z82B2Z1  aff3928535f48  1541037532  ...     None        None   None\n",
              "4  00RL8Z82B2Z1  aff3928535f48  1541037532  ...     None        None   None\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "R6swP2ZlWOgO",
        "outputId": "c09e3785-23ad-4add-e8b6-efb6bde76c5b"
      },
      "source": [
        "df_test = trivago.load_test()\n",
        "df_test.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>USERID</th>\n",
              "      <th>SESSIONID</th>\n",
              "      <th>TIMESTAMP</th>\n",
              "      <th>STEP</th>\n",
              "      <th>EVENTTYPE</th>\n",
              "      <th>REFERENCE</th>\n",
              "      <th>PLATFORM</th>\n",
              "      <th>CITY</th>\n",
              "      <th>DEVICE</th>\n",
              "      <th>FILTERS</th>\n",
              "      <th>IMPRESSIONS</th>\n",
              "      <th>PRICES</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>004A07DM0IDW</td>\n",
              "      <td>1d688ec168932</td>\n",
              "      <td>1541555614</td>\n",
              "      <td>1</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>2059240</td>\n",
              "      <td>CO</td>\n",
              "      <td>Santa Marta, Colombia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>004A07DM0IDW</td>\n",
              "      <td>1d688ec168932</td>\n",
              "      <td>1541555614</td>\n",
              "      <td>2</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>2059240</td>\n",
              "      <td>CO</td>\n",
              "      <td>Santa Marta, Colombia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "      <td>None</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>004A07DM0IDW</td>\n",
              "      <td>1d688ec168932</td>\n",
              "      <td>1541555696</td>\n",
              "      <td>3</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>1050068</td>\n",
              "      <td>CO</td>\n",
              "      <td>Santa Marta, Colombia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>2059240|2033381|1724779|127131|399441|103357|1...</td>\n",
              "      <td>70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>004A07DM0IDW</td>\n",
              "      <td>1d688ec168932</td>\n",
              "      <td>1541555707</td>\n",
              "      <td>4</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>1050068</td>\n",
              "      <td>CO</td>\n",
              "      <td>Santa Marta, Colombia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>2059240|2033381|1724779|127131|399441|103357|1...</td>\n",
              "      <td>70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>004A07DM0IDW</td>\n",
              "      <td>1d688ec168932</td>\n",
              "      <td>1541555717</td>\n",
              "      <td>5</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>1050068</td>\n",
              "      <td>CO</td>\n",
              "      <td>Santa Marta, Colombia</td>\n",
              "      <td>mobile</td>\n",
              "      <td>None</td>\n",
              "      <td>2059240|2033381|1724779|127131|399441|103357|1...</td>\n",
              "      <td>70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         USERID  ...                                             PRICES\n",
              "0  004A07DM0IDW  ...                                               None\n",
              "1  004A07DM0IDW  ...                                               None\n",
              "2  004A07DM0IDW  ...  70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...\n",
              "3  004A07DM0IDW  ...  70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...\n",
              "4  004A07DM0IDW  ...  70|46|48|76|65|65|106|66|87|43|52|44|60|61|50|...\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RmdHIfMuixzw"
      },
      "source": [
        "### Sample load"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rS9gGutmizNT"
      },
      "source": [
        "sample_session_data = Session(version='trivago')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "id": "EbnKHqJzizHq",
        "outputId": "ce2c24c6-f754-472a-8420-0e3dda57afc1"
      },
      "source": [
        "df_train = sample_session_data.train()\n",
        "df_train"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>action_type</th>\n",
              "      <th>reference</th>\n",
              "      <th>impressions</th>\n",
              "      <th>prices</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f89</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>5001</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f89</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>5002</td>\n",
              "      <td>5014|5002|5010</td>\n",
              "      <td>100|125|120</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f89</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>interaction item info</td>\n",
              "      <td>5003</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f89</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>filter selection</td>\n",
              "      <td>unknown</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>64BLF</td>\n",
              "      <td>4504h9</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>5010</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>64BLF</td>\n",
              "      <td>4504h9</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>5001</td>\n",
              "      <td>5001|5023|5040|5005</td>\n",
              "      <td>75|110|65|210</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>5504hFL</td>\n",
              "      <td>7</td>\n",
              "      <td>1</td>\n",
              "      <td>filter selection</td>\n",
              "      <td>unknown</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>5504hFL</td>\n",
              "      <td>8</td>\n",
              "      <td>2</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>5004</td>\n",
              "      <td>5010|5001|5023|5004|5002|5008</td>\n",
              "      <td>120|89|140|126|86|110</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>5504hFL</td>\n",
              "      <td>9</td>\n",
              "      <td>3</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>5001</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>5504hFL</td>\n",
              "      <td>10</td>\n",
              "      <td>4</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>5001</td>\n",
              "      <td>5010|5001|5023|5004|5002|5008</td>\n",
              "      <td>120|89|140|126|86|110</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  user_id session_id  ...                    impressions                 prices\n",
              "0  64BL89    3579f89  ...                            NaN                    NaN\n",
              "1  64BL89    3579f89  ...                 5014|5002|5010            100|125|120\n",
              "2  64BL89    3579f89  ...                            NaN                    NaN\n",
              "3  64BL89    3579f89  ...                            NaN                    NaN\n",
              "4   64BLF     4504h9  ...                            NaN                    NaN\n",
              "5   64BLF     4504h9  ...            5001|5023|5040|5005          75|110|65|210\n",
              "6  64BL89    5504hFL  ...                            NaN                    NaN\n",
              "7  64BL89    5504hFL  ...  5010|5001|5023|5004|5002|5008  120|89|140|126|86|110\n",
              "8  64BL89    5504hFL  ...                            NaN                    NaN\n",
              "9  64BL89    5504hFL  ...  5010|5001|5023|5004|5002|5008  120|89|140|126|86|110\n",
              "\n",
              "[10 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JWSfYEIbnapv"
      },
      "source": [
        "1. User ```64BL89``` viewed the image of item 5001.\n",
        "2. User ```64BL89``` clicked on item 5002. 3 items (5014, 5002, 5010) were shown to the user and user clicked on 2nd item which is 5002. It was the costliest among the 3 items.\n",
        "3. User ```64BL89``` viewed more info of item 5003.\n",
        "4. User ```64BL89``` selected an unknown filter.\n",
        "5. User ```64BLF``` viewed image of item 5010.\n",
        "6. User ```64BLF``` clicked item 5001. It was on top of the recommended list.\n",
        "7. User ```64BL89``` again came back after some time. A new session started for this user. User directly started with applying unknown filter.\n",
        "8. User ```64BL89``` clicked on item 5004. It was at 4th position.\n",
        "9. User ```64BL89``` again viewed the image of item 5001.\n",
        "10. User ```64BL89``` now clicked item 5001 this time. It was at 2nd position.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "lv7z9wLkizEL",
        "outputId": "b1275cf8-0dc5-4a55-dd10-df1d6728503e"
      },
      "source": [
        "df_test = sample_session_data.test()\n",
        "df_test"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>action_type</th>\n",
              "      <th>reference</th>\n",
              "      <th>impressions</th>\n",
              "      <th>prices</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>interaction item image</td>\n",
              "      <td>5023</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5002|5003|5010|5004|5001|5023</td>\n",
              "      <td>120|75|110|105|89|99</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>9</td>\n",
              "      <td>1</td>\n",
              "      <td>interaction item info</td>\n",
              "      <td>5010</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>clickout item</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5001|5004|5010|5014</td>\n",
              "      <td>76|102|115|124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>11</td>\n",
              "      <td>3</td>\n",
              "      <td>filter selection</td>\n",
              "      <td>unknown</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  ...                    impressions                prices\n",
              "0    64BL89    3579f90  ...                            NaN                   NaN\n",
              "1    64BL89    3579f90  ...  5002|5003|5010|5004|5001|5023  120|75|110|105|89|99\n",
              "2  64BL91F2    3779f92  ...                            NaN                   NaN\n",
              "3  64BL91F2    3779f92  ...            5001|5004|5010|5014        76|102|115|124\n",
              "4  64BL91F2    3779f92  ...                            NaN                   NaN\n",
              "\n",
              "[5 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "id": "e0F3Fc57kxHs",
        "outputId": "3c1dba2a-9862-4bc7-9e9b-5b0efe17f1ca"
      },
      "source": [
        "df_items = sample_session_data.items()\n",
        "df_items"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>item_id</th>\n",
              "      <th>properties</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>5001</td>\n",
              "      <td>Wifi|Croissant|TV</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>5002</td>\n",
              "      <td>Wifi|TV</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>5003</td>\n",
              "      <td>Croissant</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>5004</td>\n",
              "      <td>Shoe dryer</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  item_id         properties\n",
              "0    5001  Wifi|Croissant|TV\n",
              "1    5002            Wifi|TV\n",
              "2    5003          Croissant\n",
              "3    5004         Shoe dryer"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5FlKnZO5qVA6"
      },
      "source": [
        "## Utilities"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g8BEGNFMqOZB"
      },
      "source": [
        "def explode(df, col_expl):\n",
        "    \"\"\"Separate string in column col_expl and explode elements into multiple rows.\"\"\"\n",
        "\n",
        "    s = df[col_expl].str.split('|', expand=True).stack()\n",
        "    i = s.index.get_level_values(0)\n",
        "    df2 = df.loc[i].copy()\n",
        "    df2[col_expl] = s.values\n",
        "\n",
        "    return df2\n",
        "\n",
        "\n",
        "def explode_mult(df_in, col_list):\n",
        "    \"\"\"Explode each column in col_list into multiple rows.\"\"\"\n",
        "\n",
        "    df = df_in.copy()\n",
        "\n",
        "    for col in col_list:\n",
        "        df.loc[:, col] = df.loc[:, col].str.split(\"|\")\n",
        "\n",
        "    df_out = pd.DataFrame(\n",
        "        {col: np.repeat(df[col].to_numpy(),\n",
        "                        df[col_list[0]].str.len())\n",
        "         for col in df.columns.drop(col_list)}\n",
        "    )\n",
        "\n",
        "    for col in col_list:\n",
        "        df_out.loc[:, col] = np.concatenate(df.loc[:, col].to_numpy())\n",
        "\n",
        "    return df_out\n",
        "\n",
        "\n",
        "def group_concat(df, gr_cols, col_concat):\n",
        "    \"\"\"Concatenate multiple rows into one.\"\"\"\n",
        "\n",
        "    df_out = (\n",
        "        df\n",
        "        .groupby(gr_cols)[col_concat]\n",
        "        .apply(lambda x: ' '.join(x))\n",
        "        .to_frame()\n",
        "        .reset_index()\n",
        "    )\n",
        "\n",
        "    return df_out\n",
        "\n",
        "\n",
        "def get_target_rows(df):\n",
        "    \"\"\"Restrict data frame to rows for which a prediction needs to be made.\"\"\"\n",
        "    \n",
        "    df_target = df[\n",
        "        (df.action_type == \"clickout item\") & \n",
        "        (df[\"reference\"].isna())\n",
        "    ]\n",
        "\n",
        "    return df_target\n",
        "\n",
        "\n",
        "def summarize_recs(df, rec_col):\n",
        "    \"\"\"Bring the data frame into submission format.\"\"\"\n",
        "\n",
        "    df_rec = (\n",
        "        df\n",
        "        .sort_values(by=[\"user_id\", \"session_id\", \"timestamp\", \"step\", rec_col],\n",
        "                        ascending=[True, True, True, True, False])\n",
        "        .groupby([\"user_id\", \"session_id\", \"timestamp\", \"step\"])[\"impressed_item\"]\n",
        "        .apply(lambda x: ' '.join(x))\n",
        "        .to_frame()\n",
        "        .reset_index()\n",
        "        .rename(columns={'impressed_item': 'item_recommendations'})\n",
        "    )\n",
        "\n",
        "    return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "38H0VRY0qOQi"
      },
      "source": [
        "def print_time(s):\n",
        "    \"\"\"Print string s and current time.\"\"\"\n",
        "\n",
        "    t = time.localtime()\n",
        "    current_time = time.strftime(\"%H:%M:%S\", t)\n",
        "    print(f\"{current_time} | {s}\")\n",
        "\n",
        "\n",
        "def print_header(s):\n",
        "    \"\"\"Print a nice header for string s.\"\"\"\n",
        "\n",
        "    print()\n",
        "    print(f\"##{'#'*len(s)}##\")\n",
        "    print(f\"# {s} #\")\n",
        "    print(f\"##{'#'*len(s)}##\")\n",
        "    print()\n",
        "\n",
        "\n",
        "def validate_model_name(model_name):\n",
        "    \"\"\"Check if the inserted model name is valid.\"\"\"\n",
        "\n",
        "    model_names = [\n",
        "        'gbm_rank', 'logistic_regression',\n",
        "        'nn_interaction', 'nn_item',\n",
        "        'pop_abs', 'pop_user', \n",
        "        'position', 'random'\n",
        "    ]\n",
        "\n",
        "    try:\n",
        "        if model_name not in model_names: raise NameError\n",
        "    except NameError:\n",
        "        print(\"No such model. Please choose a valid one.\")\n",
        "        sys.exit(1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3nwHa4Y2wTsT"
      },
      "source": [
        "## Feature engineering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lhP4qzy2qOUs"
      },
      "source": [
        "def build_features(df):\n",
        "    \"\"\"Build features for the lightGBM and logistic regression model.\"\"\"\n",
        "\n",
        "    # Select columns that are of interest for this method\n",
        "    print_time(\"start\")\n",
        "    cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "            'action_type', 'reference', 'impressions', 'prices']\n",
        "    df_cols = df.loc[:, cols] \n",
        "\n",
        "    # We are only interested in action types, for wich the reference is an item ID\n",
        "    print_time(\"filter interactions\")\n",
        "    item_interactions = [\n",
        "        'clickout item', 'interaction item deals', 'interaction item image',\n",
        "        'interaction item info', 'interaction item rating', 'search for item'\n",
        "    ]\n",
        "    df_actions = (\n",
        "        df_cols\n",
        "        .loc[df_cols.action_type.isin(item_interactions), :]\n",
        "        .copy()\n",
        "        .rename(columns={'reference': 'referenced_item'})\n",
        "    )\n",
        "\n",
        "    print_time(\"cleaning\")\n",
        "    # Clean of instances that have no reference\n",
        "    idx_rm = (df_actions.action_type != \"clickout item\") & (df_actions.referenced_item.isna())\n",
        "    df_actions = df_actions[~idx_rm]\n",
        "\n",
        "    # Get item ID of previous interaction of a user in a session\n",
        "    print_time(\"previous interactions\")\n",
        "    df_actions.loc[:, \"previous_item\"] = (\n",
        "        df_actions\n",
        "        .sort_values(by=[\"user_id\", \"session_id\", \"timestamp\", \"step\"],\n",
        "                        ascending=[True, True, True, True])\n",
        "        .groupby([\"user_id\"])[\"referenced_item\"]\n",
        "        .shift(1)\n",
        "    )\n",
        "\n",
        "    # Combine the impressions and item column, they both contain item IDs\n",
        "    # and we can expand the impression lists in the next step to get the total\n",
        "    # interaction count for an item\n",
        "    print_time(\"combining columns - impressions\")\n",
        "    df_actions.loc[:, \"interacted_item\"] = np.where(\n",
        "        df_actions.impressions.isna(),\n",
        "        df_actions.referenced_item,\n",
        "        df_actions.impressions\n",
        "    )\n",
        "    df_actions = df_actions.drop(columns=\"impressions\")\n",
        "\n",
        "    # Price array expansion will get easier without NAs\n",
        "    print_time(\"combining columns - prices\")\n",
        "    df_actions.loc[:, \"prices\"] = np.where(\n",
        "        df_actions.prices.isna(),\n",
        "        \"\",\n",
        "        df_actions.prices\n",
        "    )\n",
        "\n",
        "    # Convert pipe separated lists into columns\n",
        "    print_time(\"explode arrays\")\n",
        "    df_items = explode_mult(df_actions, [\"interacted_item\", \"prices\"]).copy()\n",
        "\n",
        "    # Feature: Number of previous interactions with an item\n",
        "    print_time(\"interaction count\")\n",
        "    df_items.loc[:, \"interaction_count\"] = (\n",
        "        df_items\n",
        "        .groupby([\"user_id\", \"interacted_item\"])\n",
        "        .cumcount()\n",
        "    )\n",
        "\n",
        "    # Reduce to impression level again \n",
        "    print_time(\"reduce to impressions\")\n",
        "    df_impressions = (\n",
        "        df_items[df_items.action_type == \"clickout item\"]\n",
        "        .copy()\n",
        "        .drop(columns=\"action_type\")\n",
        "        .rename(columns={\"interacted_item\": \"impressed_item\"})\n",
        "    )\n",
        "\n",
        "    # Feature: Position of item in the original list.\n",
        "    # Items are in original order after the explode for each index\n",
        "    print_time(\"position feature\")\n",
        "    df_impressions.loc[:, \"position\"] = (\n",
        "        df_impressions\n",
        "        .groupby([\"user_id\", \"session_id\", \"timestamp\", \"step\"])\n",
        "        .cumcount()+1\n",
        "    )\n",
        "\n",
        "    # Feature: Is the impressed item the last interacted item\n",
        "    print_time(\"last interacted item feature\")\n",
        "    df_impressions.loc[:, \"is_last_interacted\"] = (\n",
        "        df_impressions[\"previous_item\"] == df_impressions[\"impressed_item\"]\n",
        "    ).astype(int)\n",
        "\n",
        "    print_time(\"change price datatype\")\n",
        "    df_impressions.loc[:, \"prices\"] = df_impressions.prices.astype(int)\n",
        "\n",
        "    return_cols = [\n",
        "        \"user_id\",\n",
        "        \"session_id\",\n",
        "        \"timestamp\",\n",
        "        \"step\",\n",
        "        \"position\",\n",
        "        \"prices\",\n",
        "        \"interaction_count\",\n",
        "        \"is_last_interacted\",\n",
        "        \"referenced_item\",\n",
        "        \"impressed_item\",\n",
        "    ]\n",
        "\n",
        "    df_return = df_impressions[return_cols]\n",
        "\n",
        "    return df_return"
      ],
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HQ5CBMSrgrKo"
      },
      "source": [
        "## Models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c_qWk0GNreBc"
      },
      "source": [
        "### List of models\n",
        "    - gbm_rank: lightGBM model\n",
        "    - log_reg: Logistic regression\n",
        "    - nn_interaction: kNN w/ session co-occurrence\n",
        "    - nn_item: kNN w/ metadata similarity\n",
        "    - pop_abs: Popularity - total clicks\n",
        "    - pop_user: Popularity - distinct users\n",
        "    - position: Original display position\n",
        "    - random: Random order"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DHHpskPBrgo0"
      },
      "source": [
        "### random"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i-aIjKUBrUp_"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelRandom():\n",
        "    \"\"\"\n",
        "    Model class for the random ordering model.\n",
        "    Methods\n",
        "        fit(df): Not needed. Only added for consistency with other model classes\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "    def fit(self, _):\n",
        "        pass\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Randomly sort the impressions list.\"\"\"\n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        print_time(\"target rows\")\n",
        "        df_target = get_target_rows(df.copy())\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        random.seed(10121)\n",
        "        df_target.loc[:, \"item_recs_list\"] = (\n",
        "            df_target\n",
        "            .loc[:, \"impressions\"].str.split(\"|\")\n",
        "            .map(lambda x: sorted(x, key=lambda k: random.random()))\n",
        "        )\n",
        "\n",
        "        df_target.loc[:, \"item_recommendations\"] = (\n",
        "           df_target[\"item_recs_list\"]\n",
        "           .map(lambda arr: ' '.join(arr))\n",
        "        )\n",
        "\n",
        "        cols_rec = [\"user_id\", \"session_id\", \"timestamp\", \"step\", \"item_recommendations\"]\n",
        "        df_rec = df_target.loc[:, cols_rec]\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nu3Hxeb6rxX0",
        "outputId": "191405a7-5ddf-45f5-c116-751cf3b84c6c"
      },
      "source": [
        "model = ModelRandom()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "14:47:52 | target rows\n",
            "14:47:52 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "Uo82KIhgrzkN",
        "outputId": "53e8e774-b55a-4fa5-a641-7ec2327fdb2b"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5004 5010 5001 5002 5003 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5010 5014 5004</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "1    64BL89    3579f90          6     2  5004 5010 5001 5002 5003 5023\n",
              "3  64BL91F2    3779f92         10     2            5001 5010 5014 5004"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J09qiuF6sQcY"
      },
      "source": [
        "### position"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b_lcsrHMsI0M"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelPosition():\n",
        "    \"\"\"\n",
        "    Model class for the model based on the original position in displayed list.\n",
        "    Methods\n",
        "        fit(df): Not needed. Only added for consistency with other model classes\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "    def fit(self, _):\n",
        "        pass\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Return items in impressions list in original order.\"\"\"\n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        print_time(\"target rows\")\n",
        "        df_target = get_target_rows(df.copy())\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        df_target[\"item_recommendations\"] = (\n",
        "            df_target\n",
        "            .apply(lambda x: x.impressions.replace(\"|\", \" \"), axis=1)\n",
        "        )\n",
        "\n",
        "        cols_rec = [\"user_id\", \"session_id\", \"timestamp\", \"step\", \"item_recommendations\"]\n",
        "        df_rec = df_target.loc[:, cols_rec]\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-CvVzvKzslSy",
        "outputId": "04074250-0ff3-4407-b7c7-f69393afc413"
      },
      "source": [
        "model = ModelPosition()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "14:51:15 | target rows\n",
            "14:51:15 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "f8cNjbSwsrbS",
        "outputId": "08fa97d5-db46-44ee-d4e6-b7f6252772ad"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5002 5003 5010 5004 5001 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "1    64BL89    3579f90          6     2  5002 5003 5010 5004 5001 5023\n",
              "3  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ndeQGvQFtWHh"
      },
      "source": [
        "### pop_abs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cxjT_bvBtHAK"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelPopAbs():\n",
        "    \"\"\"\n",
        "    Model class for the popularity model based on total number of clicks.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Count the number of clicks for each item.\"\"\"\n",
        "\n",
        "        # Select columns that are of interest for this method\n",
        "        print_time(\"start\")\n",
        "        cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "                'action_type', 'reference']\n",
        "        df_cols = df.loc[:, cols] \n",
        "\n",
        "        # We only need to count clickouts per item\n",
        "        print_time(\"clicks per item\")\n",
        "        df_item_clicks = (\n",
        "            df_cols\n",
        "            .loc[df_cols[\"action_type\"] == \"clickout item\", :]\n",
        "            .groupby(\"reference\")\n",
        "            .size()\n",
        "            .reset_index(name=\"n_clicks\")\n",
        "            .rename(columns={\"reference\": \"item\"})\n",
        "        )\n",
        "\n",
        "        self.df_pop = df_item_clicks\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Sort the impression list by number of clicks in the training phase.\"\"\"\n",
        "\n",
        "        # Select columns that are of interest for this method\n",
        "        print_time(\"start\")\n",
        "        cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "                'action_type', 'reference', \"impressions\"]\n",
        "        df_cols = df.loc[:, cols] \n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        print_time(\"target rows\")\n",
        "        df_target = get_target_rows(df_cols)\n",
        "\n",
        "        # Explode to impression level\n",
        "        print_time(\"explode impression array\")\n",
        "        df_impressions = (\n",
        "            explode(df_target, \"impressions\")\n",
        "            .rename(columns={\"impressions\": \"impressed_item\"})\n",
        "        )\n",
        "        df_impressions = (\n",
        "            df_impressions\n",
        "            .merge(\n",
        "                self.df_pop,\n",
        "                left_on=\"impressed_item\",\n",
        "                right_on=\"item\",\n",
        "                how=\"left\"\n",
        "            )\n",
        "        )\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        df_rec = summarize_recs(df_impressions, \"n_clicks\")\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z1zRlpp7tda5",
        "outputId": "a7fcdd29-988c-4341-fe58-bcf49394b950"
      },
      "source": [
        "model = ModelPopAbs()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "14:57:30 | start\n",
            "14:57:30 | clicks per item\n",
            "14:57:30 | start\n",
            "14:57:30 | target rows\n",
            "14:57:30 | explode impression array\n",
            "14:57:30 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "WseXnTsSteDP",
        "outputId": "211341f2-24e6-49bd-ff64-bef3a1006091"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5002 5004 5003 5010 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5001 5002 5004 5003 5010 5023\n",
              "1  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MYCNr7aNtnw_"
      },
      "source": [
        "### pop_user"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O8uui6u-tnxA"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelPopUsers():\n",
        "    \"\"\"\n",
        "    Model class for the popularity model based on distinct users.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Count the number of distinct users that click on an item.\"\"\"\n",
        "\n",
        "        # Select columns that are of interest for this method\n",
        "        print_time(\"start\")\n",
        "        cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "                'action_type', 'reference']\n",
        "        df_cols = df.loc[:, cols] \n",
        "\n",
        "        # We only need to count clickouts per item\n",
        "        print_time(\"clicks per item\")\n",
        "        df_item_clicks = (\n",
        "            df_cols\n",
        "            .loc[df_cols[\"action_type\"] == \"clickout item\", :]\n",
        "            .groupby(\"reference\")\n",
        "            .user_id\n",
        "            .nunique()\n",
        "            .reset_index(name=\"n_users\")\n",
        "            .rename(columns={\"reference\": \"item\"})\n",
        "        )\n",
        "\n",
        "        self.df_pop = df_item_clicks\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Sort the impression list by number of distinct users in the training phase.\"\"\"\n",
        "\n",
        "        # Select columns that are of interest for this method\n",
        "        print_time(\"start\")\n",
        "        cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "                'action_type', 'reference', \"impressions\"]\n",
        "        df_cols = df.loc[:, cols] \n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        print_time(\"target rows\")\n",
        "        df_target = get_target_rows(df_cols)\n",
        "\n",
        "        # Explode to impression level\n",
        "        print_time(\"explode impression array\")\n",
        "        df_impressions = (\n",
        "            explode(df_target, \"impressions\")\n",
        "            .rename(columns={\"impressions\": \"impressed_item\"})\n",
        "        )\n",
        "        df_impressions = (\n",
        "            df_impressions\n",
        "            .merge(\n",
        "                self.df_pop,\n",
        "                left_on=\"impressed_item\",\n",
        "                right_on=\"item\",\n",
        "                how=\"left\"\n",
        "            )\n",
        "        )\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        df_rec = summarize_recs(df_impressions, \"n_users\")\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pUPXq_tltnxB",
        "outputId": "e80bddcb-1e55-44cc-dae2-53500f2ffc84"
      },
      "source": [
        "model = ModelPopUsers()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "14:59:09 | start\n",
            "14:59:09 | clicks per item\n",
            "14:59:09 | start\n",
            "14:59:09 | target rows\n",
            "14:59:09 | explode impression array\n",
            "14:59:09 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "zwHJYLQNtnxB",
        "outputId": "0882c7b8-1d3c-4f17-fcfb-c9d5c693a6be"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5002 5004 5003 5010 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5001 5002 5004 5003 5010 5023\n",
              "1  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cfJsTtwutog5"
      },
      "source": [
        "### nn_item"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0RXX6i_gv8v_"
      },
      "source": [
        "#collapse-hide\n",
        "def calc_item_sims(df, item_col, reference_col):\n",
        "    \"\"\"Calculate similarity of items based on nearest neighbor algorithm.\n",
        "    The final data frame will have similarity scores for pairs of items.\n",
        "    :param df: Data frame of training data\n",
        "    :param item_col: Name of data frame column that contains the item ID\n",
        "    :param reference_col: Name of the reference column, depending on the model either\n",
        "        1. session_id for the similarity based on session co-occurrences\n",
        "        2. properties for the similarity based on item metadata\n",
        "    :return: Data frame with item pairs and similarity scores\n",
        "    \"\"\"\n",
        "\n",
        "    # Create data frame with item and reference indices\n",
        "    print_time(\"item and reference indices\")\n",
        "    unique_items = df[item_col].unique()\n",
        "    unique_refs = df[reference_col].unique()\n",
        "\n",
        "    d_items = {item_col: unique_items, 'item_idx': range(0, len(unique_items))}\n",
        "    d_refs = {reference_col: unique_refs, 'ref_idx': range(0, len(unique_refs))}\n",
        "\n",
        "    df_items = pd.DataFrame(data=d_items)\n",
        "    df_refs = pd.DataFrame(data=d_refs)\n",
        "\n",
        "    df = (\n",
        "        df\n",
        "        .merge(\n",
        "            df_items,\n",
        "            how=\"inner\",\n",
        "            on=item_col\n",
        "        )\n",
        "        .merge(\n",
        "            df_refs,\n",
        "            how=\"inner\",\n",
        "            on=reference_col\n",
        "        )\n",
        "    )\n",
        "\n",
        "    df_idx = (\n",
        "        df\n",
        "        .loc[:, [\"item_idx\", \"ref_idx\"]]\n",
        "        .assign(data=lambda x: 1.)\n",
        "        .drop_duplicates()\n",
        "    )\n",
        "\n",
        "    # Build item co-ooccurrence matrix\n",
        "    print_time(\"item co-occurrence matrix\")\n",
        "    mat_coo = sparse.coo_matrix((df_idx.data, (df_idx.item_idx, df_idx.ref_idx)))\n",
        "    mat_item_coo = mat_coo.T.dot(mat_coo)\n",
        "\n",
        "    # Calculate Cosine similarities\n",
        "    print_time(\"Cosine similarity\")\n",
        "    inv_occ = np.sqrt(1 / mat_item_coo.diagonal())\n",
        "    cosine_sim = mat_item_coo.multiply(inv_occ)\n",
        "    cosine_sim = cosine_sim.T.multiply(inv_occ)\n",
        "\n",
        "    # Create item similarity data frame\n",
        "    print_time(\"item similarity data frame\")\n",
        "    idx_ref, idx_item, sim = sparse.find(cosine_sim)\n",
        "    d_item_sim = {'idx_ref': idx_ref, 'idx_item': idx_item, 'similarity': sim}\n",
        "    df_item_sim = pd.DataFrame(data=d_item_sim)\n",
        "\n",
        "    df_item_sim = (\n",
        "        df_item_sim\n",
        "        .merge(\n",
        "            df_items.assign(item_ref=df_items[item_col]),\n",
        "            how=\"inner\",\n",
        "            left_on=\"idx_ref\",\n",
        "            right_on=\"item_idx\"\n",
        "        )\n",
        "        .merge(\n",
        "            df_items.assign(item_sim=df_items[item_col]),\n",
        "            how=\"inner\",\n",
        "            left_on=\"idx_item\",\n",
        "            right_on=\"item_idx\"\n",
        "        )\n",
        "        .loc[:, [\"item_ref\", \"item_sim\", \"similarity\"]]\n",
        "    )\n",
        "\n",
        "    return df_item_sim\n",
        "\n",
        "\n",
        "def predict_nn(df, df_item_sim):\n",
        "    \"\"\"Calculate predictions based on the item similarity scores.\"\"\"\n",
        "\n",
        "    # Select columns that are of interest for this function\n",
        "    print_time(\"start\")\n",
        "    cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "            'action_type', 'reference', 'impressions']\n",
        "    df_cols = df.loc[:, cols] \n",
        "\n",
        "    # Get previous reference per user\n",
        "    print_time(\"previous reference\")\n",
        "    df_cols[\"previous_reference\"] = (\n",
        "        df_cols\n",
        "        .sort_values(by=[\"user_id\", \"session_id\", \"timestamp\"],\n",
        "                     ascending=[True, True, True])\n",
        "        .groupby([\"user_id\"])[\"reference\"]\n",
        "        .shift(1)\n",
        "    )\n",
        "\n",
        "    # Target row, withheld item ID that needs to be predicted\n",
        "    print_time(\"target rows\")\n",
        "    df_target = get_target_rows(df_cols)\n",
        "\n",
        "    # Explode to impression level\n",
        "    print_time(\"explode impression array\")\n",
        "    df_impressions = explode(df_target, \"impressions\")\n",
        "\n",
        "    df_item_sim[\"item_ref\"] = df_item_sim[\"item_ref\"].astype(str)\n",
        "    df_item_sim[\"item_sim\"] = df_item_sim[\"item_sim\"].astype(str)\n",
        "\n",
        "    # Get similarities\n",
        "    print_time(\"get similarities\")\n",
        "    df_impressions = (\n",
        "        df_impressions\n",
        "        .merge(\n",
        "            df_item_sim,\n",
        "            how=\"left\",\n",
        "            left_on=[\"previous_reference\", \"impressions\"],\n",
        "            right_on=[\"item_ref\", \"item_sim\"]\n",
        "        )\n",
        "        .fillna(value={'similarity': 0})\n",
        "        .sort_values(by=[\"user_id\", \"timestamp\", \"step\", \"similarity\"],\n",
        "                        ascending=[True, True, True, False])\n",
        "    )\n",
        "\n",
        "    # Summarize recommendations\n",
        "    print_time(\"summarize recommendations\")\n",
        "    df_rec = group_concat(\n",
        "        df_impressions, [\"user_id\", \"session_id\", \"timestamp\", \"step\"], \n",
        "        \"impressions\"\n",
        "    )\n",
        "\n",
        "    df_rec = (\n",
        "        df_rec\n",
        "        .rename(columns={'impressions': 'item_recommendations'})\n",
        "        .loc[:, [\"user_id\", \"session_id\", \"timestamp\", \"step\", \"item_recommendations\"]]\n",
        "    )\n",
        "\n",
        "    return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lpt9T1uftog6"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelNNItem():\n",
        "    \"\"\"\n",
        "    Model class for the item metadata nearest neighbor model.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Calculate item similarity based on item metadata.\"\"\"\n",
        "\n",
        "        # Explode property arrays\n",
        "        print_time(\"explode properties\")\n",
        "        df_properties = explode(df, \"properties\")\n",
        "\n",
        "        df_item_sim = calc_item_sims(df_properties, \"item_id\", \"properties\")\n",
        "\n",
        "        self.df_item_sim = df_item_sim\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Sort impression list by similarity.\"\"\"\n",
        "\n",
        "        df_rec = predict_nn(df, self.df_item_sim)\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n2guldDTtog7",
        "outputId": "13205c1d-fe07-483a-e30e-d3c5b119fb29"
      },
      "source": [
        "model = ModelNNItem()\n",
        "model.fit(df_items)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15:07:50 | explode properties\n",
            "15:07:50 | item and reference indices\n",
            "15:07:50 | item co-occurrence matrix\n",
            "15:07:50 | Cosine similarity\n",
            "15:07:50 | item similarity data frame\n",
            "15:07:50 | start\n",
            "15:07:50 | previous reference\n",
            "15:07:50 | target rows\n",
            "15:07:50 | explode impression array\n",
            "15:07:50 | get similarities\n",
            "15:07:50 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "fh5UB1X6tog8",
        "outputId": "4896e452-1a9d-4292-b320-ffaeae7e54f6"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5002 5003 5010 5004 5001 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5002 5003 5010 5004 5001 5023\n",
              "1  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GfyyfSBktpIp"
      },
      "source": [
        "### nn_interaction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FqfcAfjwtpIq"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelNNInteraction():\n",
        "    \"\"\"\n",
        "    Model class for the session co-occurrence nearest neighbor model.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Calculate item similarity based on session co-occurrence.\"\"\"\n",
        "\n",
        "        # Select columns that are of interest for this method\n",
        "        print_time(\"start\")\n",
        "        cols = ['user_id', 'session_id', 'timestamp', 'step',\n",
        "                'action_type', 'reference']\n",
        "        df_cols = df.loc[:, cols] \n",
        "\n",
        "        # We are only interested in action types, for wich the reference is an item ID\n",
        "        print_time(\"filter interactions\")\n",
        "        item_interactions = [\n",
        "            'clickout item', 'interaction item deals', 'interaction item image',\n",
        "            'interaction item info', 'interaction item rating', 'search for item'\n",
        "        ]\n",
        "        df_actions = (\n",
        "            df_cols\n",
        "            .loc[df_cols.action_type.isin(item_interactions), :]\n",
        "            .rename(columns={'reference': 'item'})\n",
        "            .drop(columns='action_type')\n",
        "        )\n",
        "\n",
        "        df_item_sim = calc_item_sims(df_actions, \"item\", \"session_id\")\n",
        "\n",
        "        self.df_item_sim = df_item_sim\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Sort impression list by similarity.\"\"\"\n",
        "\n",
        "        df_rec = predict_nn(df, self.df_item_sim)\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uXUiInYetpIr",
        "outputId": "ccd00752-1248-43bd-f7e1-8ae43155a669"
      },
      "source": [
        "model = ModelNNInteraction()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15:09:13 | start\n",
            "15:09:13 | filter interactions\n",
            "15:09:13 | item and reference indices\n",
            "15:09:13 | item co-occurrence matrix\n",
            "15:09:13 | Cosine similarity\n",
            "15:09:13 | item similarity data frame\n",
            "15:09:13 | start\n",
            "15:09:13 | previous reference\n",
            "15:09:13 | target rows\n",
            "15:09:13 | explode impression array\n",
            "15:09:13 | get similarities\n",
            "15:09:13 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "FXJhGxTltpIs",
        "outputId": "feaa355d-4e96-4bac-a8df-5c5b67fbd005"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5002 5003 5010 5004 5001 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5002 5003 5010 5004 5001 5023\n",
              "1  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JDHr8CGAtpSX"
      },
      "source": [
        "### log_reg"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_gieDTNbtpSY"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelLogReg():\n",
        "    \"\"\"\n",
        "    Model class for the logistic regression model.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Train the logistic regression model.\"\"\"\n",
        "\n",
        "        df_impressions = build_features(df)\n",
        "\n",
        "        # Target column, item that was clicked\n",
        "        print_time(\"target column\")\n",
        "        df_impressions.loc[:, \"is_clicked\"] = (\n",
        "            df_impressions[\"referenced_item\"] == df_impressions[\"impressed_item\"]\n",
        "        ).astype(int)\n",
        "\n",
        "        features = [\n",
        "            \"position\",\n",
        "            \"prices\",\n",
        "            \"interaction_count\",\n",
        "            \"is_last_interacted\",\n",
        "        ]\n",
        "\n",
        "        X = df_impressions[features]\n",
        "        y = df_impressions.is_clicked\n",
        "\n",
        "        # Training the actual model\n",
        "        print_time(\"training logistic regression model\")\n",
        "        self.logreg = LogisticRegression(solver=\"lbfgs\", max_iter=100, tol=1e-11, C=1e10).fit(X, y)\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Calculate click probability based on trained logistic regression model.\"\"\"\n",
        "\n",
        "        df_impressions = build_features(df)\n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        df_impressions = df_impressions[df_impressions.referenced_item.isna()]\n",
        "\n",
        "        features = [\n",
        "            \"position\",\n",
        "            \"prices\",\n",
        "            \"interaction_count\",\n",
        "            \"is_last_interacted\"\n",
        "        ]\n",
        "\n",
        "        # Predict clickout probabilities for each impressed item\n",
        "        print_time(\"predict clickout item\")\n",
        "        df_impressions.loc[:, \"click_probability\"] = (\n",
        "            self\n",
        "            .logreg\n",
        "            .predict_proba(df_impressions[features])[:, 1]\n",
        "        )\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        df_rec = summarize_recs(df_impressions, \"click_probability\")\n",
        "\n",
        "        return df_rec"
      ],
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EG6cppXHtpSZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4a79dcd6-6c54-41a7-897a-f3b326f41671"
      },
      "source": [
        "model = ModelLogReg()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15:16:04 | start\n",
            "15:16:04 | filter interactions\n",
            "15:16:04 | cleaning\n",
            "15:16:04 | previous interactions\n",
            "15:16:04 | combining columns - impressions\n",
            "15:16:04 | combining columns - prices\n",
            "15:16:04 | explode arrays\n",
            "15:16:04 | interaction count\n",
            "15:16:04 | reduce to impressions\n",
            "15:16:04 | position feature\n",
            "15:16:04 | last interacted item feature\n",
            "15:16:04 | change price datatype\n",
            "15:16:04 | target column\n",
            "15:16:04 | training logistic regression model\n",
            "15:16:04 | start\n",
            "15:16:04 | filter interactions\n",
            "15:16:04 | cleaning\n",
            "15:16:04 | previous interactions\n",
            "15:16:04 | combining columns - impressions\n",
            "15:16:04 | combining columns - prices\n",
            "15:16:04 | explode arrays\n",
            "15:16:04 | interaction count\n",
            "15:16:04 | reduce to impressions\n",
            "15:16:04 | position feature\n",
            "15:16:04 | last interacted item feature\n",
            "15:16:04 | change price datatype\n",
            "15:16:04 | predict clickout item\n",
            "15:16:04 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WYb8RtQ1tpSZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "b2b25ee0-e759-4d25-e519-fbee972b95d4"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5023 5002 5003 5010 5004 5001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5010 5001 5004 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5023 5002 5003 5010 5004 5001\n",
              "1  64BL91F2    3779f92         10     2            5010 5001 5004 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q9r_boM1tpf6"
      },
      "source": [
        "### gbm_rank"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QatREEBjtpf7"
      },
      "source": [
        "#collapse-hide\n",
        "class ModelGbmRank():\n",
        "    \"\"\"\n",
        "    Model class for the lightGBM model.\n",
        "    Methods\n",
        "        fit(df): Fit the model on training data\n",
        "        predict(df): Calculate recommendations for test data        \n",
        "    \"\"\"\n",
        "\n",
        "    def fit(self, df):\n",
        "        \"\"\"Train the lightGBM model.\"\"\"\n",
        "\n",
        "        df_impressions = build_features(df)\n",
        "\n",
        "        # Target column, item that was clicked\n",
        "        print_time(\"target column\")\n",
        "        df_impressions.loc[:, \"is_clicked\"] = (\n",
        "            df_impressions[\"referenced_item\"] == df_impressions[\"impressed_item\"]\n",
        "        ).astype(int)\n",
        "\n",
        "        features = [\n",
        "            \"position\",\n",
        "            \"prices\",\n",
        "            \"interaction_count\",\n",
        "            \"is_last_interacted\",\n",
        "        ]\n",
        "\n",
        "        # Bring to format suitable for lightGBM\n",
        "        print_time(\"lightGBM format\")\n",
        "        X = df_impressions[features]\n",
        "        y = df_impressions.is_clicked\n",
        "\n",
        "        q = (\n",
        "            df_impressions\n",
        "            .groupby([\"user_id\", \"session_id\", \"timestamp\", \"step\"])\n",
        "            .size()\n",
        "            .reset_index(name=\"query_length\")\n",
        "            .query_length\n",
        "        )\n",
        "\n",
        "        # Training the actual model\n",
        "        print_time(\"training lightGBM model\")\n",
        "        self.gbm = lgb.LGBMRanker()\n",
        "        self.gbm.fit(X, y, group=q, verbose=True)\n",
        "\n",
        "\n",
        "    def predict(self, df):\n",
        "        \"\"\"Calculate item ranking based on trained lightGBM model.\"\"\"\n",
        "\n",
        "        df_impressions = build_features(df)\n",
        "\n",
        "        # Target row, withheld item ID that needs to be predicted\n",
        "        df_impressions = df_impressions[df_impressions.referenced_item.isna()]\n",
        "\n",
        "        features = [\n",
        "            \"position\",\n",
        "            \"prices\",\n",
        "            \"interaction_count\",\n",
        "            \"is_last_interacted\"\n",
        "        ]\n",
        "\n",
        "        df_impressions.loc[:, \"click_propensity\"] = self.gbm.predict(df_impressions[features])\n",
        "\n",
        "        # Summarize recommendations\n",
        "        print_time(\"summarize recommendations\")\n",
        "        df_rec = summarize_recs(df_impressions, \"click_propensity\")\n",
        "         \n",
        "        return df_rec"
      ],
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "meVWs09Ptpf7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b59757d7-90ae-41b8-e1df-43ab5f0ba0b1"
      },
      "source": [
        "model = ModelGbmRank()\n",
        "model.fit(df_train)\n",
        "df_recommendations = model.predict(df_test)"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15:16:53 | start\n",
            "15:16:53 | filter interactions\n",
            "15:16:53 | cleaning\n",
            "15:16:53 | previous interactions\n",
            "15:16:53 | combining columns - impressions\n",
            "15:16:53 | combining columns - prices\n",
            "15:16:53 | explode arrays\n",
            "15:16:53 | interaction count\n",
            "15:16:53 | reduce to impressions\n",
            "15:16:53 | position feature\n",
            "15:16:53 | last interacted item feature\n",
            "15:16:53 | change price datatype\n",
            "15:16:53 | target column\n",
            "15:16:53 | lightGBM format\n",
            "15:16:53 | training lightGBM model\n",
            "15:16:53 | start\n",
            "15:16:53 | filter interactions\n",
            "15:16:53 | cleaning\n",
            "15:16:53 | previous interactions\n",
            "15:16:53 | combining columns - impressions\n",
            "15:16:53 | combining columns - prices\n",
            "15:16:53 | explode arrays\n",
            "15:16:53 | interaction count\n",
            "15:16:53 | reduce to impressions\n",
            "15:16:53 | position feature\n",
            "15:16:53 | last interacted item feature\n",
            "15:16:53 | change price datatype\n",
            "15:16:53 | summarize recommendations\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RrtkaemEtpf7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "69d8d504-a4c1-43ec-9aef-3a180bdabeb1"
      },
      "source": [
        "df_recommendations"
      ],
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>session_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>step</th>\n",
              "      <th>item_recommendations</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>64BL89</td>\n",
              "      <td>3579f90</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>5002 5003 5010 5004 5001 5023</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64BL91F2</td>\n",
              "      <td>3779f92</td>\n",
              "      <td>10</td>\n",
              "      <td>2</td>\n",
              "      <td>5001 5004 5010 5014</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    user_id session_id  timestamp  step           item_recommendations\n",
              "0    64BL89    3579f90          6     2  5002 5003 5010 5004 5001 5023\n",
              "1  64BL91F2    3779f92         10     2            5001 5004 5010 5014"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 70
        }
      ]
    }
  ]
}