Address reviewer feedback: use MONAI utilities and simplify data setup

- Replace custom as_numpy() with monai.utils.convert_data_type
- Replace custom ensure_channel_first_2d() with EnsureChannelFirstd
- Remove MONAI_DATA_DIRECTORY/tempfile pattern, use direct path
- Add error handling for missing data directory and .h5 files
- Fix Part 4 slice mismatch for multi-slice MONAI output
- Remove cleanup cell (no longer needed)

Signed-off-by: Vidya Sagar <[email protected]>

Author: ViSaReVe

reconstruction/MRI_reconstruction/tutorials/01_kspace_basics_fastmri_knee.ipynb

@@ -65,8 +65,6 @@
    "outputs": [],
    "source": [
     "import os\n",
-    "import shutil\n",
-    "import tempfile\n",
     "\n",
     "import h5py\n",
     "import matplotlib.pyplot as plt\n",
@@ -83,39 +81,17 @@
     "from monai.transforms import (\n",
     "    CenterSpatialCropd,\n",
     "    Compose,\n",
+    "    EnsureChannelFirstd,\n",
     "    EnsureTyped,\n",
     "    Lambdad,\n",
     "    LoadImaged,\n",
     "    ThresholdIntensityd,\n",
     ")\n",
+    "from monai.utils.type_conversion import convert_data_type\n",
     "\n",
     "print_config()"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Setup data directory\n",
-    "\n",
-    "You can specify a directory with the `MONAI_DATA_DIRECTORY` environment variable.\n",
-    "This allows you to save results and reuse downloads.\n",
-    "If not specified a temporary directory will be used."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "directory = os.environ.get(\"MONAI_DATA_DIRECTORY\")\n",
-    "if directory is not None:\n",
-    "    os.makedirs(directory, exist_ok=True)\n",
-    "root_dir = tempfile.mkdtemp() if directory is None else directory\n",
-    "print(root_dir)"
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -132,7 +108,7 @@
     "**How to obtain the data:**\n",
     "1. Register at https://fastmri.org/dataset\n",
     "2. Download the knee single-coil validation set (`knee_singlecoil_val.tar.gz`)\n",
-    "3. Extract to a folder (e.g., `<MONAI_DATA_DIRECTORY>/knee_singlecoil_val/`)\n",
+    "3. Extract to a folder and update `data_path` in the cell below\n",
     "\n",
     "**Note:** This dataset is under a non-commercial license. You may not use it for commercial purposes.\n",
     "\n",
@@ -149,11 +125,24 @@
    "source": [
     "# Update this path to where your fastMRI knee single-coil data is stored.\n",
     "# You only need ONE .h5 file from the knee_singlecoil_val set.\n",
-    "data_path = os.path.join(root_dir, \"knee_singlecoil_val\")\n",
+    "data_path = os.path.join(\"YOUR_DIR_HERE\", \"knee_singlecoil_val\")\n",
+    "\n",
+    "if not os.path.isdir(data_path):\n",
+    "    raise FileNotFoundError(\n",
+    "        f\"Data directory not found: {data_path}\\n\"\n",
+    "        \"Please download the fastMRI knee single-coil validation set from \"\n",
+    "        \"https://fastmri.org/dataset and update the path above.\"\n",
+    "    )\n",
     "\n",
     "sample_files = sorted(\n",
     "    [f for f in os.listdir(data_path) if f.endswith(\".h5\")]\n",
     ")\n",
+    "if len(sample_files) == 0:\n",
+    "    raise FileNotFoundError(\n",
+    "        f\"No .h5 files found in {data_path}\\n\"\n",
+    "        \"Please place at least one .h5 file from the knee_singlecoil_val set.\"\n",
+    "    )\n",
+    "\n",
     "sample_file = os.path.join(data_path, sample_files[0])\n",
     "print(f\"Using sample file: {sample_file}\")\n",
     "print(f\"Total .h5 files found: {len(sample_files)}\")"
@@ -460,10 +449,9 @@
    "outputs": [],
    "source": [
     "def as_numpy(x):\n",
-    "    \"\"\"Convert torch tensor or array to numpy.\"\"\"\n",
-    "    if isinstance(x, torch.Tensor):\n",
-    "        return x.detach().cpu().numpy()\n",
-    "    return np.asarray(x)\n",
+    "    \"\"\"Convert torch tensor or array to numpy using MONAI utility.\"\"\"\n",
+    "    arr, *_ = convert_data_type(x, np.ndarray)\n",
+    "    return arr\n",
     "\n",
     "\n",
     "fig, axes = plt.subplots(2, 3, figsize=(18, 12))\n",
@@ -476,7 +464,10 @@
     "axes[0, 0].set_title(\"Random mask pattern\")\n",
     "axes[0, 0].axis(\"off\")\n",
     "\n",
-    "rand_recon = np.abs(as_numpy(random_result[\"kspace_masked_ifft\"]).squeeze())\n",
+    "rand_recon = as_numpy(random_result[\"kspace_masked_ifft\"])\n",
+    "if rand_recon.ndim == 3:\n",
+    "    rand_recon = rand_recon[rand_recon.shape[0] // 2]\n",
+    "rand_recon = np.abs(rand_recon.squeeze())\n",
     "axes[0, 1].imshow(rand_recon, cmap=\"gray\")\n",
     "axes[0, 1].set_title(\"Zero-filled recon (random 4x)\")\n",
     "axes[0, 1].axis(\"off\")\n",
@@ -494,9 +485,10 @@
     "axes[1, 0].set_title(\"Equispaced mask pattern\")\n",
     "axes[1, 0].axis(\"off\")\n",
     "\n",
-    "equi_recon = np.abs(\n",
-    "    as_numpy(equispaced_result[\"kspace_masked_ifft\"]).squeeze()\n",
-    ")\n",
+    "equi_recon = as_numpy(equispaced_result[\"kspace_masked_ifft\"])\n",
+    "if equi_recon.ndim == 3:\n",
+    "    equi_recon = equi_recon[equi_recon.shape[0] // 2]\n",
+    "equi_recon = np.abs(equi_recon.squeeze())\n",
     "axes[1, 1].imshow(equi_recon, cmap=\"gray\")\n",
     "axes[1, 1].set_title(\"Zero-filled recon (equispaced 4x)\")\n",
     "axes[1, 1].axis(\"off\")\n",
@@ -558,21 +550,6 @@
     "    return x\n",
     "\n",
     "\n",
-    "def to_numpy(x):\n",
-    "    \"\"\"Convert torch tensor to numpy safely.\"\"\"\n",
-    "    if isinstance(x, torch.Tensor):\n",
-    "        return x.detach().cpu().numpy()\n",
-    "    return np.asarray(x)\n",
-    "\n",
-    "\n",
-    "def ensure_channel_first_2d(x):\n",
-    "    \"\"\"Promote (H, W) to (1, H, W) for CenterSpatialCropd.\"\"\"\n",
-    "    x = np.asarray(x)\n",
-    "    if x.ndim == 2:\n",
-    "        return x[None, ...]\n",
-    "    return x\n",
-    "\n",
-    "\n",
     "def complex_to_magnitude(x):\n",
     "    \"\"\"Take magnitude if complex; otherwise return as-is.\"\"\"\n",
     "    x = np.asarray(x)\n",
@@ -609,11 +586,10 @@
     "\n",
     "# Stage 2: Shape fixes, crop to 320x320, normalize, clamp\n",
     "post_transform = Compose([\n",
-    "    Lambdad(\n",
-    "        keys=[\"kspace_masked_ifft\", \"reconstruction_esc\"],\n",
-    "        func=to_numpy,\n",
+    "    EnsureChannelFirstd(\n",
+    "        keys=[\"kspace_masked_ifft\"],\n",
+    "        channel_dim=\"no_channel\",\n",
     "    ),\n",
-    "    Lambdad(keys=[\"kspace_masked_ifft\"], func=ensure_channel_first_2d),\n",
     "    CenterSpatialCropd(\n",
     "        keys=[\"kspace_masked_ifft\", \"reconstruction_esc\"],\n",
     "        roi_size=(320, 320),\n",
@@ -677,10 +653,8 @@
    "source": [
     "def as_numpy_2d(x):\n",
     "    \"\"\"Return a 2D numpy image from (1, H, W) torch/numpy.\"\"\"\n",
-    "    if isinstance(x, torch.Tensor):\n",
-    "        x = x.detach().cpu().numpy()\n",
-    "    x = np.asarray(x)\n",
-    "    return x[0] if x.ndim == 3 else x\n",
+    "    arr, *_ = convert_data_type(x, np.ndarray)\n",
+    "    return arr[0] if arr.ndim == 3 else arr\n",
     "\n",
     "\n",
     "input_img = as_numpy_2d(result[\"kspace_masked_ifft\"])\n",
@@ -751,26 +725,6 @@
     "\n",
     "Data used in the preparation of this tutorial were obtained from the NYU fastMRI Initiative database (fastmri.med.nyu.edu). [Citation: Knoll et al., Radiol Artif Intell. 2020 Jan 29;2(1):e190007. doi: 10.1148/ryai.2020190007. (https://pubs.rsna.org/doi/10.1148/ryai.2020190007), and the arXiv paper: https://arxiv.org/abs/1811.08839] As such, NYU fastMRI investigators provided data but did not participate in analysis or writing of this tutorial. A listing of NYU fastMRI investigators, subject to updates, can be found at: fastmri.med.nyu.edu. The primary goal of fastMRI is to test whether machine learning can aid in the reconstruction of medical images."
    ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Cleanup data directory\n",
-    "\n",
-    "Remove directory if a temporary was used."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "if directory is None:\n",
-    "    shutil.rmtree(root_dir)\n",
-    "print(\"Cleanup complete!\")"
-   ]
   }
  ],
  "metadata": {