MOTOR_Tutorial_Notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wlFGbW3cj2cN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d0634c08-ccea-45ce-bbbf-91f61f65ab86"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting jax[cpu]==0.4.8\n",
            "  Downloading jax-0.4.8.tar.gz (1.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.2/1.2 MB\u001b[0m \u001b[31m5.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
            "  Preparing metadata (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: ml-dtypes>=0.0.3 in /usr/local/lib/python3.10/dist-packages (from jax[cpu]==0.4.8) (0.2.0)\n",
            "Requirement already satisfied: numpy>=1.21 in /usr/local/lib/python3.10/dist-packages (from jax[cpu]==0.4.8) (1.25.2)\n",
            "Requirement already satisfied: opt-einsum in /usr/local/lib/python3.10/dist-packages (from jax[cpu]==0.4.8) (3.3.0)\n",
            "Requirement already satisfied: scipy>=1.7 in /usr/local/lib/python3.10/dist-packages (from jax[cpu]==0.4.8) (1.11.4)\n",
            "Collecting jaxlib==0.4.7 (from jax[cpu]==0.4.8)\n",
            "  Downloading jaxlib-0.4.7-cp310-cp310-manylinux2014_x86_64.whl (66.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m66.1/66.1 MB\u001b[0m \u001b[31m9.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hBuilding wheels for collected packages: jax\n",
            "  Building wheel for jax (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for jax: filename=jax-0.4.8-py3-none-any.whl size=1439670 sha256=9504571df0355d5d38021a10287160344dc93c9517fe11ea92708695a85ebf91\n",
            "  Stored in directory: /root/.cache/pip/wheels/09/6f/35/a8fac8b61de8e0d9eb07988481528898561923e260b1fa7d2f\n",
            "Successfully built jax\n",
            "Installing collected packages: jaxlib, jax\n",
            "  Attempting uninstall: jaxlib\n",
            "    Found existing installation: jaxlib 0.4.26+cuda12.cudnn89\n",
            "    Uninstalling jaxlib-0.4.26+cuda12.cudnn89:\n",
            "      Successfully uninstalled jaxlib-0.4.26+cuda12.cudnn89\n",
            "  Attempting uninstall: jax\n",
            "    Found existing installation: jax 0.4.26\n",
            "    Uninstalling jax-0.4.26:\n",
            "      Successfully uninstalled jax-0.4.26\n",
            "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
            "chex 0.1.86 requires jax>=0.4.16, but you have jax 0.4.8 which is incompatible.\n",
            "flax 0.8.2 requires jax>=0.4.19, but you have jax 0.4.8 which is incompatible.\n",
            "orbax-checkpoint 0.4.4 requires jax>=0.4.9, but you have jax 0.4.8 which is incompatible.\u001b[0m\u001b[31m\n",
            "\u001b[0mSuccessfully installed jax-0.4.8 jaxlib-0.4.7\n",
            "Collecting femr[models]==0.1.16\n",
            "  Downloading femr-0.1.16-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.1/1.1 MB\u001b[0m \u001b[31m5.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.22 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (1.25.2)\n",
            "Requirement already satisfied: scipy>=1.6 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (1.11.4)\n",
            "Requirement already satisfied: scikit-learn>=0.24 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (1.2.2)\n",
            "Requirement already satisfied: tqdm>=4.60.0 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (4.66.2)\n",
            "Collecting zstandard>=0.18 (from femr[models]==0.1.16)\n",
            "  Downloading zstandard-0.22.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.4 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.4/5.4 MB\u001b[0m \u001b[31m12.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting icecream==2.1.3 (from femr[models]==0.1.16)\n",
            "  Downloading icecream-2.1.3-py2.py3-none-any.whl (8.4 kB)\n",
            "Collecting nptyping==2.4.1 (from femr[models]==0.1.16)\n",
            "  Downloading nptyping-2.4.1-py3-none-any.whl (36 kB)\n",
            "Requirement already satisfied: msgpack>=1.0.5 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (1.0.8)\n",
            "Collecting optax==0.1.4 (from femr[models]==0.1.16)\n",
            "  Downloading optax-0.1.4-py3-none-any.whl (154 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m154.9/154.9 kB\u001b[0m \u001b[31m17.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting dm-haiku==0.0.9 (from femr[models]==0.1.16)\n",
            "  Downloading dm_haiku-0.0.9-py3-none-any.whl (352 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m352.1/352.1 kB\u001b[0m \u001b[31m25.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: jax==0.4.8 in /usr/local/lib/python3.10/dist-packages (from femr[models]==0.1.16) (0.4.8)\n",
            "Requirement already satisfied: absl-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from dm-haiku==0.0.9->femr[models]==0.1.16) (1.4.0)\n",
            "Collecting jmp>=0.0.2 (from dm-haiku==0.0.9->femr[models]==0.1.16)\n",
            "  Downloading jmp-0.0.4-py3-none-any.whl (18 kB)\n",
            "Requirement already satisfied: tabulate>=0.8.9 in /usr/local/lib/python3.10/dist-packages (from dm-haiku==0.0.9->femr[models]==0.1.16) (0.9.0)\n",
            "Collecting colorama>=0.3.9 (from icecream==2.1.3->femr[models]==0.1.16)\n",
            "  Downloading colorama-0.4.6-py2.py3-none-any.whl (25 kB)\n",
            "Requirement already satisfied: pygments>=2.2.0 in /usr/local/lib/python3.10/dist-packages (from icecream==2.1.3->femr[models]==0.1.16) (2.16.1)\n",
            "Collecting executing>=0.3.1 (from icecream==2.1.3->femr[models]==0.1.16)\n",
            "  Downloading executing-2.0.1-py2.py3-none-any.whl (24 kB)\n",
            "Collecting asttokens>=2.0.1 (from icecream==2.1.3->femr[models]==0.1.16)\n",
            "  Downloading asttokens-2.4.1-py2.py3-none-any.whl (27 kB)\n",
            "Requirement already satisfied: ml-dtypes>=0.0.3 in /usr/local/lib/python3.10/dist-packages (from jax==0.4.8->femr[models]==0.1.16) (0.2.0)\n",
            "Requirement already satisfied: opt-einsum in /usr/local/lib/python3.10/dist-packages (from jax==0.4.8->femr[models]==0.1.16) (3.3.0)\n",
            "Requirement already satisfied: chex>=0.1.5 in /usr/local/lib/python3.10/dist-packages (from optax==0.1.4->femr[models]==0.1.16) (0.1.86)\n",
            "Requirement already satisfied: jaxlib>=0.1.37 in /usr/local/lib/python3.10/dist-packages (from optax==0.1.4->femr[models]==0.1.16) (0.4.7)\n",
            "Requirement already satisfied: typing-extensions>=3.10.0 in /usr/local/lib/python3.10/dist-packages (from optax==0.1.4->femr[models]==0.1.16) (4.11.0)\n",
            "Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn>=0.24->femr[models]==0.1.16) (1.4.0)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn>=0.24->femr[models]==0.1.16) (3.4.0)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from asttokens>=2.0.1->icecream==2.1.3->femr[models]==0.1.16) (1.16.0)\n",
            "INFO: pip is looking at multiple versions of chex to determine which version is compatible with other requirements. This could take a while.\n",
            "Collecting chex>=0.1.5 (from optax==0.1.4->femr[models]==0.1.16)\n",
            "  Downloading chex-0.1.85-py3-none-any.whl (95 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m95.1/95.1 kB\u001b[0m \u001b[31m13.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading chex-0.1.84-py3-none-any.whl (94 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m94.8/94.8 kB\u001b[0m \u001b[31m13.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading chex-0.1.83-py3-none-any.whl (94 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m94.7/94.7 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading chex-0.1.82-py3-none-any.whl (94 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m94.8/94.8 kB\u001b[0m \u001b[31m12.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: toolz>=0.9.0 in /usr/local/lib/python3.10/dist-packages (from chex>=0.1.5->optax==0.1.4->femr[models]==0.1.16) (0.12.1)\n",
            "Installing collected packages: zstandard, nptyping, jmp, executing, colorama, asttokens, icecream, dm-haiku, femr, chex, optax\n",
            "  Attempting uninstall: chex\n",
            "    Found existing installation: chex 0.1.86\n",
            "    Uninstalling chex-0.1.86:\n",
            "      Successfully uninstalled chex-0.1.86\n",
            "  Attempting uninstall: optax\n",
            "    Found existing installation: optax 0.2.2\n",
            "    Uninstalling optax-0.2.2:\n",
            "      Successfully uninstalled optax-0.2.2\n",
            "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
            "flax 0.8.2 requires jax>=0.4.19, but you have jax 0.4.8 which is incompatible.\u001b[0m\u001b[31m\n",
            "\u001b[0mSuccessfully installed asttokens-2.4.1 chex-0.1.82 colorama-0.4.6 dm-haiku-0.0.9 executing-2.0.1 femr-0.1.16 icecream-2.1.3 jmp-0.0.4 nptyping-2.4.1 optax-0.1.4 zstandard-0.22.0\n"
          ]
        }
      ],
      "source": [
        "# Step 1: Installing the software\n",
        "\n",
        "# MOTOR relies upon femr v1, a package for machine learning on EHR records.\n",
        "\n",
        "# In particular, MOTOR requires FEMR 0.1.16: https://github.com/som-shahlab/femr/tree/0.1.16\n",
        "\n",
        "# There are two versions of femr, CPU and GPU\n",
        "# For this tutorial we will use the CPU version\n",
        "\n",
        "# hhttps://github.com/som-shahlab/femr/tree/0.1.16\n",
        "\n",
        "# How to install CPU version\n",
        "!pip install --upgrade \"jax[cpu]==0.4.8\"\n",
        "!pip install \"femr[models] == 0.1.16\"\n",
        "\n",
        "\n",
        "# How to install GPU version\n",
        "#!pip install --upgrade \"jax[cuda11_pip]==0.4.8\" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html\n",
        "#!pip install \"femr_cuda[models] == 0.1.16\""
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 2: Downloading the artifacts\n",
        "\n",
        "# The real model is stored on HuggingFace and can be downloaded by cloning https://huggingface.co/StanfordShahLab/motor-t-base\n",
        "\n",
        "# Real data can be obtained by processing data with the FEMR 1.16 library: https://github.com/som-shahlab/femr/tree/0.1.16?tab=readme-ov-file#getting-started\n",
        "\n",
        "# However, this tutorial is going to be using a synthetic model + synthetic data\n",
        "\n",
        "# The commands to download the synthetic data\n",
        "!gdown https://drive.google.com/uc?id=1Yhsz47BG8SZGoIFx0Q_UGEKsfhi2aiDs -O synthetic_data.zip\n",
        "!unzip -oq synthetic_data.zip"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yh9cB9gL7H1P",
        "outputId": "1d24f690-0d2f-48da-eedc-b46f1a34f643"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1Yhsz47BG8SZGoIFx0Q_UGEKsfhi2aiDs\n",
            "To: /content/synthetic_data.zip\n",
            "100% 22.6M/22.6M [00:00<00:00, 61.6MB/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 3: Understanding the artifacts\n",
        "\n",
        "!ls synthetic_data\n",
        "\n",
        "# clmbr / motor are the models\n",
        "\n",
        "# extract_lite is the features\n"
      ],
      "metadata": {
        "id": "Dq89rEGG7-a3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bc4ab13c-fa51-43bd-b1fd-d90d419341a5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "clmbr  extract_lite  motor  train_person_ids.csv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 3b: Understanding extract_lite\n",
        "\n",
        "import femr.datasets\n",
        "\n",
        "\n",
        "# The database contains all the features used for the patients\n",
        "database = femr.datasets.PatientDatabase('synthetic_data/extract_lite')\n",
        "\n",
        "patients = list(database)\n",
        "\n",
        "\n",
        "# For example, we can load a single patient with patient_id 110\n",
        "print(patients[10])\n",
        "\n",
        "patient = database[patients[10]]\n",
        "\n",
        "# Print out all the events for the patients\n",
        "for event in patient.events:\n",
        "  print(event)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zp2VgjdLUJ7L",
        "outputId": "357553d8-bd1b-409a-c73d-06b3b25d6eca"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "110\n",
            "Event(start=1992-02-04 00:00:00, code=Birth/Birth, value=None, visit_ids=1, clarity_source=PATIENT)\n",
            "Event(start=1992-02-04 00:00:00, code=Gender/Gender, value=F, visit_ids=1, clarity_source=PATIENT)\n",
            "Event(start=1992-02-04 00:00:00, code=Race/Race, value=White, visit_ids=1, clarity_source=PATIENT)\n",
            "Event(start=1992-04-03 00:00:00, code=ICD10CM/VX5IL6SO, value=None, visit_ids=13, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-05-08 00:00:00, code=ICD10CM/NJ9WA5CL, value=None, visit_ids=16, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-06-01 00:00:00, code=ICD10CM/9NFBZAEY, value=None, visit_ids=9, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-07-26 00:00:00, code=ICD10CM/VGT3XQIL, value=None, visit_ids=4, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-08-25 00:00:00, code=ICD10CM/57LXJUED, value=None, visit_ids=13, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-11-10 00:00:00, code=ICD10CM/5TPJBYZD, value=None, visit_ids=12, clarity_source=DIAGNOSIS)\n",
            "Event(start=1992-12-22 00:00:00, code=ICD10CM/TZZCVHUD, value=None, visit_ids=20, clarity_source=DIAGNOSIS)\n",
            "Event(start=1993-01-11 00:00:00, code=CPT/KERFNLH3, value=None, visit_ids=19, clarity_source=PROCEDURES)\n",
            "Event(start=1993-02-16 00:00:00, code=CPT/VC2GE5DV, value=None, visit_ids=3, clarity_source=PROCEDURES)\n",
            "Event(start=1993-04-25 00:00:00, code=CPT/O8VXPCRZ, value=None, visit_ids=6, clarity_source=PROCEDURES)\n",
            "Event(start=1993-07-02 00:00:00, code=CPT/P8P3ZNVX, value=None, visit_ids=3, clarity_source=PROCEDURES)\n",
            "Event(start=1993-08-26 00:00:00, code=CPT/AT8HVO0M, value=None, visit_ids=1, clarity_source=PROCEDURES)\n",
            "Event(start=1993-11-30 00:00:00, code=CPT/XZ7Q2FUL, value=None, visit_ids=10, clarity_source=PROCEDURES)\n",
            "Event(start=1994-01-04 00:00:00, code=CPT/IOSOCE9T, value=None, visit_ids=10, clarity_source=PROCEDURES)\n",
            "Event(start=1994-03-23 00:00:00, code=Drug/BK2F6I3J, value=None, dosage=45, visit_ids=5, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1994-05-02 00:00:00, code=Drug/C6RELHBH, value=None, dosage=33, visit_ids=9, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1994-05-07 00:00:00, code=Drug/M3SWQXK7, value=None, dosage=19, visit_ids=4, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1994-07-09 00:00:00, code=Drug/Q2UJDOQ7, value=None, dosage=38, visit_ids=5, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1994-09-21 00:00:00, code=Drug/K928X18C, value=None, dosage=33, visit_ids=13, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1994-10-29 00:00:00, code=Drug/P0YHRTSX, value=None, dosage=45, visit_ids=6, lab_units=mg, clarity_source=MED_ORDER)\n",
            "Event(start=1995-01-06 00:00:00, code=Vitals/KPFUIMIW, value=137.0, visit_ids=9, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1995-01-13 00:00:00, code=Vitals/3DF185D7, value=124.0, visit_ids=19, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1995-04-11 00:00:00, code=Vitals/IKA1DGNX, value=124.0, visit_ids=7, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1995-07-04 00:00:00, code=Vitals/WLEU9LQT, value=81.0, visit_ids=12, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1995-09-08 00:00:00, code=Vitals/NSBUV2VX, value=189.0, visit_ids=19, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1995-11-09 00:00:00, code=Vitals/N7X8J96D, value=80.0, visit_ids=18, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1996-02-13 00:00:00, code=Vitals/JYLI4IM8, value=145.0, visit_ids=12, lab_units=mmHg, clarity_source=LAB_RESULT)\n",
            "Event(start=1996-03-23 00:00:00, code=Vitals/IN89QRIE, value=115.0, visit_ids=0, lab_units=mmHg, clarity_source=LAB_RESULT)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Generate synthetic prediction times\n",
        "\n",
        "import random\n",
        "import csv\n",
        "\n",
        "random.seed(4533421)\n",
        "\n",
        "with open('prediction_times.csv', 'w') as f:\n",
        "  writer = csv.writer(f)\n",
        "  writer.writerow(['patient_id', 'prediction_time'])\n",
        "\n",
        "  for patient_id in database:\n",
        "    patient = database[patient_id]\n",
        "    times = [event.start for event in patient.events][4:]\n",
        "    chosen_times = list(set(random.choices(times, k=2)))\n",
        "\n",
        "    for chosen_time in chosen_times:\n",
        "      writer.writerow([patient_id, chosen_time.isoformat()])"
      ],
      "metadata": {
        "id": "17TsOWFa8Myf",
        "cellView": "form"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 4: Choosing prediction times\n",
        "\n",
        "!head prediction_times.csv\n",
        "\n",
        "# We expect a csv with a simple format with two columns patient_id, prediciton_time\n",
        "\n",
        "# For OMOP datasets, patient_id = person_id\n",
        "# prediction_time indicates what features to use for that patient\n",
        "\n",
        "!wc -l prediction_times.csv\n",
        "\n",
        "# Note that a single patient can have multiple prediction times"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8yDdRJlOYpog",
        "outputId": "ccb8a172-bee8-44ed-f315-641e20fe0f1a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "patient_id,prediction_time\r\n",
            "3,2020-08-09T00:00:00\r\n",
            "3,2022-06-05T00:00:00\r\n",
            "100,1991-03-03T00:00:00\r\n",
            "100,1992-05-26T00:00:00\r\n",
            "101,1992-11-02T00:00:00\r\n",
            "101,1993-11-09T00:00:00\r\n",
            "102,1993-03-10T00:00:00\r\n",
            "102,1994-02-05T00:00:00\r\n",
            "104,1991-08-13T00:00:00\r\n",
            "393 prediction_times.csv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# For example, let's consider patient 3\n",
        "\n",
        "patient = database[3]\n",
        "\n",
        "# Print out all the events for the patients\n",
        "for event in patient.events:\n",
        "  print(event)\n",
        "\n",
        "# We have two prediction times for this patient 2020-08-09T00:00:00 and 2022-06-05T00:00:00\n",
        "\n",
        "# The first will use all data up until the Vitals/HbA1c event\n",
        "\n",
        "# The second will use all data up till Drug/Atorvastatin"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5zjRzSdAhF96",
        "outputId": "73aeafa5-1f27-4846-ff82-8b25fcd461ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Event(start=1970-01-07 00:00:00, code=Birth/Birth, value=None)\n",
            "Event(start=1990-01-07 00:00:00, code=Gender/Gender, value=Female)\n",
            "Event(start=1990-01-07 00:00:00, code=Race/Race, value=White)\n",
            "Event(start=2020-07-09 00:00:00, code=Vitals/Blood Pressure, value=160.0, units=mmHg)\n",
            "Event(start=2020-08-09 00:00:00, code=Vitals/HbA1c, value=7.0, dosage=%)\n",
            "Event(start=2022-05-03 00:00:00, code=ICD10CM/E11.4, value=None)\n",
            "Event(start=2022-06-05 00:00:00, code=ICD10CM/E10.1, value=None)\n",
            "Event(start=2022-06-05 00:00:00, code=Note/ProgressNote, value=Patient Bob came to the clinic today)\n",
            "Event(start=2022-06-05 00:00:00, code=Drug/Atorvastatin, value=None, units=mg, dosage=50)\n",
            "Event(start=2022-06-06 00:00:00, code=Note/ProgressNote, value=Complicated notes generally need escaping , \"\n",
            " example)\n",
            "Event(start=2022-07-06 00:00:00, code=Drug/Multivitamins, value=None, units=ml, dosage=5)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 5: Computing representations\n",
        "\n",
        "!ls synthetic_data/motor\n",
        "\n",
        "# Note that this will be much faster if ran on a GPU\n",
        "\n",
        "!femr_compute_representations --data_path synthetic_data/extract_lite --model_path synthetic_data/motor --prediction_times_path prediction_times.csv --batch_size 1024 motor_reprs.pkl"
      ],
      "metadata": {
        "id": "A_Ee3auV8og_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "299b5a78-c9e7-4b53-fdc6-8131714ebfca"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dictionary  model\n",
            "2024-04-30 16:44:20,485 [MainThread  ] [INFO ]  Preparing batches with Namespace(directory='/tmp/tmpemjcawvy/task_batches', data_path='synthetic_data/extract_lite', dictionary_path='synthetic_data/motor/dictionary', task='labeled_patients', transformer_vocab_size=2048, clmbr_survival_dictionary_path=None, labeled_patients_path='prediction_times.csv', is_hierarchical=False, seed=97, val_start=70, test_start=85, batch_size=1024, note_embedding_data=None, limit_to_patients_file=None, limit_before_date=None, num_clmbr_tasks=8192)\n",
            "2024-04-30 16:44:20,529 [MainThread  ] [INFO ]  Wrote config ...\n",
            "2024-04-30 16:44:20,529 [MainThread  ] [INFO ]  Starting to load\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "0 out of 2048\n",
            "When mapping codes, dropped When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped When mapping codes, dropped 0 out of 02048 out of 2048\n",
            "\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "2024-04-30 16:44:20,868 [MainThread  ] [INFO ]  Loaded\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "2024-04-30 16:44:21,007 [MainThread  ] [INFO ]  Number of train patients 5\n",
            "No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)\n",
            "When mapping codes, dropped 0 out of 2048\n",
            "Computing reps 2024-04-30 16:44:21.871497\n",
            "Starting batches train 2024-04-30 16:44:21.871834\n",
            "Starting to process train 2024-04-30 16:44:21.871860\n",
            "Compiling the transformer ... (1024,) (256,)\n",
            "Starting batches dev 2024-04-30 16:44:25.026443\n",
            "Starting to process dev 2024-04-30 16:44:25.026506\n",
            "Starting batches test 2024-04-30 16:44:25.560646\n",
            "Starting to process test 2024-04-30 16:44:25.560706\n",
            "About to concat 1 2024-04-30 16:44:25.984356\n",
            "About to concat 2 2024-04-30 16:44:26.069108\n",
            "About to concat 3 2024-04-30 16:44:26.069199\n",
            "About to sort labels 2024-04-30 16:44:26.069547 392\n",
            "Done sorting labels 2024-04-30 16:44:26.074552\n",
            "About to sort representations 2024-04-30 16:44:26.074638 323\n",
            "Done sorting representations 2024-04-30 16:44:26.074729\n",
            "Done sorting labels and representations 2024-04-30 16:44:26.074792\n",
            "Creating representations\n",
            "Computing dates\n",
            "About to save 2024-04-30 16:44:26.370424\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Inspect the format of the result\n",
        "\n",
        "import pickle\n",
        "\n",
        "with open('motor_reprs.pkl', 'rb') as f:\n",
        "    motor_reprs = pickle.load(f)\n",
        "\n",
        "\n",
        "# We have three keys here, patient_ids, prediction_times, and representations\n",
        "\n",
        "# Note the representations for this model are of size 257\n",
        "\n",
        "print(motor_reprs.keys())\n",
        "\n",
        "print('-' * 80)\n",
        "\n",
        "for k, v in motor_reprs.items():\n",
        "    print(k, v.shape)\n",
        "\n",
        "print('-' * 80)\n",
        "\n",
        "for k, v in motor_reprs.items():\n",
        "    if len(v.shape) == 2:\n",
        "        print(k, v[:10, :])\n",
        "    else:\n",
        "        print(k, v[:10])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7cCM702WAbEs",
        "outputId": "366d672e-fc42-494a-aaa3-8f3f88a10385"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dict_keys(['representations', 'patient_ids', 'prediction_times'])\n",
            "--------------------------------------------------------------------------------\n",
            "representations (392, 257)\n",
            "patient_ids (392,)\n",
            "prediction_times (392,)\n",
            "--------------------------------------------------------------------------------\n",
            "representations [[ 0.1548  0.1927  1.495  ...  1.805  -0.927   1.    ]\n",
            " [ 0.64   -1.572   0.3352 ... -0.4333 -1.084   1.    ]\n",
            " [-0.8384  0.31    1.224  ... -0.2367  1.393   1.    ]\n",
            " ...\n",
            " [-0.8564  1.88    0.6196 ...  0.317   0.2336  1.    ]\n",
            " [ 1.054   0.2644 -0.1779 ... -1.586  -1.703   1.    ]\n",
            " [ 1.054   0.2644 -0.1779 ... -1.586  -1.703   1.    ]]\n",
            "patient_ids [  3   3 100 100 101 101 102 102 103 103]\n",
            "prediction_times [datetime.datetime(2020, 8, 9, 0, 0) datetime.datetime(2022, 6, 5, 0, 0)\n",
            " datetime.datetime(1991, 3, 3, 0, 0) datetime.datetime(1992, 5, 26, 0, 0)\n",
            " datetime.datetime(1992, 11, 2, 0, 0) datetime.datetime(1993, 11, 9, 0, 0)\n",
            " datetime.datetime(1993, 3, 10, 0, 0) datetime.datetime(1994, 2, 5, 0, 0)\n",
            " datetime.datetime(1993, 8, 6, 0, 0) datetime.datetime(1996, 4, 29, 0, 0)]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Generate synthetic labels\n",
        "\n",
        "import random\n",
        "import csv\n",
        "import numpy as np\n",
        "import scipy\n",
        "\n",
        "random.seed(4533421)\n",
        "np.random.seed(3421342)\n",
        "\n",
        "representation_size = motor_reprs['representations'].shape[1]\n",
        "\n",
        "weight = np.random.normal(size=(representation_size,))\n",
        "\n",
        "logits = np.dot(motor_reprs['representations'], weight)\n",
        "\n",
        "probabilities = scipy.special.expit(logits)\n",
        "\n",
        "sampled_probabilties = np.random.rand(len(logits))\n",
        "\n",
        "labels = sampled_probabilties < probabilities\n",
        "\n",
        "with open('labels.csv', 'w') as f:\n",
        "  writer = csv.writer(f)\n",
        "  writer.writerow(['patient_id', 'prediction_time', 'value'])\n",
        "\n",
        "  for patient_id, prediction_time, label in zip(motor_reprs['patient_ids'], motor_reprs['prediction_times'], labels):\n",
        "    writer.writerow([patient_id, prediction_time, label])"
      ],
      "metadata": {
        "cellView": "form",
        "id": "pH_7UDdyZIcG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 6: Computing labels\n",
        "\n",
        "# We have to align our labels with our representations\n",
        "\n",
        "import datetime\n",
        "\n",
        "\n",
        "# We start by assuming labels have already been generated\n",
        "!wc -l labels.csv\n",
        "\n",
        "# Simple csv with a colum value for the label\n",
        "\n",
        "!head labels.csv\n",
        "\n",
        "# Load the labels into a Python dictionary\n",
        "\n",
        "label_map = {}\n",
        "\n",
        "with open('labels.csv') as f:\n",
        "    reader = csv.DictReader(f)\n",
        "    for row in reader:\n",
        "        label_map[int(row['patient_id']), datetime.datetime.fromisoformat(row['prediction_time'])] = row['value'] == 'True'\n",
        "\n",
        "# Now we have to align the labels with our representations\n",
        "\n",
        "labels = []\n",
        "\n",
        "for patient_id, prediction_time in zip(motor_reprs['patient_ids'], motor_reprs['prediction_times']):\n",
        "    labels.append(label_map[patient_id, prediction_time])\n",
        "\n",
        "print(len(labels))\n",
        "\n",
        "labels = np.array(labels)"
      ],
      "metadata": {
        "id": "cfoZyAX5DLa5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3425200b-53cf-454e-c71e-7cc207a50bc5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "393 labels.csv\n",
            "patient_id,prediction_time,value\n",
            "3,2020-08-09 00:00:00,True\n",
            "3,2022-06-05 00:00:00,False\n",
            "100,1991-03-03 00:00:00,True\n",
            "100,1992-05-26 00:00:00,False\n",
            "101,1992-11-02 00:00:00,True\n",
            "101,1993-11-09 00:00:00,True\n",
            "102,1993-03-10 00:00:00,True\n",
            "102,1994-02-05 00:00:00,False\n",
            "103,1993-08-06 00:00:00,True\n",
            "392\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 7: Constructing a held-out test set\n",
        "\n",
        "patient_ids = list(set(motor_reprs['patient_ids']))\n",
        "\n",
        "random.seed(43534)\n",
        "patient_ids.sort()\n",
        "random.shuffle(patient_ids)\n",
        "\n",
        "test_patients = patient_ids[:len(patient_ids) // 2]\n",
        "\n",
        "print(\"Chosen test patients\", len(test_patients))"
      ],
      "metadata": {
        "id": "gBqeHze5DNen",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5bbf867e-bce7-4304-a1ca-d655adec56fc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Chosen test patients 196\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 8: Splitting features/labels into train / held out test\n",
        "\n",
        "test_mask = np.isin(motor_reprs['patient_ids'], test_patients)\n",
        "\n",
        "train_features = motor_reprs['representations'][~test_mask, :]\n",
        "train_labels =labels[~test_mask]\n",
        "\n",
        "test_features = motor_reprs['representations'][test_mask, :]\n",
        "test_labels = labels[test_mask]\n",
        "\n",
        "print(\"Train size\", len(train_labels))\n",
        "\n",
        "print(\"Test size\", len(test_labels))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "brbPzkjdfVS7",
        "outputId": "93c32055-6ece-481c-81f5-6e691fd2794a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train size 81\n",
            "Test size 311\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 8: Training a logistic regression model\n",
        "\n",
        "import sklearn.linear_model\n",
        "\n",
        "model = sklearn.linear_model.LogisticRegressionCV()\n",
        "model.fit(train_features, train_labels)"
      ],
      "metadata": {
        "id": "5gESjyAJ8qWX",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 74
        },
        "outputId": "7d2e5570-f26f-4b5c-af2b-f1fca82fad0b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LogisticRegressionCV()"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>LogisticRegressionCV()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">LogisticRegressionCV</label><div class=\"sk-toggleable__content\"><pre>LogisticRegressionCV()</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 9: Performance analysis\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "probabilities = model.predict_log_proba(test_features)[:, 1]\n",
        "\n",
        "\n",
        "display =  sklearn.metrics.PrecisionRecallDisplay.from_predictions(\n",
        "    labels[test_mask], probabilities,\n",
        ")\n",
        "plt.title(\"Precision Recall Curve\")\n",
        "\n",
        "\n",
        "display =  sklearn.metrics.RocCurveDisplay.from_predictions(\n",
        "    labels[test_mask], probabilities,\n",
        ")\n",
        "plt.title(\"Reciever Operator Curve\")"
      ],
      "metadata": {
        "id": "52rDFNz_DVRI",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 944
        },
        "outputId": "b8460716-a7b5-4831-e8ed-bf1581600c99"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Reciever Operator Curve')"
            ]
          },
          "metadata": {},
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}