model_performance_vs_existing_segs.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Performance\n",
    "When the AIMI 2024 project was initiated there were no existing segmentations in the IDC PROSTATE-MRI-US-BIOPSY collection. Since the, other segmentations have been added. We can compare our model performance to these segmentations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "%pip install numpy pandas pydicom pydicom-seg scikit-learn seaborn requests SimpleITK matplotlib google-cloud-bigquery db-dtypes tqdm\n",
    "%pip install segmentationmetrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import required python packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "import io\n",
    "import os\n",
    "import subprocess\n",
    "import zipfile\n",
    "from pathlib import Path\n",
    "from typing import Union\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import pydicom\n",
    "import pydicom_seg\n",
    "import requests\n",
    "import scipy.stats as stats\n",
    "import seaborn as sns\n",
    "import segmentationmetrics as sm\n",
    "import SimpleITK as sitk\n",
    "from google.cloud import bigquery\n",
    "from segmentationmetrics.surface_distance import compute_surface_dice_at_tolerance\n",
    "from tqdm.auto import tqdm\n",
    "\n",
    "sns.set_theme(style=\"whitegrid\")\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt_colors = {\n",
    "    \"ne2\": \"#5eceb0\",\n",
    "    \"rad1\": \"#9e70e1\",\n",
    "    \"tp\": \"k\",\n",
    "    \"fp\": \"b\",\n",
    "    \"fn\": \"r\",\n",
    "}\n",
    "\n",
    "revewer_cmap = sns.color_palette([plt_colors[\"ne2\"], plt_colors[\"rad1\"]])\n",
    "vol_cmap = sns.color_palette([plt_colors[\"fp\"], plt_colors[\"fn\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_dcm_seg(seg_path: Path, label: Union[str, int] = 1):\n",
    "    dcm = pydicom.dcmread(str(seg_path))\n",
    "    reader = pydicom_seg.SegmentReader()\n",
    "    result = reader.read(dcm)\n",
    "    if isinstance(label, str):\n",
    "        for segment_number in result.available_segments:\n",
    "            if result.segment_infos[segment_number].SegmentLabel == label:\n",
    "                label = segment_number\n",
    "                break\n",
    "        assert isinstance(label, int), f\"Label {label} not found in segmentation\"\n",
    "    image = result.segment_image(label)\n",
    "    image = sitk.Cast(image, sitk.sitkUInt8)\n",
    "    return image\n",
    "\n",
    "\n",
    "def resize_label(img: sitk.Image, ref_img: sitk.Image, interp=sitk.sitkNearestNeighbor):\n",
    "    resampler = sitk.ResampleImageFilter()\n",
    "    resampler.SetReferenceImage(ref_img)\n",
    "    resampler.SetInterpolator(interp)\n",
    "    resampler.SetDefaultPixelValue(0)\n",
    "    resampled_img = resampler.Execute(img)\n",
    "    return resampled_img\n",
    "\n",
    "\n",
    "def calc_metrics_for_label(\n",
    "    ai_img, qa_img, series_uid, label_value=1, version=\"aimiv2\", label_suffix=\"\"\n",
    "):\n",
    "    qa_img = resize_label(qa_img, ai_img)  # match the size of the ai_img\n",
    "    ai_arr = sitk.GetArrayFromImage(ai_img)\n",
    "    qa_arr = sitk.GetArrayFromImage(qa_img)\n",
    "    spacing = ai_img.GetSpacing()[::-1]  # numpy is reversed dimensions from sitk\n",
    "\n",
    "    qa_img = resize_label(qa_img, ai_img)  # match the size of the ai_img\n",
    "    m = sm.SegmentationMetrics(ai_arr, qa_arr, spacing)\n",
    "    metrics = {\n",
    "        \"dice\": m.dice,\n",
    "        \"hausdorff_distance_95\": m.hausdorff_distance,\n",
    "        \"mean_surface_distance\": m.mean_surface_distance,\n",
    "        \"mean_surface_distance_tol_4\": compute_surface_dice_at_tolerance(\n",
    "            m._surface_dist, 4\n",
    "        ),\n",
    "        \"SeriesUID\": series_uid,  # from medical segmentation decathlon\n",
    "        \"label\": label_value,\n",
    "        \"version\": version,\n",
    "    }\n",
    "\n",
    "    if label_suffix:\n",
    "        metrics = {f\"{k}_{label_suffix}\": v for k, v in metrics.items()}\n",
    "    return metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download the AIMI segmentations results from zenodo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "qa_dir = Path(\"qa-results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def download_inference(\n",
    "    extract_dir=\"qa-results\", record=\"13244892\", filename=\"prostate-mr.zip\"\n",
    "):\n",
    "    url = f\"https://zenodo.org/record/{record}/files/{filename}\"\n",
    "    # download the zip file and extract it\n",
    "    r = requests.get(url)\n",
    "    z = zipfile.ZipFile(io.BytesIO(r.content))\n",
    "    z.extractall(extract_dir)\n",
    "\n",
    "\n",
    "if not qa_dir.exists() or not (qa_dir / \"ai-segmentations-dcm\").exists():\n",
    "    download_inference(qa_dir)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download `Prostate-MRI-US-Biopsy-DICOM-Annotations` segmentations from IDC\n",
    "Create a file `gcp_project.txt` with the GCP project name to use to download the segmentations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate s5 download script\n",
    "s5_download_script = Path(\"Prostate-MRI-US-Biopsy-DICOM-Annotations.s5cmd\")\n",
    "\n",
    "if not s5_download_script.exists():\n",
    "    with open(\"gcp_project.txt\", \"r\") as file:\n",
    "        project_id = file.read().replace(\"\\n\", \"\")\n",
    "\n",
    "    # python API is the most flexible way to query IDC BigQuery metadata tables\n",
    "    bq_client = bigquery.Client(project_id)\n",
    "\n",
    "    selection_query = f\"\"\"\n",
    "  SELECT\n",
    "    # Organize the files in-place right after downloading\n",
    "    ANY_VALUE(CONCAT(\"cp s3\",REGEXP_SUBSTR(gcs_url, \"(://.*)/\"),\"/* \",analysis_result_id,\"/\",PatientID,\"/\",StudyInstanceUID,\"/\",SeriesInstanceUID)) AS s5cmd_command\n",
    "  FROM\n",
    "    `bigquery-public-data.idc_current.dicom_all`\n",
    "  WHERE\n",
    "    collection_id=\"prostate_mri_us_biopsy\" AND Modality=\"SEG\" AND analysis_result_id='Prostate-MRI-US-Biopsy-DICOM-Annotations'\n",
    "  GROUP BY\n",
    "    SeriesInstanceUID\n",
    "  \"\"\"\n",
    "\n",
    "    selection_result = bq_client.query(selection_query)\n",
    "    selection_df = selection_result.result().to_dataframe()\n",
    "    selection_df.to_csv(s5_download_script, header=False, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Download the Prostate-MRI-US-Biopsy-DICOM-Annotations dataset\n",
    "s5_download_dir = Path(\"Prostate-MRI-US-Biopsy-DICOM-Annotations\")\n",
    "if not s5_download_dir.exists():\n",
    "    subprocess.run(\n",
    "        f\"s5cmd --no-sign-request --endpoint-url https://storage.googleapis.com run {s5_download_script}\",\n",
    "        shell=True,\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c874a9ca496e413ea7dcb9e5dc34bb3b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/2328 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# segmentation file referenced series instance UID -> filepath, if that seg file has a label 'prostate'\n",
    "segment_file_map = {}\n",
    "\n",
    "files = list((s5_download_dir.rglob(\"*.dcm\")))\n",
    "for f in tqdm(files):\n",
    "    dcm = pydicom.dcmread(f)\n",
    "    reader = pydicom_seg.SegmentReader()\n",
    "    result = reader.read(dcm)\n",
    "    for segment_number in result.available_segments:\n",
    "        if result.segment_infos[segment_number].SegmentLabel == \"Prostate\":\n",
    "            segment_file_map[dcm.ReferencedSeriesSequence[0].SeriesInstanceUID] = f\n",
    "            break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Performance on other datasets\n",
    "\n",
    "Lets check how the model performs on:\n",
    "* Task 5 of the [Medical Segmentation Decathlon](http://medicaldecathlon.com/)\n",
    "* [PROMISE12](https://promise12.grand-challenge.org/)\n",
    "* [QIN-Prostate-Repeatability](https://wiki.cancerimagingarchive.net/display/Public/QIN-PROSTATE-Repeatability)\n",
    "\n",
    "Use the `prostate_mr_qa_datasets.ipynb` notebook to process and download these datasets in nifti format.\n",
    "\n",
    "Extract the zip files and you should have the following folders:\n",
    "\n",
    "* MSD/Task05_Prostate/T2imagesTr/\n",
    "* MSD/Task05_Prostate/prostateLabelsTr/\n",
    "* promise12/imagesTr/\n",
    "* promise12/labelsTr/\n",
    "* qin-prostate-repeatability/imagesTr/\n",
    "* qin-prostate-repeatability/labelsTr/\n",
    "\n",
    "### Run model\n",
    "\n",
    "Run the model that was published to [mhub.ai](https://mhub.ai/models/bamf_nnunet_mr_prostate).\n",
    "\n",
    "```bash\n",
    "docker run --rm -t --gpus all -v $in:/app/data/input_data -v $out:/app/data/output_data mhubai/bamf_nnunet_mr_prostate\n",
    "\n",
    "```\n",
    "Run the following commands to run the model in the container and predict the prostate segmentation for the three datasets.\n",
    "\n",
    "* MSD\n",
    "\n",
    "```bash\n",
    "mkdir -p {REPO_DIR}/MSD/Task05_Prostate/predTr\n",
    "docker run --gpus all -v {REPO_DIR}/MSD/Task05_Prostate/T2imagesTr:/data/input -v {REPO_DIR}/MSD/Task05_Prostate/predTr:/data/output bamf_prostate_mr:latest --nifti\n",
    "```\n",
    "\n",
    "* PROMISE12\n",
    "\n",
    "```bash\n",
    "mkdir -p {REPO_DIR}/promise12/predTr\n",
    "docker run --gpus all -v {REPO_DIR}/promise12/imagesTr:/data/input -v {REPO_DIR}/promise12/predTr:/data/output bamf_prostate_mr:latest --nifti\n",
    "```\n",
    "\n",
    "* QIN-Prostate-Repeatability (QPR)\n",
    "\n",
    "```bash\n",
    "mkdir -p {REPO_DIR}/qin-prostate-repeatability/predTr\n",
    "docker run --gpus all -v {REPO_DIR}/qin-prostate-repeatability/imagesTr:/data/input -v {REPO_DIR}/qin-prostate-repeatability/predTr:/data/output bamf_prostate_mr:latest --nifti\n",
    "```\n",
    "\n",
    "Then we can calcuate metrics for these datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compare segmentations\n",
    "Compare the segmentations from the AIMI 2024 project to the segmentations from the IDC PROSTATE-MRI-US-BIOPSY collection."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(qa_dir / \"qa-results.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cde58a26371f425ebdc0317d33e5963c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/817 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.read_csv(qa_dir / \"qa-results.csv\")\n",
    "df = df[df[\"Collection\"] == \"prostate_mri_us_biopsy\"]\n",
    "metrics = []\n",
    "\n",
    "for _, row in tqdm(df.iterrows(), total=len(df)):\n",
    "    seg_a_file = qa_dir / \"ai-segmentations-dcm\" / row[\"AISegmentation\"]\n",
    "    assert seg_a_file.exists(), f\"AI segmentation {seg_a_file} does not exist\"\n",
    "    seg_b_file = segment_file_map.get(row[\"SeriesInstanceUID\"])\n",
    "    if seg_b_file is None:\n",
    "        print(f\"QA segmentation not found for {row['SeriesInstanceUID']}\")\n",
    "        continue\n",
    "    seg_qa_file = None\n",
    "    if not pd.isna(row[\"CorrectedSegmentation\"]):\n",
    "        seg_qa_file = qa_dir / \"qa-segmentations-dcm\" / row[\"CorrectedSegmentation\"]\n",
    "        assert seg_qa_file.exists(), f\"QA segmentation {seg_qa_file} does not exist\"\n",
    "\n",
    "    img_a = load_dcm_seg(seg_a_file, 1)\n",
    "    img_b = load_dcm_seg(seg_b_file, \"Prostate\")\n",
    "    metric = calc_metrics_for_label(\n",
    "        img_a,\n",
    "        img_b,\n",
    "        row.SeriesInstanceUID,\n",
    "        version=\"Prostate-MRI-US-Biopsy-DICOM-Annotations\",\n",
    "    )\n",
    "\n",
    "    metrics.append(metric)\n",
    "\n",
    "metrics_df = pd.DataFrame(metrics)\n",
    "\n",
    "# Replace infinite values with NaN and drop NaN values\n",
    "metrics_df = metrics_df.replace([np.inf, -np.inf], np.nan).dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_df.to_csv(\"metrics-other.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">dice</th>\n",
       "      <th colspan=\"2\" halign=\"left\">hausdorff_distance_95</th>\n",
       "      <th colspan=\"2\" halign=\"left\">mean_surface_distance</th>\n",
       "      <th colspan=\"2\" halign=\"left\">mean_surface_distance_tol_4</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>label</th>\n",
       "      <th>version</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <th>Prostate-MRI-US-Biopsy-DICOM-Annotations</th>\n",
       "      <td>0.89</td>\n",
       "      <td>0.05</td>\n",
       "      <td>4.14</td>\n",
       "      <td>3.35</td>\n",
       "      <td>1.15</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.96</td>\n",
       "      <td>0.06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                dice        \\\n",
       "                                                mean   std   \n",
       "label version                                                \n",
       "1     Prostate-MRI-US-Biopsy-DICOM-Annotations  0.89  0.05   \n",
       "\n",
       "                                               hausdorff_distance_95        \\\n",
       "                                                                mean   std   \n",
       "label version                                                                \n",
       "1     Prostate-MRI-US-Biopsy-DICOM-Annotations                  4.14  3.35   \n",
       "\n",
       "                                               mean_surface_distance        \\\n",
       "                                                                mean   std   \n",
       "label version                                                                \n",
       "1     Prostate-MRI-US-Biopsy-DICOM-Annotations                  1.15  0.66   \n",
       "\n",
       "                                               mean_surface_distance_tol_4  \\\n",
       "                                                                      mean   \n",
       "label version                                                                \n",
       "1     Prostate-MRI-US-Biopsy-DICOM-Annotations                        0.96   \n",
       "\n",
       "                                                      \n",
       "                                                 std  \n",
       "label version                                         \n",
       "1     Prostate-MRI-US-Biopsy-DICOM-Annotations  0.06  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Group by 'label' and 'version' and aggregate metrics\n",
    "aggregated_df = (\n",
    "    metrics_df.groupby([\"label\", \"version\"])\n",
    "    .agg(\n",
    "        {\n",
    "            \"dice\": [\"mean\", \"std\"],\n",
    "            \"hausdorff_distance_95\": [\"mean\", \"std\"],\n",
    "            \"mean_surface_distance\": [\"mean\", \"std\"],\n",
    "            \"mean_surface_distance_tol_4\": [\"mean\", \"std\"],\n",
    "        }\n",
    "    )\n",
    "    .round(2)\n",
    ")\n",
    "\n",
    "aggregated_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyError",
     "evalue": "'mean_surface_distance_tol_7'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "File \u001b[0;32m~/projects/aimi-prostate-mr/.venv/lib/python3.11/site-packages/pandas/core/indexes/base.py:3790\u001b[0m, in \u001b[0;36mIndex.get_loc\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   3789\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m-> 3790\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_engine\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mget_loc\u001b[49m\u001b[43m(\u001b[49m\u001b[43mcasted_key\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   3791\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mKeyError\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m err:\n",
      "File \u001b[0;32mindex.pyx:152\u001b[0m, in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32mindex.pyx:181\u001b[0m, in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32mpandas/_libs/hashtable_class_helper.pxi:7080\u001b[0m, in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32mpandas/_libs/hashtable_class_helper.pxi:7088\u001b[0m, in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;31mKeyError\u001b[0m: 'mean_surface_distance_tol_7'",
      "\nThe above exception was the direct cause of the following exception:\n",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[14], line 21\u001b[0m\n\u001b[1;32m     18\u001b[0m ax \u001b[38;5;241m=\u001b[39m axs[row, col]\n\u001b[1;32m     20\u001b[0m \u001b[38;5;66;03m# Extract data for current metric\u001b[39;00m\n\u001b[0;32m---> 21\u001b[0m label1_data \u001b[38;5;241m=\u001b[39m \u001b[43mlabel1_df\u001b[49m\u001b[43m[\u001b[49m\u001b[43mmetric\u001b[49m\u001b[43m]\u001b[49m\n\u001b[1;32m     23\u001b[0m \u001b[38;5;66;03m# Check if label2_data has values and is not empty\u001b[39;00m\n\u001b[1;32m     24\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m label1_data\u001b[38;5;241m.\u001b[39mempty:\n\u001b[1;32m     25\u001b[0m     \u001b[38;5;66;03m# Convert data to numeric if necessary\u001b[39;00m\n",
      "File \u001b[0;32m~/projects/aimi-prostate-mr/.venv/lib/python3.11/site-packages/pandas/core/frame.py:3896\u001b[0m, in \u001b[0;36mDataFrame.__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   3894\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mcolumns\u001b[38;5;241m.\u001b[39mnlevels \u001b[38;5;241m>\u001b[39m \u001b[38;5;241m1\u001b[39m:\n\u001b[1;32m   3895\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_getitem_multilevel(key)\n\u001b[0;32m-> 3896\u001b[0m indexer \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcolumns\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mget_loc\u001b[49m\u001b[43m(\u001b[49m\u001b[43mkey\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   3897\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m is_integer(indexer):\n\u001b[1;32m   3898\u001b[0m     indexer \u001b[38;5;241m=\u001b[39m [indexer]\n",
      "File \u001b[0;32m~/projects/aimi-prostate-mr/.venv/lib/python3.11/site-packages/pandas/core/indexes/base.py:3797\u001b[0m, in \u001b[0;36mIndex.get_loc\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   3792\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(casted_key, \u001b[38;5;28mslice\u001b[39m) \u001b[38;5;129;01mor\u001b[39;00m (\n\u001b[1;32m   3793\u001b[0m         \u001b[38;5;28misinstance\u001b[39m(casted_key, abc\u001b[38;5;241m.\u001b[39mIterable)\n\u001b[1;32m   3794\u001b[0m         \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;28many\u001b[39m(\u001b[38;5;28misinstance\u001b[39m(x, \u001b[38;5;28mslice\u001b[39m) \u001b[38;5;28;01mfor\u001b[39;00m x \u001b[38;5;129;01min\u001b[39;00m casted_key)\n\u001b[1;32m   3795\u001b[0m     ):\n\u001b[1;32m   3796\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m InvalidIndexError(key)\n\u001b[0;32m-> 3797\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mKeyError\u001b[39;00m(key) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01merr\u001b[39;00m\n\u001b[1;32m   3798\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mTypeError\u001b[39;00m:\n\u001b[1;32m   3799\u001b[0m     \u001b[38;5;66;03m# If we have a listlike key, _check_indexing_error will raise\u001b[39;00m\n\u001b[1;32m   3800\u001b[0m     \u001b[38;5;66;03m#  InvalidIndexError. Otherwise we fall through and re-raise\u001b[39;00m\n\u001b[1;32m   3801\u001b[0m     \u001b[38;5;66;03m#  the TypeError.\u001b[39;00m\n\u001b[1;32m   3802\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_check_indexing_error(key)\n",
      "\u001b[0;31mKeyError\u001b[0m: 'mean_surface_distance_tol_7'"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x1000 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label1_df = metrics_df[metrics_df.label == 1]\n",
    "\n",
    "# Create a figure and axis for subplots\n",
    "fig, axs = plt.subplots(nrows=2, ncols=2, figsize=(12, 10))\n",
    "\n",
    "# List of metrics to plot\n",
    "metrics = [\n",
    "    \"dice\",\n",
    "    \"hausdorff_distance_95\",\n",
    "    \"mean_surface_distance\",\n",
    "    \"mean_surface_distance_tol_7\",\n",
    "]\n",
    "\n",
    "# Plot each metric\n",
    "for i, metric in enumerate(metrics):\n",
    "    row = i // 2\n",
    "    col = i % 2\n",
    "    ax = axs[row, col]\n",
    "\n",
    "    # Extract data for current metric\n",
    "    label1_data = label1_df[metric]\n",
    "\n",
    "    # Check if label2_data has values and is not empty\n",
    "    if not label1_data.empty:\n",
    "        # Convert data to numeric if necessary\n",
    "        label1_data = pd.to_numeric(label1_data, errors=\"coerce\")\n",
    "\n",
    "        # Drop NaN values if any (optional)\n",
    "        label1_data = label1_data.dropna()\n",
    "\n",
    "        # Check if after dropping NaN, there's still data\n",
    "        if not label1_data.empty:\n",
    "            # Create boxplots side by side\n",
    "            ax.boxplot([label1_data], patch_artist=True, showmeans=True)\n",
    "            ax.set_title(f\"{metric.split('aimiv2_')[-1]}\")\n",
    "            ax.set_xticklabels([\"Prostate\"])  # Update x-axis labels to \"Lesions\"\n",
    "        else:\n",
    "            # Handle case where data might be dropped to empty\n",
    "            print(f\"No valid data available for {metric} after cleaning.\")\n",
    "    else:\n",
    "        # If label2_data is empty, handle accordingly (e.g., print a message)\n",
    "        print(f\"No data available for {metric} in label2 (Lesions).\")\n",
    "\n",
    "# Adjust layout\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}