test_performances.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "195fc131-35dc-4fac-91c8-2dbfe57188ca",
   "metadata": {},
   "source": [
    "# Test performances"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4553c745-73d8-4dc4-871f-0a55388eea58",
   "metadata": {},
   "source": [
    "## Test data I/O"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "253f926d-274e-4e64-82d1-311151801da6",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "### Test dense data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "067bc416-902c-4b8d-bdad-e262e2cda634",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-06-12 09:49:20.203886: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-06-12 09:49:20.304351: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2023-06-12 09:49:20.328509: E tensorflow/stream_executor/cuda/cuda_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
      "<frozen importlib._bootstrap>:228: RuntimeWarning: scipy._lib.messagestream.MessageStream size changed, may indicate binary incompatibility. Expected 56 from C header, got 64 from PyObject\n",
      "2023-06-12 09:49:29.630871: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-06-12 09:49:30.888566: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1616] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 14783 MB memory:  -> device: 0, name: Quadro RTX 5000, pci bus id: 0000:17:00.0, compute capability: 7.5\n",
      "2023-06-12 09:49:30.889421: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1616] Created device /job:localhost/replica:0/task:0/device:GPU:1 with 14485 MB memory:  -> device: 1, name: Quadro RTX 5000, pci bus id: 0000:65:00.0, compute capability: 7.5\n"
     ]
    }
   ],
   "source": [
    "import importlib\n",
    "import numpy as np\n",
    "import nibabel as nib\n",
    "import tensorflow as tf\n",
    "\n",
    "from utils import is_equal, load_data\n",
    "from utils.med_utils import data_io, pre_processing, load_medical_data\n",
    "\n",
    "path = '/storage/Totalsegmentator_dataset/s0001/ct'\n",
    "\n",
    "data = nib.load(path + '.nii.gz')\n",
    "val  = data.get_fdata('unchanged')\n",
    "\n",
    "np.save(path + '-1.5-1.5-1.5.npy', val)\n",
    "np.savez(path + '.npz', data = val, pixdim = data.header['pixdim'][1:4])\n",
    "tf.io.write_file(path + '-1.5-1.5-1.5.tensor', tf.io.serialize_tensor(tf.cast(val, tf.float32)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e846b5a8-d62a-4874-b5d5-58c52aa9083a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Test with kwargs : {}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=3>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=3>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=3>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>, 'slice_axis': <tf.Tensor: shape=(), dtype=int32, numpy=0>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n"
     ]
    }
   ],
   "source": [
    "importlib.reload(pre_processing)\n",
    "importlib.reload(data_io)\n",
    "\n",
    "tf_load_medical_data = tf.function(data_io.load_medical_data)\n",
    "\n",
    "test_kwargs = [\n",
    "    {},\n",
    "    {'slice_start' : tf.cast(3, tf.int32)},\n",
    "    {'slice_end' : tf.cast(6, tf.int32)},\n",
    "    {'slice_start' : tf.cast(3, tf.int32), 'slice_end' : tf.cast(6, tf.int32)},\n",
    "    {'slice_start' : tf.cast(3, tf.int32), 'slice_end' : tf.cast(6, tf.int32), 'slice_axis' : tf.cast(0, tf.int32)}\n",
    "]\n",
    "for kwargs in test_kwargs:\n",
    "    print('\\nTest with kwargs : {}'.format(kwargs))\n",
    "    r1 = data_io.load_medical_data(path + '.nii.gz', ** kwargs)\n",
    "    r2 = data_io.load_medical_data(path + '-1.5-1.5-1.5.npy', ** kwargs)\n",
    "    r3 = data_io.load_medical_data(path + '.npz', ** kwargs)\n",
    "    r4 = data_io.load_medical_data(path + '-1.5-1.5-1.5.tensor', ** kwargs)\n",
    "\n",
    "    r5 = data_io.load_medical_data(tf.cast(path + '.nii.gz', tf.string), ** kwargs)\n",
    "    r6 = data_io.load_medical_data(tf.cast(path + '-1.5-1.5-1.5.npy', tf.string), ** kwargs)\n",
    "    r7 = data_io.load_medical_data(tf.cast(path + '.npz', tf.string), ** kwargs)\n",
    "    r8 = data_io.load_medical_data(tf.cast(path + '-1.5-1.5-1.5.tensor', tf.string), ** kwargs)\n",
    "\n",
    "    r9  = tf_load_medical_data(tf.cast(path + '.nii.gz', tf.string), ** kwargs)\n",
    "    r10 = tf_load_medical_data(tf.cast(path + '-1.5-1.5-1.5.npy', tf.string), ** kwargs)\n",
    "    r11 = tf_load_medical_data(tf.cast(path + '.npz', tf.string), ** kwargs)\n",
    "    r12 = tf_load_medical_data(tf.cast(path + '-1.5-1.5-1.5.tensor', tf.string), ** kwargs)\n",
    "\n",
    "    print(is_equal(r1, r2)[1])\n",
    "    print(is_equal(r1, r3)[1])\n",
    "    print(is_equal(r1, r4)[1])\n",
    "\n",
    "    print(is_equal(r1, r5)[1])\n",
    "    print(is_equal(r1, r6)[1])\n",
    "    print(is_equal(r1, r7)[1])\n",
    "    print(is_equal(r1, r8)[1])\n",
    "\n",
    "    print(is_equal(r1, r9)[1])\n",
    "    print(is_equal(r1, r10)[1])\n",
    "    print(is_equal(r1, r11)[1])\n",
    "    print(is_equal(r1, r12)[1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "dabb495b-2729-4137-b3b7-a389904e07e3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "182 ms ± 57.3 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "25 ms ± 174 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "141 ms ± 37 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "32 ms ± 148 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "\n",
      "226 ms ± 175 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "69.4 ms ± 340 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "186 ms ± 189 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "32.1 ms ± 167 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "\n",
      "232 ms ± 674 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "74.2 ms ± 519 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "190 ms ± 243 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "31.4 ms ± 218 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit load_medical_data(path + '.nii.gz')\n",
    "%timeit load_medical_data(path + '-1.5-1.5-1.5.npy')\n",
    "%timeit load_medical_data(path + '.npz')\n",
    "%timeit load_medical_data(path + '-1.5-1.5-1.5.tensor')\n",
    "print()\n",
    "%timeit load_medical_data(tf.cast(path + '.nii.gz', tf.string))\n",
    "%timeit load_medical_data(tf.cast(path + '-1.5-1.5-1.5.npy', tf.string))\n",
    "%timeit load_medical_data(tf.cast(path + '.npz', tf.string))\n",
    "%timeit load_medical_data(tf.cast(path + '-1.5-1.5-1.5.tensor', tf.string))\n",
    "print()\n",
    "%timeit tf_load_medical_data(tf.cast(path + '.nii.gz', tf.string))\n",
    "%timeit tf_load_medical_data(tf.cast(path + '-1.5-1.5-1.5.npy', tf.string))\n",
    "%timeit tf_load_medical_data(tf.cast(path + '.npz', tf.string))\n",
    "%timeit tf_load_medical_data(tf.cast(path + '-1.5-1.5-1.5.tensor', tf.string))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5aac0031-ceaa-43e2-8feb-b05f1e63ef63",
   "metadata": {},
   "source": [
    "### Test sparse data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d4e97622-a81d-436a-b459-d04ee6e9bd80",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-06-12 09:41:14.336458: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-06-12 09:41:14.433925: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2023-06-12 09:41:14.458382: E tensorflow/stream_executor/cuda/cuda_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
      "<frozen importlib._bootstrap>:228: RuntimeWarning: scipy._lib.messagestream.MessageStream size changed, may indicate binary incompatibility. Expected 56 from C header, got 64 from PyObject\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting to sparse\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-06-12 09:41:25.698124: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving...\n",
      "CPU times: user 0 ns, sys: 2.61 s, total: 2.61 s\n",
      "Wall time: 2.62 s\n",
      "CPU times: user 41 ms, sys: 30.9 ms, total: 71.9 ms\n",
      "Wall time: 71.9 ms\n",
      "CPU times: user 26.4 ms, sys: 31.3 ms, total: 57.7 ms\n",
      "Wall time: 57.8 ms\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import importlib\n",
    "import numpy as np\n",
    "import nibabel as nib\n",
    "import tensorflow as tf\n",
    "\n",
    "from utils import is_equal, load_data\n",
    "from utils.med_utils.label_utils import TOTALSEGMENTATOR_LABELS\n",
    "from utils.med_utils import data_io, pre_processing, label_utils, sparse_utils, load_medical_data\n",
    "\n",
    "tf.config.set_visible_devices([], 'GPU')\n",
    "\n",
    "path = '/storage/Totalsegmentator_dataset/s0001/masks'\n",
    "\n",
    "labels = sorted([f[:-7] for f in os.listdir(path[:-5] + 'segmentations/')])\n",
    "\n",
    "data = nib.load(path + '.nii.gz')\n",
    "val  = data.get_fdata('unchanged')\n",
    "\n",
    "print('Converting to sparse')\n",
    "path_infos = '-1.5-1.5-1.5-{}'.format('-'.join([str(s) for s in data.shape]))\n",
    "with tf.device('cpu'):\n",
    "    tensor = tf.cast(val, tf.uint8)\n",
    "    sparse = tf.sparse.from_dense(tensor)\n",
    "\n",
    "print('Saving...')\n",
    "%time np.save(path + path_infos + '.npy', val)\n",
    "#%time np.savez(path + '.npz', data = val, pixdim = data.header['pixdim'][1:4], labels = labels)\n",
    "%time np.savez(path + '-sparse.npz', mask = sparse.indices, pixdim = data.header['pixdim'][1:4], labels = labels, shape = val.shape)\n",
    "#%time tf.io.write_file(path + path_infos + '.tensor', tf.io.serialize_tensor(tensor))\n",
    "%time tf.io.write_file(path + path_infos + '.stensor', tf.io.serialize_tensor(sparse.indices))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2c8b52a8-ac9c-452b-b110-e456155c1f57",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Test with kwargs : {'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=3>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>}\n",
      "WARNING:tensorflow:5 out of the last 9 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "5 out of the last 9 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "WARNING:tensorflow:6 out of the last 11 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "6 out of the last 11 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=3>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=6>, 'slice_axis': <tf.Tensor: shape=(), dtype=int32, numpy=0>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=8>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=64>, 'labels': ['adrenal_gland_left', 'adrenal_gland_right', 'aorta', 'autochthon_left', 'autochthon_right', 'brain', 'clavicula_left', 'clavicula_right', 'colon', 'duodenum', 'esophagus', 'face', 'femur_left', 'femur_right', 'gallbladder', 'gluteus_maximus_left', 'gluteus_maximus_right', 'gluteus_medius_left', 'gluteus_medius_right', 'gluteus_minimus_left', 'gluteus_minimus_right', 'heart_atrium_left', 'heart_atrium_right', 'heart_myocardium', 'heart_ventricle_left', 'heart_ventricle_right', 'hip_left', 'hip_right', 'humerus_left', 'humerus_right', 'iliac_artery_left', 'iliac_artery_right', 'iliac_vena_left', 'iliac_vena_right', 'iliopsoas_left', 'iliopsoas_right', 'inferior_vena_cava', 'kidney_left', 'kidney_right', 'liver', 'lung_lower_lobe_left', 'lung_lower_lobe_right', 'lung_middle_lobe_right', 'lung_upper_lobe_left', 'lung_upper_lobe_right', 'pancreas', 'portal_vein_and_splenic_vein', 'pulmonary_artery', 'rib_left_1', 'rib_left_10', 'rib_left_11', 'rib_left_12', 'rib_left_2', 'rib_left_3', 'rib_left_4', 'rib_left_5', 'rib_left_6', 'rib_left_7', 'rib_left_8', 'rib_left_9', 'rib_right_1', 'rib_right_10', 'rib_right_11', 'rib_right_12', 'rib_right_2', 'rib_right_3', 'rib_right_4', 'rib_right_5', 'rib_right_6', 'rib_right_7', 'rib_right_8', 'rib_right_9', 'sacrum', 'scapula_left', 'scapula_right', 'small_bowel', 'spleen', 'stomach', 'trachea', 'urinary_bladder', 'vertebrae_C1', 'vertebrae_C2', 'vertebrae_C3', 'vertebrae_C4', 'vertebrae_C5', 'vertebrae_C6', 'vertebrae_C7', 'vertebrae_L1', 'vertebrae_L2', 'vertebrae_L3', 'vertebrae_L4', 'vertebrae_L5', 'vertebrae_T1', 'vertebrae_T10', 'vertebrae_T11', 'vertebrae_T12', 'vertebrae_T2', 'vertebrae_T3', 'vertebrae_T4', 'vertebrae_T5', 'vertebrae_T6', 'vertebrae_T7', 'vertebrae_T8', 'vertebrae_T9'], 'mapping': [None, 'adrenal_gland_left', 'adrenal_gland_right', 'aorta', 'autochthon_left', 'autochthon_right', 'brain', 'clavicula_left', 'clavicula_right', 'colon', 'duodenum', 'esophagus', 'face', 'femur_left', 'femur_right', 'gallbladder', 'gluteus_maximus_left', 'gluteus_maximus_right', 'gluteus_medius_left', 'gluteus_medius_right', 'gluteus_minimus_left', 'gluteus_minimus_right', 'heart_atrium_left', 'heart_atrium_right', 'heart_myocardium', 'heart_ventricle_left', 'heart_ventricle_right', 'hip_left', 'hip_right', 'humerus_left', 'humerus_right', 'iliac_artery_left', 'iliac_artery_right', 'iliac_vena_left', 'iliac_vena_right', 'iliopsoas_left', 'iliopsoas_right', 'inferior_vena_cava', 'kidney_left', 'kidney_right', 'liver', 'lung_lower_lobe_left', 'lung_lower_lobe_right', 'lung_middle_lobe_right', 'lung_upper_lobe_left', 'lung_upper_lobe_right', 'pancreas', 'portal_vein_and_splenic_vein', 'pulmonary_artery', 'rib_left_1', 'rib_left_10', 'rib_left_11', 'rib_left_12', 'rib_left_2', 'rib_left_3', 'rib_left_4', 'rib_left_5', 'rib_left_6', 'rib_left_7', 'rib_left_8', 'rib_left_9', 'rib_right_1', 'rib_right_10', 'rib_right_11', 'rib_right_12', 'rib_right_2', 'rib_right_3', 'rib_right_4', 'rib_right_5', 'rib_right_6', 'rib_right_7', 'rib_right_8', 'rib_right_9', 'sacrum', 'scapula_left', 'scapula_right', 'small_bowel', 'spleen', 'stomach', 'trachea', 'urinary_bladder', 'vertebrae_C1', 'vertebrae_C2', 'vertebrae_C3', 'vertebrae_C4', 'vertebrae_C5', 'vertebrae_C6', 'vertebrae_C7', 'vertebrae_L1', 'vertebrae_L2', 'vertebrae_L3', 'vertebrae_L4', 'vertebrae_L5', 'vertebrae_T1', 'vertebrae_T10', 'vertebrae_T11', 'vertebrae_T12', 'vertebrae_T2', 'vertebrae_T3', 'vertebrae_T4', 'vertebrae_T5', 'vertebrae_T6', 'vertebrae_T7', 'vertebrae_T8', 'vertebrae_T9']}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=8>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=64>, 'labels': ['adrenal_gland_left', 'adrenal_gland_right', 'aorta', 'autochthon_left', 'autochthon_right', 'brain', 'clavicula_left', 'clavicula_right', 'colon', 'duodenum', 'esophagus', 'face', 'femur_left', 'femur_right', 'gallbladder', 'gluteus_maximus_left', 'gluteus_maximus_right', 'gluteus_medius_left', 'gluteus_medius_right', 'gluteus_minimus_left', 'gluteus_minimus_right', 'heart_atrium_left', 'heart_atrium_right', 'heart_myocardium', 'heart_ventricle_left', 'heart_ventricle_right', 'hip_left', 'hip_right', 'humerus_left', 'humerus_right', 'iliac_artery_left', 'iliac_artery_right', 'iliac_vena_left', 'iliac_vena_right', 'iliopsoas_left', 'iliopsoas_right', 'inferior_vena_cava', 'kidney_left', 'kidney_right', 'liver', 'lung_lower_lobe_left', 'lung_lower_lobe_right', 'lung_middle_lobe_right', 'lung_upper_lobe_left', 'lung_upper_lobe_right', 'pancreas', 'portal_vein_and_splenic_vein', 'pulmonary_artery', 'rib_left_1', 'rib_left_10', 'rib_left_11', 'rib_left_12', 'rib_left_2', 'rib_left_3', 'rib_left_4', 'rib_left_5', 'rib_left_6', 'rib_left_7', 'rib_left_8', 'rib_left_9', 'rib_right_1', 'rib_right_10', 'rib_right_11', 'rib_right_12', 'rib_right_2', 'rib_right_3', 'rib_right_4', 'rib_right_5', 'rib_right_6', 'rib_right_7', 'rib_right_8', 'rib_right_9', 'sacrum', 'scapula_left', 'scapula_right', 'small_bowel', 'spleen', 'stomach', 'trachea', 'urinary_bladder', 'vertebrae_C1', 'vertebrae_C2', 'vertebrae_C3', 'vertebrae_C4', 'vertebrae_C5', 'vertebrae_C6', 'vertebrae_C7', 'vertebrae_L1', 'vertebrae_L2', 'vertebrae_L3', 'vertebrae_L4', 'vertebrae_L5', 'vertebrae_T1', 'vertebrae_T10', 'vertebrae_T11', 'vertebrae_T12', 'vertebrae_T2', 'vertebrae_T3', 'vertebrae_T4', 'vertebrae_T5', 'vertebrae_T6', 'vertebrae_T7', 'vertebrae_T8', 'vertebrae_T9'], 'mapping': [None, 'spleen', 'kidney_right', 'kidney_left', 'gallbladder', 'liver', 'stomach', 'aorta', 'inferior_vena_cava', 'portal_vein_and_splenic_vein', 'pancreas', 'adrenal_gland_right', 'adrenal_gland_left', 'lung_upper_lobe_left', 'lung_lower_lobe_left', 'lung_upper_lobe_right', 'lung_middle_lobe_right', 'lung_lower_lobe_right', 'vertebrae_L5', 'vertebrae_L4', 'vertebrae_L3', 'vertebrae_L2', 'vertebrae_L1', 'vertebrae_T12', 'vertebrae_T11', 'vertebrae_T10', 'vertebrae_T9', 'vertebrae_T8', 'vertebrae_T7', 'vertebrae_T6', 'vertebrae_T5', 'vertebrae_T4', 'vertebrae_T3', 'vertebrae_T2', 'vertebrae_T1', 'vertebrae_C7', 'vertebrae_C6', 'vertebrae_C5', 'vertebrae_C4', 'vertebrae_C3', 'vertebrae_C2', 'vertebrae_C1', 'esophagus', 'trachea', 'heart_myocardium', 'heart_atrium_left', 'heart_ventricle_left', 'heart_atrium_right', 'heart_ventricle_right', 'pulmonary_artery', 'brain', 'iliac_artery_left', 'iliac_artery_right', 'iliac_vena_left', 'iliac_vena_right', 'small_bowel', 'duodenum', 'colon', 'rib_left_1', 'rib_left_2', 'rib_left_3', 'rib_left_4', 'rib_left_5', 'rib_left_6', 'rib_left_7', 'rib_left_8', 'rib_left_9', 'rib_left_10', 'rib_left_11', 'rib_left_12', 'rib_right_1', 'rib_right_2', 'rib_right_3', 'rib_right_4', 'rib_right_5', 'rib_right_6', 'rib_right_7', 'rib_right_8', 'rib_right_9', 'rib_right_10', 'rib_right_11', 'rib_right_12', 'humerus_left', 'humerus_right', 'scapula_left', 'scapula_right', 'clavicula_left', 'clavicula_right', 'femur_left', 'femur_right', 'hip_left', 'hip_right', 'sacrum', 'face', 'gluteus_maximus_left', 'gluteus_maximus_right', 'gluteus_medius_left', 'gluteus_medius_right', 'gluteus_minimus_left', 'gluteus_minimus_right', 'autochthon_left', 'autochthon_right', 'iliopsoas_left', 'iliopsoas_right', 'urinary_bladder']}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=8>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=64>, 'labels': <tf.Tensor: shape=(104,), dtype=string, numpy=\n",
      "array([b'adrenal_gland_left', b'adrenal_gland_right', b'aorta',\n",
      "       b'autochthon_left', b'autochthon_right', b'brain',\n",
      "       b'clavicula_left', b'clavicula_right', b'colon', b'duodenum',\n",
      "       b'esophagus', b'face', b'femur_left', b'femur_right',\n",
      "       b'gallbladder', b'gluteus_maximus_left', b'gluteus_maximus_right',\n",
      "       b'gluteus_medius_left', b'gluteus_medius_right',\n",
      "       b'gluteus_minimus_left', b'gluteus_minimus_right',\n",
      "       b'heart_atrium_left', b'heart_atrium_right', b'heart_myocardium',\n",
      "       b'heart_ventricle_left', b'heart_ventricle_right', b'hip_left',\n",
      "       b'hip_right', b'humerus_left', b'humerus_right',\n",
      "       b'iliac_artery_left', b'iliac_artery_right', b'iliac_vena_left',\n",
      "       b'iliac_vena_right', b'iliopsoas_left', b'iliopsoas_right',\n",
      "       b'inferior_vena_cava', b'kidney_left', b'kidney_right', b'liver',\n",
      "       b'lung_lower_lobe_left', b'lung_lower_lobe_right',\n",
      "       b'lung_middle_lobe_right', b'lung_upper_lobe_left',\n",
      "       b'lung_upper_lobe_right', b'pancreas',\n",
      "       b'portal_vein_and_splenic_vein', b'pulmonary_artery',\n",
      "       b'rib_left_1', b'rib_left_10', b'rib_left_11', b'rib_left_12',\n",
      "       b'rib_left_2', b'rib_left_3', b'rib_left_4', b'rib_left_5',\n",
      "       b'rib_left_6', b'rib_left_7', b'rib_left_8', b'rib_left_9',\n",
      "       b'rib_right_1', b'rib_right_10', b'rib_right_11', b'rib_right_12',\n",
      "       b'rib_right_2', b'rib_right_3', b'rib_right_4', b'rib_right_5',\n",
      "       b'rib_right_6', b'rib_right_7', b'rib_right_8', b'rib_right_9',\n",
      "       b'sacrum', b'scapula_left', b'scapula_right', b'small_bowel',\n",
      "       b'spleen', b'stomach', b'trachea', b'urinary_bladder',\n",
      "       b'vertebrae_C1', b'vertebrae_C2', b'vertebrae_C3', b'vertebrae_C4',\n",
      "       b'vertebrae_C5', b'vertebrae_C6', b'vertebrae_C7', b'vertebrae_L1',\n",
      "       b'vertebrae_L2', b'vertebrae_L3', b'vertebrae_L4', b'vertebrae_L5',\n",
      "       b'vertebrae_T1', b'vertebrae_T10', b'vertebrae_T11',\n",
      "       b'vertebrae_T12', b'vertebrae_T2', b'vertebrae_T3',\n",
      "       b'vertebrae_T4', b'vertebrae_T5', b'vertebrae_T6', b'vertebrae_T7',\n",
      "       b'vertebrae_T8', b'vertebrae_T9'], dtype=object)>, 'mapping': [None, 'adrenal_gland_left', 'adrenal_gland_right', 'aorta', 'autochthon_left', 'autochthon_right', 'brain', 'clavicula_left', 'clavicula_right', 'colon', 'duodenum', 'esophagus', 'face', 'femur_left', 'femur_right', 'gallbladder', 'gluteus_maximus_left', 'gluteus_maximus_right', 'gluteus_medius_left', 'gluteus_medius_right', 'gluteus_minimus_left', 'gluteus_minimus_right', 'heart_atrium_left', 'heart_atrium_right', 'heart_myocardium', 'heart_ventricle_left', 'heart_ventricle_right', 'hip_left', 'hip_right', 'humerus_left', 'humerus_right', 'iliac_artery_left', 'iliac_artery_right', 'iliac_vena_left', 'iliac_vena_right', 'iliopsoas_left', 'iliopsoas_right', 'inferior_vena_cava', 'kidney_left', 'kidney_right', 'liver', 'lung_lower_lobe_left', 'lung_lower_lobe_right', 'lung_middle_lobe_right', 'lung_upper_lobe_left', 'lung_upper_lobe_right', 'pancreas', 'portal_vein_and_splenic_vein', 'pulmonary_artery', 'rib_left_1', 'rib_left_10', 'rib_left_11', 'rib_left_12', 'rib_left_2', 'rib_left_3', 'rib_left_4', 'rib_left_5', 'rib_left_6', 'rib_left_7', 'rib_left_8', 'rib_left_9', 'rib_right_1', 'rib_right_10', 'rib_right_11', 'rib_right_12', 'rib_right_2', 'rib_right_3', 'rib_right_4', 'rib_right_5', 'rib_right_6', 'rib_right_7', 'rib_right_8', 'rib_right_9', 'sacrum', 'scapula_left', 'scapula_right', 'small_bowel', 'spleen', 'stomach', 'trachea', 'urinary_bladder', 'vertebrae_C1', 'vertebrae_C2', 'vertebrae_C3', 'vertebrae_C4', 'vertebrae_C5', 'vertebrae_C6', 'vertebrae_C7', 'vertebrae_L1', 'vertebrae_L2', 'vertebrae_L3', 'vertebrae_L4', 'vertebrae_L5', 'vertebrae_T1', 'vertebrae_T10', 'vertebrae_T11', 'vertebrae_T12', 'vertebrae_T2', 'vertebrae_T3', 'vertebrae_T4', 'vertebrae_T5', 'vertebrae_T6', 'vertebrae_T7', 'vertebrae_T8', 'vertebrae_T9']}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=8>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=64>, 'labels': <tf.Tensor: shape=(104,), dtype=string, numpy=\n",
      "array([b'adrenal_gland_left', b'adrenal_gland_right', b'aorta',\n",
      "       b'autochthon_left', b'autochthon_right', b'brain',\n",
      "       b'clavicula_left', b'clavicula_right', b'colon', b'duodenum',\n",
      "       b'esophagus', b'face', b'femur_left', b'femur_right',\n",
      "       b'gallbladder', b'gluteus_maximus_left', b'gluteus_maximus_right',\n",
      "       b'gluteus_medius_left', b'gluteus_medius_right',\n",
      "       b'gluteus_minimus_left', b'gluteus_minimus_right',\n",
      "       b'heart_atrium_left', b'heart_atrium_right', b'heart_myocardium',\n",
      "       b'heart_ventricle_left', b'heart_ventricle_right', b'hip_left',\n",
      "       b'hip_right', b'humerus_left', b'humerus_right',\n",
      "       b'iliac_artery_left', b'iliac_artery_right', b'iliac_vena_left',\n",
      "       b'iliac_vena_right', b'iliopsoas_left', b'iliopsoas_right',\n",
      "       b'inferior_vena_cava', b'kidney_left', b'kidney_right', b'liver',\n",
      "       b'lung_lower_lobe_left', b'lung_lower_lobe_right',\n",
      "       b'lung_middle_lobe_right', b'lung_upper_lobe_left',\n",
      "       b'lung_upper_lobe_right', b'pancreas',\n",
      "       b'portal_vein_and_splenic_vein', b'pulmonary_artery',\n",
      "       b'rib_left_1', b'rib_left_10', b'rib_left_11', b'rib_left_12',\n",
      "       b'rib_left_2', b'rib_left_3', b'rib_left_4', b'rib_left_5',\n",
      "       b'rib_left_6', b'rib_left_7', b'rib_left_8', b'rib_left_9',\n",
      "       b'rib_right_1', b'rib_right_10', b'rib_right_11', b'rib_right_12',\n",
      "       b'rib_right_2', b'rib_right_3', b'rib_right_4', b'rib_right_5',\n",
      "       b'rib_right_6', b'rib_right_7', b'rib_right_8', b'rib_right_9',\n",
      "       b'sacrum', b'scapula_left', b'scapula_right', b'small_bowel',\n",
      "       b'spleen', b'stomach', b'trachea', b'urinary_bladder',\n",
      "       b'vertebrae_C1', b'vertebrae_C2', b'vertebrae_C3', b'vertebrae_C4',\n",
      "       b'vertebrae_C5', b'vertebrae_C6', b'vertebrae_C7', b'vertebrae_L1',\n",
      "       b'vertebrae_L2', b'vertebrae_L3', b'vertebrae_L4', b'vertebrae_L5',\n",
      "       b'vertebrae_T1', b'vertebrae_T10', b'vertebrae_T11',\n",
      "       b'vertebrae_T12', b'vertebrae_T2', b'vertebrae_T3',\n",
      "       b'vertebrae_T4', b'vertebrae_T5', b'vertebrae_T6', b'vertebrae_T7',\n",
      "       b'vertebrae_T8', b'vertebrae_T9'], dtype=object)>, 'mapping': <tf.Tensor: shape=(105, 1), dtype=string, numpy=\n",
      "array([[b''],\n",
      "       [b'spleen'],\n",
      "       [b'kidney_right'],\n",
      "       [b'kidney_left'],\n",
      "       [b'gallbladder'],\n",
      "       [b'liver'],\n",
      "       [b'stomach'],\n",
      "       [b'aorta'],\n",
      "       [b'inferior_vena_cava'],\n",
      "       [b'portal_vein_and_splenic_vein'],\n",
      "       [b'pancreas'],\n",
      "       [b'adrenal_gland_right'],\n",
      "       [b'adrenal_gland_left'],\n",
      "       [b'lung_upper_lobe_left'],\n",
      "       [b'lung_lower_lobe_left'],\n",
      "       [b'lung_upper_lobe_right'],\n",
      "       [b'lung_middle_lobe_right'],\n",
      "       [b'lung_lower_lobe_right'],\n",
      "       [b'vertebrae_L5'],\n",
      "       [b'vertebrae_L4'],\n",
      "       [b'vertebrae_L3'],\n",
      "       [b'vertebrae_L2'],\n",
      "       [b'vertebrae_L1'],\n",
      "       [b'vertebrae_T12'],\n",
      "       [b'vertebrae_T11'],\n",
      "       [b'vertebrae_T10'],\n",
      "       [b'vertebrae_T9'],\n",
      "       [b'vertebrae_T8'],\n",
      "       [b'vertebrae_T7'],\n",
      "       [b'vertebrae_T6'],\n",
      "       [b'vertebrae_T5'],\n",
      "       [b'vertebrae_T4'],\n",
      "       [b'vertebrae_T3'],\n",
      "       [b'vertebrae_T2'],\n",
      "       [b'vertebrae_T1'],\n",
      "       [b'vertebrae_C7'],\n",
      "       [b'vertebrae_C6'],\n",
      "       [b'vertebrae_C5'],\n",
      "       [b'vertebrae_C4'],\n",
      "       [b'vertebrae_C3'],\n",
      "       [b'vertebrae_C2'],\n",
      "       [b'vertebrae_C1'],\n",
      "       [b'esophagus'],\n",
      "       [b'trachea'],\n",
      "       [b'heart_myocardium'],\n",
      "       [b'heart_atrium_left'],\n",
      "       [b'heart_ventricle_left'],\n",
      "       [b'heart_atrium_right'],\n",
      "       [b'heart_ventricle_right'],\n",
      "       [b'pulmonary_artery'],\n",
      "       [b'brain'],\n",
      "       [b'iliac_artery_left'],\n",
      "       [b'iliac_artery_right'],\n",
      "       [b'iliac_vena_left'],\n",
      "       [b'iliac_vena_right'],\n",
      "       [b'small_bowel'],\n",
      "       [b'duodenum'],\n",
      "       [b'colon'],\n",
      "       [b'rib_left_1'],\n",
      "       [b'rib_left_2'],\n",
      "       [b'rib_left_3'],\n",
      "       [b'rib_left_4'],\n",
      "       [b'rib_left_5'],\n",
      "       [b'rib_left_6'],\n",
      "       [b'rib_left_7'],\n",
      "       [b'rib_left_8'],\n",
      "       [b'rib_left_9'],\n",
      "       [b'rib_left_10'],\n",
      "       [b'rib_left_11'],\n",
      "       [b'rib_left_12'],\n",
      "       [b'rib_right_1'],\n",
      "       [b'rib_right_2'],\n",
      "       [b'rib_right_3'],\n",
      "       [b'rib_right_4'],\n",
      "       [b'rib_right_5'],\n",
      "       [b'rib_right_6'],\n",
      "       [b'rib_right_7'],\n",
      "       [b'rib_right_8'],\n",
      "       [b'rib_right_9'],\n",
      "       [b'rib_right_10'],\n",
      "       [b'rib_right_11'],\n",
      "       [b'rib_right_12'],\n",
      "       [b'humerus_left'],\n",
      "       [b'humerus_right'],\n",
      "       [b'scapula_left'],\n",
      "       [b'scapula_right'],\n",
      "       [b'clavicula_left'],\n",
      "       [b'clavicula_right'],\n",
      "       [b'femur_left'],\n",
      "       [b'femur_right'],\n",
      "       [b'hip_left'],\n",
      "       [b'hip_right'],\n",
      "       [b'sacrum'],\n",
      "       [b'face'],\n",
      "       [b'gluteus_maximus_left'],\n",
      "       [b'gluteus_maximus_right'],\n",
      "       [b'gluteus_medius_left'],\n",
      "       [b'gluteus_medius_right'],\n",
      "       [b'gluteus_minimus_left'],\n",
      "       [b'gluteus_minimus_right'],\n",
      "       [b'autochthon_left'],\n",
      "       [b'autochthon_right'],\n",
      "       [b'iliopsoas_left'],\n",
      "       [b'iliopsoas_right'],\n",
      "       [b'urinary_bladder']], dtype=object)>}\n",
      "WARNING:tensorflow:5 out of the last 9 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "5 out of the last 9 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "WARNING:tensorflow:6 out of the last 11 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "6 out of the last 11 calls to <function load_medical_data at 0x7f60801ca3a0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "\n",
      "Test with kwargs : {'slice_start': <tf.Tensor: shape=(), dtype=int32, numpy=8>, 'slice_end': <tf.Tensor: shape=(), dtype=int32, numpy=64>, 'labels': <tf.Tensor: shape=(104,), dtype=string, numpy=\n",
      "array([b'adrenal_gland_left', b'adrenal_gland_right', b'aorta',\n",
      "       b'autochthon_left', b'autochthon_right', b'brain',\n",
      "       b'clavicula_left', b'clavicula_right', b'colon', b'duodenum',\n",
      "       b'esophagus', b'face', b'femur_left', b'femur_right',\n",
      "       b'gallbladder', b'gluteus_maximus_left', b'gluteus_maximus_right',\n",
      "       b'gluteus_medius_left', b'gluteus_medius_right',\n",
      "       b'gluteus_minimus_left', b'gluteus_minimus_right',\n",
      "       b'heart_atrium_left', b'heart_atrium_right', b'heart_myocardium',\n",
      "       b'heart_ventricle_left', b'heart_ventricle_right', b'hip_left',\n",
      "       b'hip_right', b'humerus_left', b'humerus_right',\n",
      "       b'iliac_artery_left', b'iliac_artery_right', b'iliac_vena_left',\n",
      "       b'iliac_vena_right', b'iliopsoas_left', b'iliopsoas_right',\n",
      "       b'inferior_vena_cava', b'kidney_left', b'kidney_right', b'liver',\n",
      "       b'lung_lower_lobe_left', b'lung_lower_lobe_right',\n",
      "       b'lung_middle_lobe_right', b'lung_upper_lobe_left',\n",
      "       b'lung_upper_lobe_right', b'pancreas',\n",
      "       b'portal_vein_and_splenic_vein', b'pulmonary_artery',\n",
      "       b'rib_left_1', b'rib_left_10', b'rib_left_11', b'rib_left_12',\n",
      "       b'rib_left_2', b'rib_left_3', b'rib_left_4', b'rib_left_5',\n",
      "       b'rib_left_6', b'rib_left_7', b'rib_left_8', b'rib_left_9',\n",
      "       b'rib_right_1', b'rib_right_10', b'rib_right_11', b'rib_right_12',\n",
      "       b'rib_right_2', b'rib_right_3', b'rib_right_4', b'rib_right_5',\n",
      "       b'rib_right_6', b'rib_right_7', b'rib_right_8', b'rib_right_9',\n",
      "       b'sacrum', b'scapula_left', b'scapula_right', b'small_bowel',\n",
      "       b'spleen', b'stomach', b'trachea', b'urinary_bladder',\n",
      "       b'vertebrae_C1', b'vertebrae_C2', b'vertebrae_C3', b'vertebrae_C4',\n",
      "       b'vertebrae_C5', b'vertebrae_C6', b'vertebrae_C7', b'vertebrae_L1',\n",
      "       b'vertebrae_L2', b'vertebrae_L3', b'vertebrae_L4', b'vertebrae_L5',\n",
      "       b'vertebrae_T1', b'vertebrae_T10', b'vertebrae_T11',\n",
      "       b'vertebrae_T12', b'vertebrae_T2', b'vertebrae_T3',\n",
      "       b'vertebrae_T4', b'vertebrae_T5', b'vertebrae_T6', b'vertebrae_T7',\n",
      "       b'vertebrae_T8', b'vertebrae_T9'], dtype=object)>, 'mapping': <tf.Tensor: shape=(3, 2), dtype=string, numpy=\n",
      "array([[b'', b''],\n",
      "       [b'heart_atrium_left', b'heart_atrium_right'],\n",
      "       [b'aorta', b'']], dtype=object)>}\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n",
      "Value are equals !\n"
     ]
    }
   ],
   "source": [
    "importlib.reload(sparse_utils)\n",
    "importlib.reload(label_utils)\n",
    "importlib.reload(pre_processing)\n",
    "importlib.reload(data_io)\n",
    "\n",
    "def _to_dense(res):\n",
    "    sp, v = res\n",
    "    return tf.sparse.to_dense(sp), v\n",
    "\n",
    "sp_kwargs = {'dtype' : tf.uint8, 'use_sparse' : True} #tf.SparseTensorSpec(shape = (None, None, None, None), dtype = tf.uint8)\n",
    "\n",
    "tf_load_medical_data = tf.function(data_io.load_medical_data)\n",
    "\n",
    "test_kwargs = [\n",
    "    {},\n",
    "    {'slice_end' : tf.cast(6, tf.int32)},\n",
    "    {'slice_start' : tf.cast(3, tf.int32), 'slice_end' : tf.cast(6, tf.int32)},\n",
    "    {'slice_start' : tf.cast(3, tf.int32), 'slice_end' : tf.cast(6, tf.int32), 'slice_axis' : tf.cast(0, tf.int32)},\n",
    "    {\n",
    "        'slice_start' : tf.cast(8, tf.int32), 'slice_end' : tf.cast(64, tf.int32),\n",
    "        'labels' : labels, 'mapping' : [None] + labels\n",
    "    },\n",
    "    {\n",
    "        'slice_start' : tf.cast(8, tf.int32), 'slice_end' : tf.cast(64, tf.int32),\n",
    "        'labels' : labels, 'mapping' : TOTALSEGMENTATOR_LABELS\n",
    "    },\n",
    "    {\n",
    "        'slice_start' : tf.cast(8, tf.int32), 'slice_end' : tf.cast(64, tf.int32),\n",
    "        'labels' : tf.cast(labels, tf.string), 'mapping' : [None] + labels\n",
    "    },\n",
    "    {\n",
    "        'slice_start' : tf.cast(8, tf.int32), 'slice_end' : tf.cast(64, tf.int32),\n",
    "        'labels' : tf.cast(labels, tf.string), 'mapping' : label_utils.build_mapping(TOTALSEGMENTATOR_LABELS, 'tensor')\n",
    "    },\n",
    "    {\n",
    "        'slice_start' : tf.cast(8, tf.int32), 'slice_end' : tf.cast(64, tf.int32), 'labels' : tf.cast(labels, tf.string),\n",
    "        'mapping' : label_utils.build_mapping([None, ('heart_atrium_left', 'heart_atrium_right'), 'aorta'], 'tensor')\n",
    "    }\n",
    "][1:]\n",
    "for kwargs in test_kwargs:\n",
    "    print('\\nTest with kwargs : {}'.format(kwargs))\n",
    "    r1 = data_io.load_medical_data(path + '.nii.gz', dtype = tf.uint8, ** kwargs)\n",
    "\n",
    "    r13 = tf_load_medical_data(tf.cast(path + '.nii.gz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r14 = tf_load_medical_data(tf.cast(path + path_infos + '.npy', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    #r15 = tf_load_medical_data(tf.cast(path + '.npz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r16 = _to_dense(tf_load_medical_data(tf.cast(path + '-sparse.npz', tf.string), ** sp_kwargs, ** kwargs))\n",
    "    #r17 = tf_load_medical_data(tf.cast(path + path_infos + '.tensor', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r18 = _to_dense(tf_load_medical_data(tf.cast(path + path_infos + '.stensor', tf.string), ** sp_kwargs, ** kwargs))\n",
    "\n",
    "    print(is_equal(r1, r13)[1])\n",
    "    print(is_equal(r1, r14)[1])\n",
    "    #print(is_equal(r1, r15)[1])\n",
    "    print(is_equal(r1, r16)[1])\n",
    "    #print(is_equal(r1, r17)[1])\n",
    "    print(is_equal(r1, r18)[1])\n",
    "\n",
    "    r2 = data_io.load_medical_data(path + path_infos + '.npy', dtype = tf.uint8, ** kwargs)\n",
    "    #r3 = data_io.load_medical_data(path + '.npz', dtype = tf.uint8, ** kwargs)\n",
    "    r4 = _to_dense(data_io.load_medical_data(path + '-sparse.npz', ** sp_kwargs, ** kwargs))\n",
    "    #r5 = data_io.load_medical_data(path + path_infos + '.tensor', dtype = tf.uint8, ** kwargs)\n",
    "    r6 = _to_dense(data_io.load_medical_data(path + path_infos + '.stensor', ** sp_kwargs, ** kwargs))\n",
    "\n",
    "    print(is_equal(r1, r2)[1])\n",
    "    #print(is_equal(r1, r3)[1])\n",
    "    print(is_equal(r1, r4)[1])\n",
    "    #print(is_equal(r1, r5)[1])\n",
    "    print(is_equal(r1, r6)[1])\n",
    "\n",
    "    r7 = data_io.load_medical_data(tf.cast(path + '.nii.gz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r8 = data_io.load_medical_data(tf.cast(path + path_infos + '.npy', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    #r9 = data_io.load_medical_data(tf.cast(path + '.npz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r10 = _to_dense(data_io.load_medical_data(tf.cast(path + '-sparse.npz', tf.string), ** sp_kwargs, ** kwargs))\n",
    "    #r11 = data_io.load_medical_data(tf.cast(path + path_infos + '.tensor', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "    r12 = _to_dense(data_io.load_medical_data(tf.cast(path + path_infos + '.stensor', tf.string), ** sp_kwargs, ** kwargs))\n",
    "\n",
    "    print(is_equal(r1, r7)[1])\n",
    "    print(is_equal(r1, r8)[1])\n",
    "    #print(is_equal(r1, r9)[1])\n",
    "    print(is_equal(r1, r10)[1])\n",
    "    #print(is_equal(r1, r11)[1])\n",
    "    print(is_equal(r1, r12)[1])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d1f151c6-c26d-4775-a7bb-dab44d6e787f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.42 s ± 1.21 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "1.79 s ± 387 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "203 ms ± 575 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "79.9 ms ± 255 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "\n",
      "2.38 s ± 2.74 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "1.77 s ± 956 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "204 ms ± 774 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "79.7 ms ± 167 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "\n",
      "2.4 s ± 1.64 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "1.77 s ± 1.05 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "205 ms ± 770 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "66.5 ms ± 377 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit load_medical_data(path + '.nii.gz', dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(path + path_infos + '.npy', dtype = tf.uint8, ** kwargs)\n",
    "#%timeit load_medical_data(path + '.npz', dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(path + '-sparse.npz', ** sp_kwargs, ** kwargs)\n",
    "#%timeit load_medical_data(path + path_infos + '.tensor', dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(path + path_infos + '.stensor', ** sp_kwargs, ** kwargs)\n",
    "print()\n",
    "%timeit load_medical_data(tf.cast(path + '.nii.gz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(tf.cast(path + path_infos + '.npy', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "#%timeit load_medical_data(tf.cast(path + '.npz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(tf.cast(path + '-sparse.npz', tf.string), ** sp_kwargs, ** kwargs)\n",
    "#%timeit load_medical_data(tf.cast(path + path_infos + '.tensor', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit load_medical_data(tf.cast(path + path_infos + '.stensor', tf.string), ** sp_kwargs, ** kwargs)\n",
    "print()\n",
    "%timeit tf_load_medical_data(tf.cast(path + '.nii.gz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit tf_load_medical_data(tf.cast(path + path_infos + '.npy', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "#%timeit tf_load_medical_data(tf.cast(path + '.npz', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit tf_load_medical_data(tf.cast(path + '-sparse.npz', tf.string), ** sp_kwargs, ** kwargs)\n",
    "#%timeit tf_load_medical_data(tf.cast(path + path_infos + '.tensor', tf.string), dtype = tf.uint8, ** kwargs)\n",
    "%timeit tf_load_medical_data(tf.cast(path + path_infos + '.stensor', tf.string), ** sp_kwargs, ** kwargs)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "37b79790-6d83-43c1-9340-796df816d8ef",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor([b'a' b'b' b'c'], shape=(3,), dtype=string)\n",
      "tf.Tensor(\n",
      "[[b'' b'']\n",
      " [b'a' b'b']\n",
      " [b'c' b'']\n",
      " [b'd' b'']], shape=(4, 2), dtype=string)\n",
      "tf.Tensor(\n",
      "[[False False False]\n",
      " [ True  True False]\n",
      " [False False  True]\n",
      " [False False False]], shape=(4, 3), dtype=bool)\n"
     ]
    }
   ],
   "source": [
    "import importlib\n",
    "import tensorflow as tf\n",
    "\n",
    "from utils.med_utils import label_utils\n",
    "\n",
    "importlib.reload(label_utils)\n",
    "\n",
    "labels  = tf.cast(['a', 'b', 'c'], tf.string)\n",
    "mapping = [None, ('a', 'b'), 'c', 'd']\n",
    "\n",
    "matrix = label_utils.build_mapping(mapping, output_format = 'tf')\n",
    "print(labels)\n",
    "print(matrix)\n",
    "match = labels[tf.newaxis, tf.newaxis, :] == matrix[:, :, tf.newaxis]\n",
    "print(tf.reduce_any(match, axis = 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d1c26c4f-493f-419b-afc7-8f7a0a548b55",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}