playground/general_compression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# General compression using INIF "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ERROR:absl:Oryx not found! This library will still work but no summarywill be logged.\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "import os\n",
    "sys.path.append('..')\n",
    "import jax\n",
    "import jax.numpy as jnp\n",
    "import optax\n",
    "import ipywidgets as widgets\n",
    "from tqdm import tqdm\n",
    "from IPython.display import clear_output\n",
    "from utils import others, visualizer\n",
    "from models import INR, HEVC"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Read the raw data, set ROI and visualize the ROI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No files found in the directory.\n"
     ]
    }
   ],
   "source": [
    "# set the root directory\n",
    "root_directory = '/INIF/data'\n",
    "# set the default path, set to None if you want the first file in the directory\n",
    "defualt_path = '/INIF/data/TNNI1/TNNI1.tif'\n",
    "\n",
    "# fetch all the files in the root directory\n",
    "file_paths = []\n",
    "for subdir, dirs, files in os.walk(root_directory):\n",
    "    for file in files:\n",
    "        file_paths.append(os.path.join(subdir, file))\n",
    "\n",
    "# check if there are files in the directory\n",
    "if not file_paths:\n",
    "    print(\"No files found in the directory.\")\n",
    "else:\n",
    "    # create a dropdown menu\n",
    "    file_selector = widgets.Dropdown(\n",
    "        options=file_paths,\n",
    "        description='File path:',\n",
    "        value = defualt_path if defualt_path else file_paths[0],\n",
    "        disabled=False,\n",
    "        continuous_update=False\n",
    "    )\n",
    "    # display the dropdown menu\n",
    "    display(file_selector)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'file_selector' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[3], line 2\u001b[0m\n\u001b[1;32m      1\u001b[0m meta_data \u001b[38;5;241m=\u001b[39m {}\n\u001b[0;32m----> 2\u001b[0m raw_data \u001b[38;5;241m=\u001b[39m others\u001b[38;5;241m.\u001b[39mread_data(\u001b[43mfile_selector\u001b[49m\u001b[38;5;241m.\u001b[39mvalue, meta_data)\n",
      "\u001b[0;31mNameError\u001b[0m: name 'file_selector' is not defined"
     ]
    }
   ],
   "source": [
    "meta_data = {}\n",
    "raw_data = others.read_data(file_selector.value, meta_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of the ROI data is: {'dim0': 60, 'X': 441, 'Y': 555}\n",
      "The size of the ROI data is: 33.290459632873535 MB\n",
      "The data type of the raw data is: uint16\n",
      "The max value of the raw data is: 992, and the min value is: 34\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3b778c8a40e841169819930d75beb05c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntSlider(value=30, continuous_update=False, description='dim0', max=59),))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7f831c2b193a41019c09f3ad7950c6c8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# use slicing method to get the data you want\n",
    "ROI_data = raw_data # ex: raw_data[0:100, 0:100, 0:100]\n",
    "ROI_data = others.preprocess(ROI_data, meta_data)\n",
    "visual_tracker = visualizer.DataVisualizer(ROI_data, meta_data['shape'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Setup compression configs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e9f63d299e304617a86145fd0ae3faee",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "IntSlider(value=256, continuous_update=False, description='Compression Ratio:', max=2048, min=1, step=16, styl…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "20efaa72209f4f468326df14522c9aaa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "IntSlider(value=40000, continuous_update=False, description='Number of Iterations:', max=50000, min=500, step=…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0dba0905b5654714a31a897fb69075cc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "IntSlider(value=2000000, continuous_update=False, description='Batch Size:', max=14685300, min=1, style=Slider…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "217663bd9abf43dba2c95903b215be24",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Dropdown(description='Model Type:', index=2, options=('INIF', 'SIREN', 'HEVC'), value='HEVC')"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "27bab1fb1ba54b4dafe14ed135b8322c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Button(description='Apply', style=ButtonStyle())"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f599156756614adea62bd0340bfb2459",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ask for the compression ratio, the number of iterations, and the model type\n",
    "ratio_slider = widgets.IntSlider(\n",
    "    value=256,  # defualt value\n",
    "    min=1,      # minimum value\n",
    "    max=2048,   # maximum value    \n",
    "    step=16,     # step value\n",
    "    description=f'Compression Ratio:',\n",
    "    continuous_update=False,\n",
    "    style={'description_width': 'initial'},\n",
    "    \n",
    ")\n",
    "iteration_slider = widgets.IntSlider(\n",
    "    value=40000,  # defualt value\n",
    "    min=500,     # minimum value\n",
    "    max=50000,   # maximum value\n",
    "    step=500,    # step value\n",
    "    description='Number of Iterations:',\n",
    "    continuous_update=False,\n",
    "    style={'description_width': 'initial'}\n",
    ")\n",
    "batch_size_slider = widgets.IntSlider(\n",
    "    value = min(ROI_data.size, 2000000), # default value\n",
    "    min = 1, # minimum value\n",
    "    max = ROI_data.size, # maximum value\n",
    "    step = 1, # step value\n",
    "    description = 'Batch Size:',\n",
    "    continuous_update = False,\n",
    "    style = {'description_width': 'initial'}\n",
    ")\n",
    "model_selector = widgets.Dropdown(\n",
    "    options=['INIF', 'SIREN', 'HEVC'],\n",
    "    value='INIF', # default value\n",
    "    description='Model Type:',\n",
    "    disabled=False,\n",
    "    continuous_update=False\n",
    ")\n",
    "apply_button = widgets.Button(description='Apply')\n",
    "output = widgets.Output()\n",
    "def on_apply_button_clicked(b):\n",
    "    # apply the configuration\n",
    "    with output:\n",
    "        output.clear_output()\n",
    "        current_ratio = ratio_slider.value\n",
    "        current_iteration_number = iteration_slider.value\n",
    "        current_batch_size = batch_size_slider.value\n",
    "        current_model = model_selector.value\n",
    "        print(f\"Current compression ratio: {current_ratio}x\")\n",
    "        print(f\"Current number of iterations: {current_iteration_number}\")\n",
    "        print(f\"Current batch size: {current_batch_size} = {current_batch_size/ROI_data.size*100 :4f}% of total {ROI_data.size} points\")\n",
    "        print(f\"Current model type: {current_model}\")\n",
    "apply_button.on_click(on_apply_button_clicked)\n",
    "display(ratio_slider, iteration_slider, batch_size_slider, model_selector, apply_button, output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "meta_data['ratio'] = ratio_slider.value\n",
    "meta_data['iteration_number'] = iteration_slider.value\n",
    "meta_data['batch_size'] = batch_size_slider.value\n",
    "meta_data['model'] = {}\n",
    "meta_data['model']['name']= model_selector.value\n",
    "meta_data['visualize_idx'] = visual_tracker.get_indices()\n",
    "visualizer.display_metadata(meta_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Initialize the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "if meta_data['model']['name'] == 'HEVC':\n",
    "    ROI_data = others.denorm(ROI_data, meta_data)\n",
    "    ROI_data = ROI_data.astype(meta_data['dtype'])\n",
    "else:\n",
    "    sampler = others.Sampler(ROI_data, meta_data)\n",
    "    model, params, opt, opt_state = INR.get_model_param_opt_optstate(meta_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Define the training loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loss(params, key, model, x, y, aux_loss):\n",
    "    pred = model.apply(params, key, x)\n",
    "    mse = jnp.mean((pred - y)**2)\n",
    "    total_loss = mse + aux_loss\n",
    "    return total_loss, (mse, aux_loss)\n",
    "\n",
    "def update(params, opt, opt_state, model, x, y, aux_loss = 0):\n",
    "    (loss_val, (mse, aux_loss)), grads = jax.value_and_grad(loss, has_aux=True)(params, None, model, x, y, aux_loss)\n",
    "    updates, new_opt_state = opt.update(grads, opt_state, params, extra_args={'loss': loss_val})\n",
    "    new_params = optax.apply_updates(params, updates)\n",
    "    return loss_val, new_params, new_opt_state, mse, aux_loss\n",
    "\n",
    "def auto_pre_check_INR(meta_data, model, params, opt, opt_state, sampler):        \n",
    "    while True:\n",
    "        try:\n",
    "            # Attempt to train with the current batch size\n",
    "            print(f'Trying batch size: {sampler.batch_size}')\n",
    "            x, y = sampler.next()\n",
    "            loss_val, params, opt_state, mse, aux_loss = update(params, opt, opt_state, model, x, y)\n",
    "            break  # If successful, break out of the loop\n",
    "        except:\n",
    "            # If an exception occurs, reduce the batch size\n",
    "            print(f\"Failed. Reducing batch size automatically...\")\n",
    "            if sampler.batch_size < 1:\n",
    "                raise ValueError(\"Batch size cannot be reduced below 1.\")\n",
    "            sampler.batch_size = round(sampler.batch_size*0.9) # Set new batch size in the sampler\n",
    "\n",
    "    meta_data['batch_size'] = sampler.batch_size\n",
    "    meta_data['loop_number'] = meta_data[\"iteration_number\"]*sampler.batch_size/ROI_data.size\n",
    "    print(f'Batch size: {sampler.batch_size}, Iteration number: {meta_data[\"iteration_number\"]}')\n",
    "    print(f'We approximately loop trough the data {meta_data[\"loop_number\"] : 4f} times')\n",
    "    print('########## Metadata ##########')\n",
    "    visualizer.display_metadata(meta_data)\n",
    "        \n",
    "def auto_pre_check_HEVC(meta_data, data, crf_min=0, crf_max=51, tolerance=1, max_iterations=10):\n",
    "    target_ratio = meta_data['ratio']\n",
    "    iteration = 0\n",
    "    while crf_min < crf_max and iteration < max_iterations:\n",
    "        iteration += 1\n",
    "        crf_mid = (crf_min + crf_max) / 2  # current CRF\n",
    "        \n",
    "        # test the current CRF\n",
    "        meta_data['crf'] = crf_mid\n",
    "        actual_ratio = HEVC.compress(data, meta_data)\n",
    "        diff = abs(actual_ratio - target_ratio)\n",
    "        \n",
    "        if diff <= tolerance:  # if the difference is within the tolerance, return the CRF\n",
    "            exit(f'CRF: {crf_mid}, Actual Ratio: {actual_ratio}, Target Ratio: {target_ratio}')\n",
    "        elif actual_ratio > target_ratio:  # if the actual ratio is less than the target ratio, decrease CRF\n",
    "            crf_max = crf_mid\n",
    "        else:  # if the actual ratio is greater than the target ratio, increase CRF\n",
    "            crf_min = crf_mid\n",
    "\n",
    "    # if the loop ends without finding the CRF, return the average of the min and max CRF\n",
    "    meta_data['crf'] = (crf_min + crf_max) / 2\n",
    "    print('########## Metadata ##########')\n",
    "    visualizer.display_metadata(meta_data)\n",
    "    \n",
    "        \n",
    " \n",
    "def train(meta_data, model, params, opt, opt_state, sampler):\n",
    "    print('########## Compressing ##########')\n",
    "    iteration_number = meta_data['iteration_number']\n",
    "    progress_bar = tqdm(range(iteration_number))\n",
    "    for step in progress_bar:\n",
    "        x, y = sampler.next()\n",
    "        loss_val, params, opt_state, mse, aux_loss = update( params, opt, opt_state, model, x, y)\n",
    "        progress_bar.set_description(f'Loss: {loss_val:.6f},  MSE: {mse:.6f}, Aux Loss: {aux_loss:.6f}')\n",
    "        if step % 100 == 0:\n",
    "            # clean the previous plots\n",
    "            slice_coords, image_shape = visualizer.get_slice( meta_data, meta_data['visualize_idx'], downsample_ratio=2)\n",
    "            slice_values = ROI_data[meta_data['visualize_idx']] \n",
    "            decode_image = model.apply(params, None, slice_coords).reshape(image_shape)\n",
    "            clear_output(wait=True)\n",
    "            visualizer.plot2d(slice_values)\n",
    "            visualizer.plot2d(decode_image)\n",
    "    return meta_data, model, params\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Final check before compression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:25.5, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.0061969757080078125MB\n",
      "target ratio : 256\n",
      "actual ratio : 5372.049399815328\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:12.75, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.01897907257080078MB\n",
      "target ratio : 256\n",
      "actual ratio : 1754.0614541982816\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:6.375, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029384613037109375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1132.921491626639\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:3.1875, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029557228088378906MB\n",
      "target ratio : 256\n",
      "actual ratio : 1126.3051979479237\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:1.59375, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029720306396484375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1120.1250481324605\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.796875, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029631614685058594MB\n",
      "target ratio : 256\n",
      "actual ratio : 1123.4777445206141\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.3984375, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029631614685058594MB\n",
      "target ratio : 256\n",
      "actual ratio : 1123.4777445206141\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.19921875, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029567718505859375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1125.9055928267321\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.099609375, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029567718505859375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1125.9055928267321\n",
      "Remove ../save/test.mp4\n",
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.0498046875, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029567718505859375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1125.9055928267321\n",
      "Remove ../save/test.mp4\n",
      "########## Metadata ##########\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if meta_data['model']['name'] == 'HEVC':\n",
    "    auto_pre_check_HEVC(meta_data, ROI_data)\n",
    "else:\n",
    "    auto_pre_check_INR(meta_data, model, params, opt, opt_state, sampler)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Training the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Padded image shape (60, 442, 556, 1)\n",
      "Compressed Done!\n",
      "HEVC CRF:0.02490234375, compressed bpp:547.6526822041429\n",
      "Original_MB:33.290459632873535MB\n",
      "compressed_MB:0.029567718505859375MB\n",
      "target ratio : 256\n",
      "actual ratio : 1125.9055928267321\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1465a5e9ede14b0c8d33fb44511b3ffa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntSlider(value=30, continuous_update=False, description='dim0', max=59),))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "08ffa209304e4733bc3eec40d2a0ac03",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if meta_data['model']['name'] == 'HEVC':\n",
    "    save_name = f'../save/HEVC_name_{meta_data[\"name\"]}_ratio_{meta_data[\"ratio\"]}_crf_{meta_data[\"crf\"]}.mp4'\n",
    "    HEVC.compress(ROI_data, meta_data, save_name, test_only=False)\n",
    "    decompressed_data = HEVC.decompress(save_name,meta_data)\n",
    "    visualizer.DataVisualizer(decompressed_data, meta_data['shape'])\n",
    "else:\n",
    "    meta, model, params = train(meta_data, model, params, opt, opt_state, sampler)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Save file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Moving the compressed video to ../save/HEVC_name_TNNI1_ratio_256_crf_0.02490234375\n",
      "saved at ../save/HEVC_name_TNNI1_ratio_256_crf_0.02490234375/meta_data.pkl\n"
     ]
    }
   ],
   "source": [
    "if meta_data['model']['name'] != 'HEVC':\n",
    "    others.save(meta_data, params)\n",
    "else:\n",
    "    # we move the mp4 and the metadata to the save folder\n",
    "    folder_path = save_name.split('.mp4')[0]\n",
    "    os.makedirs(folder_path, exist_ok=True)\n",
    "    os.rename(save_name, f'{folder_path}/compressed.mp4')\n",
    "    others.save(meta_data, None, folder_path)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "silen",
   "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": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}