TestAcoustic.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "TestAcoustic.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "machine_shape": "hm"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.9"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "68326de829dd4b8d9ade8e1fdbec5d88": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_db3d8151b64f45369ef99a74b1c0d52b",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_ed0ee8dace74483a8a2cc0f89c9ba662",
              "IPY_MODEL_f09fd0d4d6914530900090398e499753"
            ]
          }
        },
        "db3d8151b64f45369ef99a74b1c0d52b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "ed0ee8dace74483a8a2cc0f89c9ba662": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_820627198c7c4a3c93fd4f03eb4699aa",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 2497,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 2497,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_0753167690c94b37a30832a47939fb67"
          }
        },
        "f09fd0d4d6914530900090398e499753": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_5c562ea513bc4def874d79a1e3a97f11",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 2497/2497 [34:02&lt;00:00,  1.22it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f34ab18d58504f7ab23c924fb482ea1c"
          }
        },
        "820627198c7c4a3c93fd4f03eb4699aa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "0753167690c94b37a30832a47939fb67": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "5c562ea513bc4def874d79a1e3a97f11": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "f34ab18d58504f7ab23c924fb482ea1c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "05a3ff11cd5349d38ec5b828530e69fa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_8f711c9d3d2e4dd8ba6c09bfee43efa6",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_b06c2ce62df34933acb83a46e84e9806",
              "IPY_MODEL_4b43f53dc12e4d4a9f329eb1a0268ebd"
            ]
          }
        },
        "8f711c9d3d2e4dd8ba6c09bfee43efa6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "b06c2ce62df34933acb83a46e84e9806": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_765e40fceaaa4e0fb7b1a89d3749587e",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 250,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 250,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_276784a609974a55b1756b0999626781"
          }
        },
        "4b43f53dc12e4d4a9f329eb1a0268ebd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_a428dac3c756458598c3079c06160cae",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 250/250 [01:39&lt;00:00,  2.50it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6e0c9890ad9748859a7720899455f04e"
          }
        },
        "765e40fceaaa4e0fb7b1a89d3749587e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "276784a609974a55b1756b0999626781": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "a428dac3c756458598c3079c06160cae": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "6e0c9890ad9748859a7720899455f04e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qo6aKD2YLKMx"
      },
      "source": [
        "# Predict wavefield borders using the network combined with an FD solver"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_daezAfmLX5c"
      },
      "source": [
        "This notebook tests the NN's ability to continually predict wavefield borders during a simulation."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3ww0SW9BIICD"
      },
      "source": [
        "**Install necessary versions of packages**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vuLCrGRKQYdC",
        "outputId": "bbfae1fe-9683-4807-88be-3d8da688db64"
      },
      "source": [
        "# Install packages if not available yet\n",
        "!pip install git+https://github.com/devitocodes/devito.git --quiet\n",
        "!pip install sympy==1.7.0 --quiet\n",
        "!pip install numpy==1.17.0 --quiet"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[K     |████████████████████████████████| 81kB 2.8MB/s \n",
            "\u001b[K     |████████████████████████████████| 102kB 7.4MB/s \n",
            "\u001b[K     |████████████████████████████████| 143kB 32.5MB/s \n",
            "\u001b[K     |████████████████████████████████| 51kB 6.9MB/s \n",
            "\u001b[K     |████████████████████████████████| 194kB 50.3MB/s \n",
            "\u001b[K     |████████████████████████████████| 71kB 8.5MB/s \n",
            "\u001b[K     |████████████████████████████████| 51kB 7.0MB/s \n",
            "\u001b[K     |████████████████████████████████| 71kB 8.6MB/s \n",
            "\u001b[?25h  Building wheel for devito (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for py-cpuinfo (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for cgen (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for codepy (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for pyrevolve (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for pytools (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for contexttimer (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[31mERROR: pytest-cov 2.12.1 has requirement coverage>=5.2.1, but you'll have coverage 3.7.1 which is incompatible.\u001b[0m\n",
            "\u001b[31mERROR: pytest-cov 2.12.1 has requirement pytest>=4.6, but you'll have pytest 3.6.4 which is incompatible.\u001b[0m\n",
            "\u001b[K     |████████████████████████████████| 5.9MB 3.0MB/s \n",
            "\u001b[K     |████████████████████████████████| 20.3MB 1.3MB/s \n",
            "\u001b[31mERROR: tensorflow 2.5.0 has requirement numpy~=1.19.2, but you'll have numpy 1.17.0 which is incompatible.\u001b[0m\n",
            "\u001b[31mERROR: kapre 0.3.5 has requirement numpy>=1.18.5, but you'll have numpy 1.17.0 which is incompatible.\u001b[0m\n",
            "\u001b[31mERROR: datascience 0.10.6 has requirement folium==0.2.1, but you'll have folium 0.8.3 which is incompatible.\u001b[0m\n",
            "\u001b[31mERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible.\u001b[0m\n",
            "\u001b[?25h"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wvKkGyUOqar5",
        "outputId": "6f03b67e-e4e4-45e0-bf27-d5b7279a104d"
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fri Jun 18 16:11:02 2021       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 465.27       Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla V100-SXM2...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   35C    P0    23W / 300W |      0MiB / 16160MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sDYMkPwEQgWT",
        "outputId": "083616ea-cb8c-4500-fd77-ada291d47056"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive', force_remount=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lo4U6gon4gt4",
        "outputId": "1519d2d2-dd80-4551-d4f8-7da398a79cd5"
      },
      "source": [
        "%cd /content/drive/My\\ Drive/ML_Ultrasound_Project/UROPdata"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/drive/My Drive/ML_Ultrasound_Project/UROPdata\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JtbSrMWd14YS",
        "outputId": "e3f2ddce-953b-4ba8-bb19-04a00b3d813c"
      },
      "source": [
        "pip install -U fvcore --quiet"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[K     |████████████████████████████████| 51kB 2.1MB/s \n",
            "\u001b[K     |████████████████████████████████| 645kB 5.4MB/s \n",
            "\u001b[?25h  Building wheel for fvcore (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u5WqU00L0Zya"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "import copy\n",
        "\n",
        "from collections import deque\n",
        "\n",
        "import pprint\n",
        "from mpl_toolkits.axes_grid1 import make_axes_locatable\n",
        "\n",
        "import devito\n",
        "from sympy import Symbol\n",
        "from devito import Operator\n",
        "from devito import Eq, solve\n",
        "from devito import TimeFunction\n",
        "\n",
        "import torch\n",
        "from torch.autograd import Variable\n",
        "from torch.utils.data import TensorDataset\n",
        "from torch.utils.data import DataLoader\n",
        "from torch import nn\n",
        "\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "import cv2\n",
        "from scipy import stats\n",
        "from scipy.ndimage import gaussian_filter\n",
        "\n",
        "from tqdm.notebook import tqdm\n",
        "\n",
        "from networks.convlstm import RNN\n",
        "\n",
        "from medical.plotting import plot_shotrecord, plot_velocity\n",
        "from medical.source import ToneBurstSource, Receiver\n",
        "from medical.source import TimeAxis\n",
        "from medical.model import Model\n",
        "\n",
        "from fvcore.nn import FlopCountAnalysis"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h2Hp-gXsEHs3"
      },
      "source": [
        "**Global settings**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T0yUJDEKEKzG"
      },
      "source": [
        "src_index = 315      # From 0 to 512\n",
        "num = 2500            # Simulation last number of time steps\n",
        "model_type = 'skull'  # water, skull or full\n",
        "\n",
        "if src_index > 448 or src_index <= 64:\n",
        "  nearest_border = 'right'\n",
        "if src_index > 64 and src_index <= 192:\n",
        "  nearest_border = 'bottom'\n",
        "if src_index > 192 and src_index <=320:\n",
        "  nearest_border = 'left'\n",
        "if src_index > 320 and src_index <= 448:\n",
        "  nearest_border = 'top'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pWqQcmpZEKxN"
      },
      "source": [
        "def create_eliptic_array(n_sources, boundary_offset, dx, shape):\n",
        "    a = (shape[1] / 2)\n",
        "    b = (shape[0] / 2)\n",
        "\n",
        "    pos = np.empty((2, n_sources), dtype=np.float32)\n",
        "    angle = np.zeros((n_sources))\n",
        "    for idx in range(n_sources):\n",
        "        angle[idx] = ((idx) * 360.0 / n_sources) * 2 * np.pi / 360.0  # transducer angle\n",
        "        pos[1, idx] = (a * np.cos(angle[idx]) + shape[1] / 2)  # spherical equation of ellipse\n",
        "        pos[0, idx] = (b * np.sin(angle[idx]) + shape[0] / 2)  # spherical equation of ellipse\n",
        "\n",
        "    pos[0, :] += boundary_offset[0]\n",
        "    pos[1, :] += boundary_offset[1]\n",
        "    pos *= dx\n",
        "    # np.array((np.divide(grid[0] ** 2, major_ax ** 2) + np.divide(grid[1] ** 2, minor_ax ** 2)) < 1, dtype=np.float32)\n",
        "    return pos"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ya9h3zbyYUjZ"
      },
      "source": [
        "class MidpointNormalize(mpl.colors.Normalize):\n",
        "    def __init__(self, vmin, vmax, midpoint=0, clip=False):\n",
        "        self.midpoint = midpoint\n",
        "        mpl.colors.Normalize.__init__(self, vmin, vmax, clip)\n",
        "\n",
        "    def __call__(self, value, clip=None):\n",
        "        normalized_min = max(0, 1 / 2 * (1 - abs((self.midpoint - self.vmin) / (self.midpoint - self.vmax))))\n",
        "        normalized_max = min(1, 1 / 2 * (1 + abs((self.vmax - self.midpoint) / (self.midpoint - self.vmin))))\n",
        "        normalized_mid = 0.5\n",
        "        x, y = [self.vmin, self.midpoint, self.vmax], [normalized_min, normalized_mid, normalized_max]\n",
        "        return np.ma.masked_array(np.interp(value, x, y))\n",
        "\n",
        "class ModMidpointNormalize(mpl.colors.Normalize):\n",
        "    def __init__(self, vmin, vmax, midpoint=0, clip=False):\n",
        "        self.midpoint = midpoint\n",
        "        mpl.colors.Normalize.__init__(self, vmin, vmax, clip)\n",
        "\n",
        "    def __call__(self, value, clip=None):\n",
        "        normalized_min = max(0, 1 / 2 * (1 - abs((self.midpoint - self.vmin) / (self.midpoint - self.vmax))))\n",
        "        normalized_max = 1 #min(1, 1 / 2 * (1 + abs((self.vmax - self.midpoint) / (self.midpoint - self.vmin))))\n",
        "        normalized_mid = 0.5\n",
        "        x, y = [self.vmin, self.midpoint, self.vmax], [normalized_min, normalized_mid, normalized_max]\n",
        "        return np.ma.masked_array(np.interp(value, x, y))\n",
        "\n",
        "class OnesMidpointNormalize(mpl.colors.Normalize):\n",
        "    def __init__(self, vmin, vmax, midpoint=0, clip=False):\n",
        "        self.midpoint = midpoint\n",
        "        mpl.colors.Normalize.__init__(self, vmin, vmax, clip)\n",
        "\n",
        "    def __call__(self, value, clip=None):\n",
        "        normalized_min = -1 #max(0, 1 / 2 * (1 - abs((self.midpoint - self.vmin) / (self.midpoint - self.vmax))))\n",
        "        normalized_max = 1 #min(1, 1 / 2 * (1 + abs((self.vmax - self.midpoint) / (self.midpoint - self.vmin))))\n",
        "        normalized_mid = 0.5\n",
        "        x, y = [self.vmin, self.midpoint, self.vmax], [normalized_min, normalized_mid, normalized_max]\n",
        "        return np.ma.masked_array(np.interp(value, x, y))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BXqVAF7sIPdR"
      },
      "source": [
        "##Model Integration with Devito"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4D6NjGehz27b",
        "outputId": "3362bf38-ccac-4618-e9d9-e7862732f76b"
      },
      "source": [
        "# Load trained RNN state dict\n",
        "\n",
        "dev = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")\n",
        "\n",
        "numLayers = 3\n",
        "numFeatures = 32\n",
        "\n",
        "rnn = RNN(numLayers, numFeatures)\n",
        "rnn.to(dev)\n",
        "rnn.load_state_dict(torch.load('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/rnn46_state_dict_zeroed_225'))\n",
        "rnn.eval()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RNN(\n",
              "  (convlstm1): CLSTM_cell(\n",
              "    (conv): Sequential(\n",
              "      (0): Conv2d(33, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
              "      (1): GroupNorm(4, 128, eps=1e-05, affine=True)\n",
              "    )\n",
              "  )\n",
              "  (convlstm2): CLSTM_cell(\n",
              "    (conv): Sequential(\n",
              "      (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
              "      (1): GroupNorm(4, 128, eps=1e-05, affine=True)\n",
              "    )\n",
              "  )\n",
              "  (convlstm3): CLSTM_cell(\n",
              "    (conv): Sequential(\n",
              "      (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
              "      (1): GroupNorm(4, 128, eps=1e-05, affine=True)\n",
              "    )\n",
              "  )\n",
              "  (reshape_output): ProcessOut(\n",
              "    (tanh): Tanh()\n",
              "    (flat): Flatten(start_dim=1, end_dim=-1)\n",
              "    (linear): Linear(in_features=152352, out_features=11, bias=True)\n",
              "  )\n",
              "  (lambda_1): Lambda()\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fHPkOenNPJ6C",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "97e18e35-19cb-441d-9755-2b10eecbb352"
      },
      "source": [
        "devito.parameters.configuration['language'] = 'openmp'\n",
        "devito.parameters.configuration['mpi'] = False\n",
        "\n",
        "# By changing the src_index you'll be selecting a different source coordinate from the different 512 that have been created\n",
        "# By changing the model type to water, skull or full you'll get different background mediums\n",
        "# The cutting box has dimensions of (485, 355) in this case. This cannot change\n",
        "# The cutting box is centred in the medium\n",
        "# The complete medium is currently (1001, 1001) but you will have to make this bigger if you need to in order to\n",
        "# make sure that the wavefield does not come back into the cutting box after hitting the edges of the complete medium\n",
        "\n",
        "run_type = 'limited'  # infinite, limited\n",
        "\n",
        "# Infinite will run the operator like when you generated your dataset. When you plot the resulting wavefield\n",
        "# at the end, 'infinite' will look as if the wave is expanding into infinity. If you run it with 'limited'\n",
        "# however, you'll see how the wave bounces back against the edge. The goal is to make the 'limited' case\n",
        "# run as if it was the 'infinite' one by using your network.\n",
        "\n",
        "\n",
        "# Load different types of background models\n",
        "if model_type == 'water':\n",
        "    model = np.ones((480, 350)) * 1500.\n",
        "\n",
        "elif model_type == 'skull':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-ExactSkull.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "elif model_type == 'full':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-TrueVp.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "y_length = 480+66+4  #480+11\n",
        "x_length = 350+66+4  #350+11\n",
        "\n",
        "# Define a physical size\n",
        "if run_type == 'infinite':\n",
        "    shape = (1001, 1001)            # Number of grid point (nx, nz) - CHANGE THIS FOR THE MEDIUM SIZE\n",
        "else:\n",
        "    shape = (y_length, x_length)\n",
        "    \n",
        "spacing = (0.5e-3, 0.5e-3)      # Grid spacing in m.\n",
        "origin = (0., 0.)               # What is the location of the top left corner. This is necessary to define\n",
        "                                # the absolute location of the source and receivers\n",
        "\n",
        "# Define a velocity profile. The velocity is in m/s\n",
        "diff_x = shape[0] - model.shape[0]\n",
        "diff_y = shape[1] - model.shape[1]\n",
        "offset_x = int(np.floor(diff_x/2))\n",
        "offset_y = int(np.floor(diff_y/2))\n",
        "\n",
        "model = np.pad(model,\n",
        "               ((int(np.floor(diff_x/2)), int(np.ceil(diff_x/2))), \n",
        "                (int(np.floor(diff_y/2)), int(np.ceil(diff_y/2)))),\n",
        "               'edge')\n",
        "\n",
        "plt.imshow(np.transpose(model), origin='lower')\n",
        "\n",
        "# With the velocity and model size defined, we can create the seismic model that\n",
        "# encapsulates this properties. We also define the size of the absorbing layer as 10 grid points\n",
        "model = Model(vp=model, origin=origin, shape=shape, spacing=spacing, space_order=4, nbpml=0)\n",
        "\n",
        "t0 = 0.                 # Simulation starts a t=0\n",
        "dt = 0.08e-6            # Time step from model grid spacing\n",
        "\n",
        "time_range = TimeAxis(start=t0, num=num, step=dt)\n",
        "\n",
        "f0 = 0.40e6  # Source peak frequency is 400kHz\n",
        "src = ToneBurstSource(name='src', grid=model.grid, f0=f0, npoint=1, time_range=time_range)\n",
        "\n",
        "# First, position source centrally in all dimensions\n",
        "src_positions = create_eliptic_array(512, (offset_x, offset_y), spacing[0], (480, 350))\n",
        "src.coordinates.data[0, :] = src_positions[:, src_index]\n",
        "\n",
        "print(src.coordinates.data)\n",
        "\n",
        "plt.scatter(src_positions[0, :]/spacing[0], src_positions[1, :]/spacing[1])\n",
        "plt.show()\n",
        "\n",
        "# We can plot the time signature to see the wavelet\n",
        "src.show()\n",
        "\n",
        "# We can now show the source and receivers within our domain:\n",
        "# Red dot: Source location\n",
        "plot_velocity(model, source=src.coordinates.data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[0.05801011 0.03945057]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2U5XN5_924RU"
      },
      "source": [
        "# In order to represent the wavefield u and the square slowness we need symbolic objects\n",
        "# corresponding to time-space-varying field (u, TimeFunction) and\n",
        "# space-varying field (m, Function)\n",
        "\n",
        "# Define the wavefield with the size of the model and the time dimension\n",
        "u = TimeFunction(name=\"u\", grid=model.grid, time_order=2, space_order=4, save=num)\n",
        "\n",
        "# Create a temporary symbol for H to avoid expensive sympy solve\n",
        "H = Symbol('H')\n",
        "\n",
        "# Define PDE\n",
        "eq = model.m * u.dt2 - H\n",
        "\n",
        "# Solve the symbolic equation for the field to be updated\n",
        "eq_time = solve(eq, u.forward)\n",
        "\n",
        "# Get the spacial FD\n",
        "biharmonic = u.biharmonic(1/model.m)\n",
        "laplacian = u.laplace + dt**2/12 * biharmonic\n",
        "\n",
        "# This discrete PDE can be solved in a time-marching way updating u(t+dt) from the previous time step\n",
        "# Devito as a shortcut for u(t+dt) which is u.forward. We can then rewrite the PDE as\n",
        "# a time marching updating equation known as a stencil using customized SymPy functions\n",
        "stencil = Eq(u.forward, eq_time.subs({H: laplacian}))\n",
        "\n",
        "# Finally we define the source injection function to generate the corresponding code\n",
        "src_term = src.inject(field=u.forward, expr=src * dt**2 / model.m)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_p9oU_vz-INJ"
      },
      "source": [
        "def execute_operator(kwargs, time_min=None, time_max=None):\n",
        "    if time_min is None:\n",
        "        time_min = 1\n",
        "\n",
        "    if time_max is None:\n",
        "        time_max = num-1\n",
        "\n",
        "    kwargs['time_m'] = time_min\n",
        "    kwargs['time_M'] = time_max\n",
        "    \n",
        "    args = op.arguments(**kwargs)\n",
        "\n",
        "    arg_values = [args[p.name] for p in op.parameters]\n",
        "    try:\n",
        "        op.cfunction(*arg_values)\n",
        "\n",
        "    except ctypes.ArgumentError as e:\n",
        "        if e.args[0].startswith(\"argument \"):\n",
        "            argnum = int(e.args[0][9:].split(':')[0]) - 1\n",
        "            newmsg = \"error in argument '%s' with value '%s': %s\" % (\n",
        "                self._operator.parameters[argnum].name,\n",
        "                arg_values[argnum],\n",
        "                e.args[0])\n",
        "            raise ctypes.ArgumentError(newmsg) from e\n",
        "        else:\n",
        "            raise\n",
        "            \n",
        "    op._postprocess_arguments(args, **kwargs)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U0Nly1P42Rzg"
      },
      "source": [
        "def get_network_inputs(w, y_len, x_len, top, right, bottom, left, t, norm):\n",
        "\n",
        "  inputs = []\n",
        "\n",
        "  # top row\n",
        "  for i in range(y_len-22):\n",
        "    square_timestep0_T = w[t-48][top][i:i+23, 11:34]\n",
        "    square_timestep1_T = w[t-42][top][i:i+23, 11:34]\n",
        "    square_timestep2_T = w[t-36][top][i:i+23, 11:34]\n",
        "    square_timestep3_T = w[t-30][top][i:i+23, 11:34]\n",
        "    square_timestep4_T = w[t-24][top][i:i+23, 11:34]\n",
        "    square_timestep5_T = w[t-18][top][i:i+23, 11:34]\n",
        "    square_timestep6_T = w[t-12][top][i:i+23, 11:34]\n",
        "    square_timestep7_T = w[t-6][top][i:i+23, 11:34]\n",
        "    square_timestep8_T = w[t][top][i:i+23, 11:34]\n",
        "\n",
        "    #link the 5 images together\n",
        "    inputs.append(np.stack((square_timestep0_T, square_timestep1_T, square_timestep2_T, square_timestep3_T, square_timestep4_T,\n",
        "                                square_timestep5_T, square_timestep6_T, square_timestep7_T, square_timestep8_T), axis = 0))\n",
        "\n",
        "  # right column\n",
        "  for i in range(x_len-22):\n",
        "    square_timestep0_T = np.rot90(w[t-48][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep1_T = np.rot90(w[t-42][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep2_T = np.rot90(w[t-36][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep3_T = np.rot90(w[t-30][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep4_T = np.rot90(w[t-24][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep5_T = np.rot90(w[t-18][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep6_T = np.rot90(w[t-12][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep7_T = np.rot90(w[t-6][right], 3)[i:i+23, 11:34]\n",
        "    square_timestep8_T = np.rot90(w[t][right], 3)[i:i+23, 11:34]\n",
        "\n",
        "    inputs.append(np.stack((square_timestep0_T, square_timestep1_T, square_timestep2_T, square_timestep3_T, square_timestep4_T,\n",
        "                                square_timestep5_T, square_timestep6_T, square_timestep7_T, square_timestep8_T), axis = 0))\n",
        "\n",
        "  # bottom row\n",
        "  for i in range(y_len-22):\n",
        "    square_timestep0_T = np.rot90(w[t-48][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep1_T = np.rot90(w[t-42][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep2_T = np.rot90(w[t-36][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep3_T = np.rot90(w[t-30][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep4_T = np.rot90(w[t-24][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep5_T = np.rot90(w[t-18][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep6_T = np.rot90(w[t-12][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep7_T = np.rot90(w[t-6][bottom], 2)[i:i+23, 11:34]\n",
        "    square_timestep8_T = np.rot90(w[t][bottom], 2)[i:i+23, 11:34]\n",
        "\n",
        "    inputs.append(np.stack((square_timestep0_T, square_timestep1_T, square_timestep2_T, square_timestep3_T, square_timestep4_T,\n",
        "                                square_timestep5_T, square_timestep6_T, square_timestep7_T, square_timestep8_T), axis = 0))\n",
        "\n",
        "  # left column\n",
        "  for i in range(x_len-22):\n",
        "    square_timestep0_T = np.rot90(w[t-48][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep1_T = np.rot90(w[t-42][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep2_T = np.rot90(w[t-36][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep3_T = np.rot90(w[t-30][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep4_T = np.rot90(w[t-24][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep5_T = np.rot90(w[t-18][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep6_T = np.rot90(w[t-12][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep7_T = np.rot90(w[t-6][left], 1)[i:i+23, 11:34]\n",
        "    square_timestep8_T = np.rot90(w[t][left], 1)[i:i+23, 11:34]\n",
        "\n",
        "    inputs.append(np.stack((square_timestep0_T, square_timestep1_T, square_timestep2_T, square_timestep3_T, square_timestep4_T,\n",
        "                                square_timestep5_T, square_timestep6_T, square_timestep7_T, square_timestep8_T), axis = 0))\n",
        "\n",
        "  return torch.from_numpy(np.asarray(inputs)/norm).cuda().view(-1, 9, 1, 23, 23)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ILLbKe84tn8y"
      },
      "source": [
        "# Generator\n",
        "class WrappedDataLoader:\n",
        "    def __init__(self, dl, func):\n",
        "        self.dl = dl\n",
        "        self.func = func\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.dl)\n",
        "\n",
        "    def __iter__(self):\n",
        "        batches = iter(self.dl)\n",
        "        for b in batches:\n",
        "            yield (self.func(*b))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7vW8-CDNtpjM"
      },
      "source": [
        "def preprocess(x):\n",
        "    return x.view(-1, 9, 1, 23, 23)#.to(dev)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n4459lhBtxfc",
        "outputId": "2e653d0c-bfbe-4227-f382-6e986205aa52"
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mon Jun 21 10:42:39 2021       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 465.27       Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla V100-SXM2...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   36C    P0    40W / 300W |   1363MiB / 16160MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67,
          "referenced_widgets": [
            "68326de829dd4b8d9ade8e1fdbec5d88",
            "db3d8151b64f45369ef99a74b1c0d52b",
            "ed0ee8dace74483a8a2cc0f89c9ba662",
            "f09fd0d4d6914530900090398e499753",
            "820627198c7c4a3c93fd4f03eb4699aa",
            "0753167690c94b37a30832a47939fb67",
            "5c562ea513bc4def874d79a1e3a97f11",
            "f34ab18d58504f7ab23c924fb482ea1c"
          ]
        },
        "id": "QobNu-7lQYdD",
        "outputId": "b6b26e1a-5758-49d3-c8d6-244ba1c6ce77"
      },
      "source": [
        "top = (slice(0,None), slice(0,34))\n",
        "right = (slice(-34,None), slice(0,None))\n",
        "bottom = (slice(0,None), slice(-34,None))\n",
        "left = (slice(0,34), slice(0,None))\n",
        "\n",
        "norm = 3.005867554861652e-08\n",
        "\n",
        "# Prepare the operator\n",
        "u.data.fill(0.)\n",
        "op = Operator([stencil] + src_term, subs=model.spacing_map)\n",
        "    \n",
        "kwargs = {'dt': dt}\n",
        "\n",
        "shape = (1001, 1001) #(1001, 1001)            # Number of grid point (nx, nz) - CHANGE THIS FOR THE MEDIUM SIZE\n",
        "size_diff = int((shape[0]-1001)/2)\n",
        "\n",
        "# Execute the operator\n",
        "for t in tqdm(range(1, num-2)):\n",
        "  \n",
        "  borders = get_network_inputs(np.asarray(u.data, dtype=np.float32), y_length, x_length, top, right, bottom, left, t+1, norm)\n",
        "\n",
        "  with torch.no_grad():\n",
        "    predictions = rnn(borders).cpu().detach()\n",
        "    predictions *= norm\n",
        "\n",
        "  execute_operator(kwargs, t, t+1)\n",
        "\n",
        "  u.data[t+1][11:y_length-11,:11] = predictions[:y_length-22]\n",
        "  u.data[t+1][-11:,11:x_length-11] = torch.rot90(predictions[y_length-22:y_length-22 + x_length-22], 1)\n",
        "  u.data[t+1][11:y_length-11,-11:] = torch.rot90(predictions[y_length-22 + x_length-22:y_length-22 + x_length-22 + y_length-22], 2)\n",
        "  u.data[t+1][:11,11:x_length-11] = torch.rot90(predictions[y_length-22 + x_length-22 + y_length-22:], 3)\n",
        "\n",
        "  # Make corners zero\n",
        "  u.data[t+1][:11,:11] = 0\n",
        "  u.data[t+1][-11:,:11] = 0\n",
        "  u.data[t+1][:11, -11:] = 0\n",
        "  u.data[t+1][-11:, -11:] = 0\n",
        "\n",
        "    # print('=========================================================================================================')\n",
        "    # print('Timestep ',t)\n",
        "\n",
        "    # plt.figure(figsize=(15,6))\n",
        "    # plt.title('Wavefield with predicted borders')\n",
        "    # plt.imshow(np.transpose(u.data[t+1][:,:]), cmap = 'seismic', norm=ModMidpointNormalize(vmin=np.max(v), vmax=np.min(v), midpoint=0))\n",
        "    # plt.colorbar()\n",
        "    # plt.show()\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "68326de829dd4b8d9ade8e1fdbec5d88",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=2497.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SIb6WhGP2PpT"
      },
      "source": [
        "borders = get_network_inputs(np.zeros((2500,y_length,x_length), dtype=np.float32), y_length, x_length, top, right, bottom, left, 50, norm)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VaettwsE2Ce_",
        "outputId": "3e159634-afe1-4667-d628-6db723c367b5"
      },
      "source": [
        "flops = FlopCountAnalysis(rnn, borders)\n",
        "flops.total()\n",
        "# flops.by_module_and_operator()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Unsupported operator aten::sigmoid encountered 81 time(s)\n",
            "Unsupported operator aten::tanh encountered 55 time(s)\n",
            "Unsupported operator aten::mul encountered 81 time(s)\n",
            "Unsupported operator aten::add encountered 27 time(s)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "109123795872"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DbbSJ1Eb09yJ"
      },
      "source": [
        "np.savez_compressed('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/modelled_waves_3x32_{}'.format(src_index), u.data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QX6gKiCz-Fm1"
      },
      "source": [
        "##Original 'Infinite' Devito Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "qq9Y_rcK-FnE",
        "outputId": "4216e8cb-3a10-41f7-c71d-bc68a8d58c8f"
      },
      "source": [
        "devito.parameters.configuration['language'] = 'openmp'\n",
        "devito.parameters.configuration['mpi'] = False\n",
        "\n",
        "# By changing the src_index you'll be selecting a different source coordinate from the different 512 that have been created\n",
        "# By changing the model type to water, skull or full you'll get different background mediums\n",
        "# The cutting box has dimensions of (485, 355) in this case. This cannot change\n",
        "# The cutting box is centred in the medium\n",
        "# The complete medium is currently (1001, 1001) but you will have to make this bigger if you need to in order to\n",
        "# make sure that the wavefield does not come back into the cutting box after hitting the edges of the complete medium\n",
        "\n",
        "run_type = 'infinite'  # infinite, limited\n",
        "\n",
        "# Infinite will run the operator like when you generated your dataset. When you plot the resulting wavefield\n",
        "# at the end, 'infinite' will look as if the wave is expanding into infinity. If you run it with 'limited'\n",
        "# however, you'll see how the wave bounces back against the edge. The goal is to make the 'limited' case\n",
        "# run as if it was the 'infinite' one by using your network.\n",
        "\n",
        "# Load different types of background models\n",
        "if model_type == 'water':\n",
        "    model = np.ones((480, 350)) * 1500.\n",
        "\n",
        "elif model_type == 'skull':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-ExactSkull.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "elif model_type == 'full':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-TrueVp.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "y_length = 480+66+4  #480+11\n",
        "x_length = 350+66+4  #350+11\n",
        "\n",
        "# Define a physical size\n",
        "if run_type == 'infinite':\n",
        "    shape = (1001, 1001)            # Number of grid point (nx, nz) - CHANGE THIS FOR THE MEDIUM SIZE\n",
        "else:\n",
        "    shape = (y_length, x_length)\n",
        "    \n",
        "spacing = (0.5e-3, 0.5e-3)      # Grid spacing in m.\n",
        "origin = (0., 0.)               # What is the location of the top left corner. This is necessary to define\n",
        "                                # the absolute location of the source and receivers\n",
        "\n",
        "# Define a velocity profile. The velocity is in m/s\n",
        "diff_x = shape[0] - model.shape[0]\n",
        "diff_y = shape[1] - model.shape[1]\n",
        "offset_x = int(np.floor(diff_x/2))\n",
        "offset_y = int(np.floor(diff_y/2))\n",
        "\n",
        "model = np.pad(model,\n",
        "               ((int(np.floor(diff_x/2)), int(np.ceil(diff_x/2))), \n",
        "                (int(np.floor(diff_y/2)), int(np.ceil(diff_y/2)))),\n",
        "               'edge')\n",
        "\n",
        "plt.imshow(np.transpose(model), origin='lower')\n",
        "\n",
        "# With the velocity and model size defined, we can create the seismic model that\n",
        "# encapsulates this properties. We also define the size of the absorbing layer as 10 grid points\n",
        "model = Model(vp=model, origin=origin, shape=shape, spacing=spacing, space_order=4, nbpml=0)\n",
        "\n",
        "t0 = 0.                 # Simulation starts a t=0\n",
        "dt = 0.08e-6            # Time step from model grid spacing\n",
        "\n",
        "time_range = TimeAxis(start=t0, num=num, step=dt)\n",
        "\n",
        "f0 = 0.40e6  # Source peak frequency is 400kHz\n",
        "src = ToneBurstSource(name='src', grid=model.grid, f0=f0, npoint=1, time_range=time_range)\n",
        "\n",
        "# First, position source centrally in all dimensions\n",
        "src_positions = create_eliptic_array(512, (offset_x, offset_y), spacing[0], (480, 350))\n",
        "src.coordinates.data[0, :] = src_positions[:, src_index]\n",
        "\n",
        "print(src.coordinates.data)\n",
        "\n",
        "plt.scatter(src_positions[0, :]/spacing[0], src_positions[1, :]/spacing[1])\n",
        "plt.show()\n",
        "\n",
        "# We can plot the time signature to see the wavelet\n",
        "src.show()\n",
        "\n",
        "# We can now show the source and receivers within our domain:\n",
        "# Red dot: Source location\n",
        "plot_velocity(model, source=src.coordinates.data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[0.17051011 0.18445057]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wTQq6mKR-FnQ"
      },
      "source": [
        "# In order to represent the wavefield u and the square slowness we need symbolic objects\n",
        "# corresponding to time-space-varying field (u, TimeFunction) and\n",
        "# space-varying field (m, Function)\n",
        "\n",
        "# Define the wavefield with the size of the model and the time dimension\n",
        "u = TimeFunction(name=\"u\", grid=model.grid, time_order=2, space_order=4, save=num)\n",
        "\n",
        "# Create a temporary symbol for H to avoid expensive sympy solve\n",
        "H = Symbol('H')\n",
        "\n",
        "# Define PDE\n",
        "eq = model.m * u.dt2 - H\n",
        "\n",
        "# Solve the symbolic equation for the field to be updated\n",
        "eq_time = solve(eq, u.forward)\n",
        "\n",
        "# Get the spacial FD\n",
        "biharmonic = u.biharmonic(1/model.m)\n",
        "laplacian = u.laplace + dt**2/12 * biharmonic\n",
        "\n",
        "# This discrete PDE can be solved in a time-marching way updating u(t+dt) from the previous time step\n",
        "# Devito as a shortcut for u(t+dt) which is u.forward. We can then rewrite the PDE as\n",
        "# a time marching updating equation known as a stencil using customized SymPy functions\n",
        "stencil = Eq(u.forward, eq_time.subs({H: laplacian}))\n",
        "\n",
        "# Finally we define the source injection function to generate the corresponding code\n",
        "src_term = src.inject(field=u.forward, expr=src * dt**2 / model.m)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oENjb2Zb-FnS"
      },
      "source": [
        "def execute_operator(kwargs, time_min=None, time_max=None):\n",
        "    if time_min is None:\n",
        "        time_min = 1\n",
        "\n",
        "    if time_max is None:\n",
        "        time_max = num-1\n",
        "\n",
        "    kwargs['time_m'] = time_min\n",
        "    kwargs['time_M'] = time_max\n",
        "    \n",
        "    args = op.arguments(**kwargs)\n",
        "\n",
        "    arg_values = [args[p.name] for p in op.parameters]\n",
        "    try:\n",
        "        op.cfunction(*arg_values)\n",
        "\n",
        "    except ctypes.ArgumentError as e:\n",
        "        if e.args[0].startswith(\"argument \"):\n",
        "            argnum = int(e.args[0][9:].split(':')[0]) - 1\n",
        "            newmsg = \"error in argument '%s' with value '%s': %s\" % (\n",
        "                self._operator.parameters[argnum].name,\n",
        "                arg_values[argnum],\n",
        "                e.args[0])\n",
        "            raise ctypes.ArgumentError(newmsg) from e\n",
        "        else:\n",
        "            raise\n",
        "            \n",
        "    op._postprocess_arguments(args, **kwargs)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "705_jdHs-FnV"
      },
      "source": [
        "# Prepare the operator\n",
        "u.data.fill(0.)\n",
        "op = Operator([stencil] + src_term, subs=model.spacing_map)\n",
        "    \n",
        "kwargs = {'dt': dt}\n",
        "\n",
        "# Execute the operator\n",
        "for t in tqdm(range(1, num-2)):\n",
        "    \n",
        "    execute_operator(kwargs, t, t+1)\n",
        "\n",
        "    v = np.copy(u.data[t+1])\n",
        "\n",
        "    # print('=========================================================================================================')\n",
        "    # print('Timestep ',t)\n",
        "    # plt.figure(figsize=(25,20))\n",
        "    # plt.title('Left column images')\n",
        "    # plt.yticks([11])\n",
        "    # plt.imshow(np.rot90(v[275:726,275:325], 3))\n",
        "    # plt.show()\n",
        "\n",
        "    u.data[t+1] = v"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "voa77HYEdVNq",
        "outputId": "e1bf2139-2d81-4cfb-8f45-9313f4e5d2ad"
      },
      "source": [
        "# Save middle\n",
        "print(u.data.shape)\n",
        "original_waves = []\n",
        "inf_slices = (slice(225, -226), slice(290, -291))\n",
        "for t in range(num):\n",
        "    original_waves.append(u.data[t][inf_slices])\n",
        "original_waves = np.array(original_waves)\n",
        "print(original_waves.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2500, 1001, 1001)\n",
            "(2500, 550, 420)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VTrONy0k829f"
      },
      "source": [
        "np.savez_compressed('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/original_waves_{}'.format(src_index), original_waves)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T313LMz863aU"
      },
      "source": [
        "##'Limited' Devito Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DrzScClB63ag"
      },
      "source": [
        "devito.parameters.configuration['language'] = 'openmp'\n",
        "devito.parameters.configuration['mpi'] = False\n",
        "\n",
        "# By changing the src_index you'll be selecting a different source coordinate from the different 512 that have been created\n",
        "# By changing the model type to water, skull or full you'll get different background mediums\n",
        "# The cutting box has dimensions of (485, 355) in this case. This cannot change\n",
        "# The cutting box is centred in the medium\n",
        "# The complete medium is currently (1001, 1001) but you will have to make this bigger if you need to in order to\n",
        "# make sure that the wavefield does not come back into the cutting box after hitting the edges of the complete medium\n",
        "\n",
        "run_type = 'limited'  # infinite, limited\n",
        "\n",
        "# Infinite will run the operator like when you generated your dataset. When you plot the resulting wavefield\n",
        "# at the end, 'infinite' will look as if the wave is expanding into infinity. If you run it with 'limited'\n",
        "# however, you'll see how the wave bounces back against the edge. The goal is to make the 'limited' case\n",
        "# run as if it was the 'infinite' one by using your network.\n",
        "\n",
        "# Load different types of background models\n",
        "if model_type == 'water':\n",
        "    model = np.ones((480, 350)) * 1500.\n",
        "\n",
        "elif model_type == 'skull':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-ExactSkull.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "elif model_type == 'full':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-TrueVp.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "y_length = 480+66+4  #480+11\n",
        "x_length = 350+66+4  #350+11\n",
        "\n",
        "# Define a physical size\n",
        "if run_type == 'infinite':\n",
        "    shape = (1001, 1001)            # Number of grid point (nx, nz) - CHANGE THIS FOR THE MEDIUM SIZE\n",
        "else:\n",
        "    shape = (y_length, x_length)\n",
        "    \n",
        "spacing = (0.5e-3, 0.5e-3)      # Grid spacing in m.\n",
        "origin = (0., 0.)               # What is the location of the top left corner. This is necessary to define\n",
        "                                # the absolute location of the source and receivers\n",
        "\n",
        "# Define a velocity profile. The velocity is in m/s\n",
        "diff_x = shape[0] - model.shape[0]\n",
        "diff_y = shape[1] - model.shape[1]\n",
        "offset_x = int(np.floor(diff_x/2))\n",
        "offset_y = int(np.floor(diff_y/2))\n",
        "\n",
        "model = np.pad(model,\n",
        "               ((int(np.floor(diff_x/2)), int(np.ceil(diff_x/2))), \n",
        "                (int(np.floor(diff_y/2)), int(np.ceil(diff_y/2)))),\n",
        "               'edge')\n",
        "\n",
        "plt.imshow(np.transpose(model), origin='lower')\n",
        "\n",
        "# With the velocity and model size defined, we can create the seismic model that\n",
        "# encapsulates this properties. We also define the size of the absorbing layer as 10 grid points\n",
        "model = Model(vp=model, origin=origin, shape=shape, spacing=spacing, space_order=4, nbpml=0)\n",
        "\n",
        "t0 = 0.                 # Simulation starts a t=0\n",
        "dt = 0.08e-6            # Time step from model grid spacing\n",
        "\n",
        "time_range = TimeAxis(start=t0, num=num, step=dt)\n",
        "\n",
        "f0 = 0.40e6  # Source peak frequency is 400kHz\n",
        "src = ToneBurstSource(name='src', grid=model.grid, f0=f0, npoint=1, time_range=time_range)\n",
        "\n",
        "# First, position source centrally in all dimensions\n",
        "src_positions = create_eliptic_array(512, (offset_x, offset_y), spacing[0], (480, 350))\n",
        "src.coordinates.data[0, :] = src_positions[:, src_index]\n",
        "\n",
        "print(src.coordinates.data)\n",
        "\n",
        "plt.scatter(src_positions[0, :]/spacing[0], src_positions[1, :]/spacing[1])\n",
        "plt.show()\n",
        "\n",
        "# We can plot the time signature to see the wavelet\n",
        "src.show()\n",
        "\n",
        "# We can now show the source and receivers within our domain:\n",
        "# Red dot: Source location\n",
        "plot_velocity(model, source=src.coordinates.data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BVU_-ayu63ah"
      },
      "source": [
        "# In order to represent the wavefield u and the square slowness we need symbolic objects\n",
        "# corresponding to time-space-varying field (u, TimeFunction) and\n",
        "# space-varying field (m, Function)\n",
        "\n",
        "# Define the wavefield with the size of the model and the time dimension\n",
        "u = TimeFunction(name=\"u\", grid=model.grid, time_order=2, space_order=4, save=num)\n",
        "\n",
        "# Create a temporary symbol for H to avoid expensive sympy solve\n",
        "H = Symbol('H')\n",
        "\n",
        "# Define PDE\n",
        "eq = model.m * u.dt2 - H\n",
        "\n",
        "# Solve the symbolic equation for the field to be updated\n",
        "eq_time = solve(eq, u.forward)\n",
        "\n",
        "# Get the spacial FD\n",
        "biharmonic = u.biharmonic(1/model.m)\n",
        "laplacian = u.laplace + dt**2/12 * biharmonic\n",
        "\n",
        "# This discrete PDE can be solved in a time-marching way updating u(t+dt) from the previous time step\n",
        "# Devito as a shortcut for u(t+dt) which is u.forward. We can then rewrite the PDE as\n",
        "# a time marching updating equation known as a stencil using customized SymPy functions\n",
        "stencil = Eq(u.forward, eq_time.subs({H: laplacian}))\n",
        "\n",
        "# Finally we define the source injection function to generate the corresponding code\n",
        "src_term = src.inject(field=u.forward, expr=src * dt**2 / model.m)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dRFa-5fE63ah"
      },
      "source": [
        "def execute_operator(kwargs, time_min=None, time_max=None):\n",
        "    if time_min is None:\n",
        "        time_min = 1\n",
        "\n",
        "    if time_max is None:\n",
        "        time_max = num-1\n",
        "\n",
        "    kwargs['time_m'] = time_min\n",
        "    kwargs['time_M'] = time_max\n",
        "    \n",
        "    args = op.arguments(**kwargs)\n",
        "\n",
        "    arg_values = [args[p.name] for p in op.parameters]\n",
        "    try:\n",
        "        op.cfunction(*arg_values)\n",
        "\n",
        "    except ctypes.ArgumentError as e:\n",
        "        if e.args[0].startswith(\"argument \"):\n",
        "            argnum = int(e.args[0][9:].split(':')[0]) - 1\n",
        "            newmsg = \"error in argument '%s' with value '%s': %s\" % (\n",
        "                self._operator.parameters[argnum].name,\n",
        "                arg_values[argnum],\n",
        "                e.args[0])\n",
        "            raise ctypes.ArgumentError(newmsg) from e\n",
        "        else:\n",
        "            raise\n",
        "            \n",
        "    op._postprocess_arguments(args, **kwargs)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XjGdbORZ63ah"
      },
      "source": [
        "# Prepare the operator\n",
        "u.data.fill(0.)\n",
        "op = Operator([stencil] + src_term, subs=model.spacing_map)\n",
        "    \n",
        "kwargs = {'dt': dt}\n",
        "\n",
        "# Execute the operator\n",
        "for t in tqdm(range(1, num-2)):\n",
        "    execute_operator(kwargs, t, t+1)\n",
        "\n",
        "    v = np.copy(u.data[t+1])\n",
        "\n",
        "    # print('=========================================================================================================')\n",
        "    # print('Timestep ',t)\n",
        "    # plt.figure(figsize=(25,20))\n",
        "    # plt.title('Left column images')\n",
        "    # plt.yticks([11])\n",
        "    # plt.imshow(np.rot90(v[275:726,275:325], 3))\n",
        "    # plt.show()\n",
        "\n",
        "    u.data[t+1] = v\n",
        "\n",
        "limited_waves = copy.deepcopy(u.data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cs2ZhJGU889b"
      },
      "source": [
        "np.savez_compressed('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/limited_waves', a=u.data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GrHMUFCj63ai"
      },
      "source": [
        "# # Plot the results, every 20 timesteps for now\n",
        "# for t in range(0, num, 100):    \n",
        "#     if run_type == 'infinite':\n",
        "#         slices = (slice(255, -255), slice(320, -320))\n",
        "#     else:\n",
        "#         slices = (slice(0, None), slice(0, None))\n",
        "    \n",
        "#     plt.imshow(np.transpose(u.data[t][slices]))\n",
        "#     plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uwtfkspc7V7j"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SwrIebaa7XB2"
      },
      "source": [
        "##Compare solutions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bgqnnV7D9EZu"
      },
      "source": [
        "src_index = 315\n",
        "\n",
        "original_waves = np.load('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/original_waves_{}.npz'.format(src_index))['arr_0']\n",
        "# limited_waves = np.load('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/limited_waves.npz')['a']\n",
        "modelled_waves = np.load('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/modelled_waves_3x32_{}.npz'.format(src_index))['arr_0']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qd81PVvqtuP2"
      },
      "source": [
        "# inf_slices = (slice(255-35, -255-35), slice(320-35, -320-35))\n",
        "inf_slices = (slice(225, -226), slice(290, -291))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_xhl8pWUtbEx",
        "outputId": "5855a88e-e9fb-4c49-9ca0-4f0cc5bdf34b"
      },
      "source": [
        "percentage_error = np.mean((original_waves-modelled_waves)*100/np.max(original_waves))\n",
        "\n",
        "print('Average % error: {}'.format(percentage_error))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Average % error: -0.01338666770607233\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xt-z0aNSTMdy"
      },
      "source": [
        "# Load different types of background models\n",
        "model_type = 'skull'\n",
        "\n",
        "if model_type == 'skull':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-ExactSkull.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "elif model_type == 'full':\n",
        "    model = np.fromfile('/content/drive/My Drive/ML_Ultrasound_Project/UROPdata/data/alpha2D-TrueVp.raw', dtype=np.float32, count=-1)\n",
        "    model = model.reshape([350, 480])[::1, ::1].transpose()\n",
        "\n",
        "y_length = 480+66+4  #480+11\n",
        "x_length = 350+66+4  #350+11\n",
        "\n",
        "# Define a physical size\n",
        "shape = (1001, 1001)            # Number of grid point (nx, nz) - CHANGE THIS FOR THE MEDIUM SIZE\n",
        "\n",
        "    \n",
        "spacing = (0.5e-3, 0.5e-3)      # Grid spacing in m.\n",
        "origin = (0., 0.)               # What is the location of the top left corner. This is necessary to define\n",
        "                                # the absolute location of the source and receivers\n",
        "\n",
        "# Define a velocity profile. The velocity is in m/s\n",
        "diff_x = shape[0] - model.shape[0]\n",
        "diff_y = shape[1] - model.shape[1]\n",
        "offset_x = int(np.floor(diff_x/2))\n",
        "offset_y = int(np.floor(diff_y/2))\n",
        "\n",
        "full_model = np.pad(model,\n",
        "               ((int(np.floor(diff_x/2)), int(np.ceil(diff_x/2))), \n",
        "                (int(np.floor(diff_y/2)), int(np.ceil(diff_y/2)))),\n",
        "               'edge')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lb-fVESqT64i"
      },
      "source": [
        "model = np.transpose(full_model[inf_slices])\n",
        "# model[np.where(model == 1500)] = 0\n",
        "\n",
        "# for src = 315, source coordinates are [78,116] -> flip to [116,78]\n",
        "source= [116,78]\n",
        "# for src = 128, sc = [515,211]\n",
        "# source= [516,210]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uL_nbyFOkfcw"
      },
      "source": [
        "def nrmse_loss(y_pred, y_true):\n",
        "    diff = np.divide(y_pred - y_true, np.max(y_true) - np.min(y_true) + 1e-6)\n",
        "    return np.sqrt(np.mean(np.square(diff)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xbPSwTZ_NLZD"
      },
      "source": [
        "**Plot and compare model vs infinite**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4QXedHkw7fA_"
      },
      "source": [
        "skull_vmin = np.min(model)\n",
        "skull_vmax = np.max(model)\n",
        "\n",
        "skull_norm = ModMidpointNormalize(vmin=skull_vmin, vmax=skull_vmax, midpoint=1500)\n",
        "\n",
        "cmap = 'seismic'\n",
        "skull_cmap = 'RdGy'\n",
        "err_cmap = 'Greys'\n",
        "\n",
        "print('total nrmse: {}'.format(nrmse_loss(modelled_waves, original_waves)))\n",
        "\n",
        "tp=0\n",
        "\n",
        "fig, ax = plt.subplots(3,6, figsize=(30,30))\n",
        "\n",
        "for t in range(200, 2500, 400):\n",
        "\n",
        "  inf_img = np.transpose(original_waves[t])/np.max(np.abs(original_waves[t]))\n",
        "  mod_img = np.transpose(modelled_waves[t])/np.max(np.abs(modelled_waves[t]))\n",
        "  # inf_img = inf_img[11:-11, 11:-11]\n",
        "  # mod_img = mod_img[11:-11, 11:-11]\n",
        "  err_img = np.abs(np.divide((inf_img-mod_img)*100, np.max(np.abs(inf_img))))\n",
        "\n",
        "  print('t: {}, nrmse: {}, max % error: {}'.format(t, nrmse_loss(mod_img, inf_img), np.max(err_img)))\n",
        "\n",
        "  divider = make_axes_locatable(ax[0,tp])\n",
        "  cax1 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "  ax[0,tp].set_title('Timestep {}'.format(t))\n",
        "  ax[0,tp].set_yticks([])\n",
        "  ax[0,tp].set_xticks([])\n",
        "  im1 = ax[0,tp].imshow(inf_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "  ax[0,tp].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  ax[0,tp].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  cbar1 = fig.colorbar(im1, cax=cax1)\n",
        "  cbar1.ax.set_ylabel('Pressure (AU)')\n",
        "\n",
        "  divider = make_axes_locatable(ax[1,tp])\n",
        "  cax2 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "  ax[1,tp].set_yticks([])\n",
        "  ax[1,tp].set_xticks([])\n",
        "  im2 = ax[1,tp].imshow(mod_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "  ax[1,tp].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  ax[1,tp].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  cbar2 = fig.colorbar(im2, cax=cax2)\n",
        "  cbar2.ax.set_ylabel('Pressure (AU)')\n",
        "\n",
        "  divider = make_axes_locatable(ax[2,tp])\n",
        "  cax3 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "  ax[2,tp].set_yticks([])\n",
        "  ax[2,tp].set_xticks([])\n",
        "  im3 = ax[2,tp].imshow(err_img, cmap=err_cmap, vmin=0, vmax=6)\n",
        "  ax[2,tp].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  ax[2,tp].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  cbar3 = fig.colorbar(im3, cax=cax3)\n",
        "  cbar3.ax.set_ylabel('Percentage Error (%)')\n",
        "\n",
        "  tp +=1\n",
        "\n",
        "y = ax[0,0].set_ylabel('Infinite', fontsize=20, rotation=0, va=\"center\")\n",
        "ax[0,0].yaxis.set_label_coords(-0.3,0.5)\n",
        "\n",
        "y = ax[1,0].set_ylabel('Network\\nPredictions', fontsize=20, rotation=0, va=\"center\")\n",
        "ax[1,0].yaxis.set_label_coords(-0.3,0.5)\n",
        "\n",
        "y = ax[2,0].set_ylabel('Errors', fontsize=20, rotation=0, va=\"center\")\n",
        "ax[2,0].yaxis.set_label_coords(-0.3,0.5)\n",
        "\n",
        "fig.subplots_adjust(bottom=0.1, top=0.4, left=0.1, right=0.8,\n",
        "                    wspace=0, hspace=0.1)\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_elGBNenf7OR"
      },
      "source": [
        "# source= [116,78]\n",
        "\n",
        "# skull_vmin = np.min(model)\n",
        "# skull_vmax = np.max(model)\n",
        "\n",
        "# skull_norm = ModMidpointNormalize(vmin=skull_vmin, vmax=skull_vmax, midpoint=1500)\n",
        "\n",
        "# cmap = 'seismic'\n",
        "# skull_cmap = 'RdGy'\n",
        "# err_cmap = 'Greys'\n",
        "\n",
        "# for t in range(1, num-2, 50):\n",
        "\n",
        "#     inf_img = np.transpose(original_waves[t])/np.max(np.abs(original_waves[t]))\n",
        "#     lim_img = np.transpose(limited_waves[t])/np.max(np.abs(limited_waves[t]))\n",
        "#     mod_img = np.transpose(modelled_waves[t])/np.max(np.abs(modelled_waves[t]))\n",
        "#     err_img = np.abs(np.divide((inf_img-mod_img)*100, np.max(np.abs(inf_img))))\n",
        "\n",
        "#     print('============================================================================================================================================================================================================')\n",
        "#     print('Timestep ',t)\n",
        "\n",
        "#     # print(np.where(inf_img != 0))\n",
        "\n",
        "#     fig, ax = plt.subplots(1,2, figsize=(18,9))\n",
        "#     # fig, ax = plt.subplots(1,4, figsize=(30,20))\n",
        "\n",
        "#     divider = make_axes_locatable(ax[0])\n",
        "#     cax0 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "#     ax[0].set_title('Limited')\n",
        "#     ax[0].set_yticks([])\n",
        "#     ax[0].set_xticks([])\n",
        "#     im0 = ax[0].imshow(lim_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "#     ax[0].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "#     ax[0].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "#     cbar0 = fig.colorbar(im0, cax=cax0)\n",
        "#     cbar0.ax.set_ylabel('Pressure (AU)')\n",
        "\n",
        "#     divider = make_axes_locatable(ax[1])\n",
        "#     cax1 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "#     ax[1].set_title('Infinite')\n",
        "#     ax[1].set_yticks([])\n",
        "#     ax[1].set_xticks([])\n",
        "#     im1 = ax[1].imshow(inf_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "#     ax[1].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "#     ax[1].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "#     cbar1 = fig.colorbar(im1, cax=cax1)\n",
        "#     cbar1.ax.set_ylabel('Pressure (AU)')\n",
        "\n",
        "# #     divider = make_axes_locatable(ax[2])\n",
        "# #     cax2 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "# #     ax[2].set_title('Network Predictions')\n",
        "# #     ax[2].set_yticks([])\n",
        "# #     ax[2].set_xticks([])\n",
        "# #     im2 = ax[2].imshow(mod_img, cmap=cmap, norm=mod_norm)\n",
        "# #     ax[2].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "# #     ax[2].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "# #     cbar2 = fig.colorbar(im2, cax=cax2)\n",
        "# #     cbar2.ax.set_ylabel('Pressure (AU)')\n",
        "\n",
        "# #     divider = make_axes_locatable(ax[3])\n",
        "# #     cax3 = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
        "# #     ax[3].set_title('Errors: max error = {:.2f}%'.format(err_vmax))\n",
        "# #     ax[3].set_yticks([])\n",
        "# #     ax[3].set_xticks([])\n",
        "# #     im3 = ax[3].imshow(err_img, cmap=err_cmap, vmin=0, vmax=err_vmax)\n",
        "# #     ax[3].scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "# #     ax[3].imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "# #     cbar3 = fig.colorbar(im3, cax=cax3)\n",
        "# #     cbar3.ax.set_ylabel('Percentage Error (%)')\n",
        "\n",
        "#     plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k-XFtQtv4ilq"
      },
      "source": [
        "**Save gifs**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GGCdGq-1Iayq"
      },
      "source": [
        "# for src = 315, source coordinates are [78,116] -> flip to [116,78]\n",
        "source= [116,78]\n",
        "# for src = 128, sc = [515,211]\n",
        "# source= [516,210]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 137,
          "referenced_widgets": [
            "05a3ff11cd5349d38ec5b828530e69fa",
            "8f711c9d3d2e4dd8ba6c09bfee43efa6",
            "b06c2ce62df34933acb83a46e84e9806",
            "4b43f53dc12e4d4a9f329eb1a0268ebd",
            "765e40fceaaa4e0fb7b1a89d3749587e",
            "276784a609974a55b1756b0999626781",
            "a428dac3c756458598c3079c06160cae",
            "6e0c9890ad9748859a7720899455f04e"
          ]
        },
        "id": "ofnUXRzA7G5c",
        "outputId": "06b63e16-b4b3-4966-fe9e-82e97d65785c"
      },
      "source": [
        "skull_vmin = np.min(model)\n",
        "skull_vmax = np.max(model)\n",
        "\n",
        "skull_norm = ModMidpointNormalize(vmin=skull_vmin, vmax=skull_vmax, midpoint=1500)\n",
        "\n",
        "cmap = 'seismic'\n",
        "skull_cmap = 'RdGy'\n",
        "err_cmap = 'Greys'\n",
        "\n",
        "for t in tqdm(range(0, len(original_waves), 10)):\n",
        "\n",
        "  inf_img = np.transpose(original_waves[t])/np.max(np.abs(original_waves[t]))\n",
        "  mod_img = np.transpose(modelled_waves[t])/np.max(np.abs(modelled_waves[t]))\n",
        "  # inf_img = inf_img[11:-11, 11:-11]\n",
        "  # mod_img = mod_img[11:-11, 11:-11]\n",
        "  err_img = np.abs(np.divide((inf_img-mod_img)*100, np.max(np.abs(inf_img))))\n",
        "\n",
        "  plt.figure()\n",
        "  plt.yticks([])\n",
        "  plt.xticks([])\n",
        "  im1 = plt.imshow(inf_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "  plt.scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  plt.imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  # plt.show()\n",
        "  plt.savefig(str(t)+'_og_frames_src'+str(src_index)+'.png')\n",
        "  plt.close()\n",
        "\n",
        "  plt.figure()\n",
        "  plt.yticks([])\n",
        "  plt.xticks([])\n",
        "  im1 = plt.imshow(mod_img, cmap=cmap, vmin=-1, vmax=1)\n",
        "  plt.scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  plt.imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  # plt.show()\n",
        "  plt.savefig(str(t)+'_mod_frames_src'+str(src_index)+'.png')\n",
        "  plt.close()\n",
        "\n",
        "  plt.figure()\n",
        "  plt.yticks([])\n",
        "  plt.xticks([])\n",
        "  im1 = plt.imshow(err_img, cmap=err_cmap, vmin=0, vmax=6)\n",
        "  plt.scatter(source[0], source[1], s=25, c='red', marker='o')\n",
        "  plt.imshow(model, alpha=.2, cmap=skull_cmap, norm=skull_norm)\n",
        "  # plt.show()\n",
        "  plt.savefig(str(t)+'_err_frames_src'+str(src_index)+'.png')\n",
        "  plt.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "05a3ff11cd5349d38ec5b828530e69fa",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=250.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:12: RuntimeWarning: invalid value encountered in true_divide\n",
            "  if sys.path[0] == '':\n",
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:13: RuntimeWarning: invalid value encountered in true_divide\n",
            "  del sys.path[0]\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V-t10AMZ-IsS"
      },
      "source": [
        "from PIL import Image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3CytVPA6AXKx"
      },
      "source": [
        "og_fp_in = '_og_frames_src'+str(src_index)+'.png'\n",
        "og_fp_out = 'og_'+str(src_index)+'.gif'\n",
        "\n",
        "mod_fp_in = '_mod_frames_src'+str(src_index)+'.png'\n",
        "mod_fp_out = 'mod_'+str(src_index)+'.gif'\n",
        "\n",
        "err_fp_in = '_err_frames_src'+str(src_index)+'.png'\n",
        "err_fp_out = 'err_'+str(src_index)+'.gif'\n",
        "\n",
        "og_img, *og_imgs = [Image.open(str(f)+og_fp_in) for f in range(0,1500,10)]\n",
        "og_img.save(fp=og_fp_out, format='GIF', append_images=og_imgs, save_all=True, duration=60, loop=0)\n",
        "\n",
        "mod_img, *mod_imgs = [Image.open(str(f)+mod_fp_in) for f in range(0,1500,10)]\n",
        "mod_img.save(fp=mod_fp_out, format='GIF', append_images=mod_imgs, save_all=True, duration=60, loop=0)\n",
        "\n",
        "err_img, *err_imgs = [Image.open(str(f)+err_fp_in) for f in range(0,1500,10)]\n",
        "err_img.save(fp=err_fp_out, format='GIF', append_images=err_imgs, save_all=True, duration=60, loop=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ER3KMywuM03k"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}