Remove deprecated AsChannelFirst and AddChannel (#1527)

### Description After Project-MONAI/MONAI#7019, `AsChannelFirst` and `AddChannel` have been deprecated, remove the usage in the tutorial. ### Checks  - [ ] Avoid including large-size files in the PR. - [ ] Clean up long text outputs from code cells in the notebook. - [ ] For security purposes, please check the contents and remove any sensitive info such as user names and private key. - [ ] Ensure (1) hyperlinks and markdown anchors are working (2) use relative paths for tutorial repo files (3) put figure and graphs in the `./figure` folder - [ ] Notebook runs automatically `./runner.sh -t <path to .ipynb file>` --------- Signed-off-by: KumoLiu <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Project-MONAI · Sep 25, 2023 · 19b42c9 · 19b42c9
1 parent 0b2638b
commit 19b42c9
Show file tree

Hide file tree

Showing 34 changed files with 651 additions and 610 deletions.
diff --git a/acceleration/distributed_training/unet_evaluation_ddp.py b/acceleration/distributed_training/unet_evaluation_ddp.py
@@ -62,7 +62,7 @@
 from monai.data import DataLoader, Dataset, create_test_image_3d, DistributedSampler, decollate_batch
 from monai.inferers import sliding_window_inference
 from monai.metrics import DiceMetric
-from monai.transforms import Activations, AsChannelFirstd, AsDiscrete, Compose, LoadImaged, ScaleIntensityd
+from monai.transforms import Activations, EnsureChannelFirstd, AsDiscrete, Compose, LoadImaged, ScaleIntensityd
 
 
 def evaluate(args):
@@ -90,7 +90,7 @@ def evaluate(args):
     val_transforms = Compose(
         [
             LoadImaged(keys=["img", "seg"]),
-            AsChannelFirstd(keys=["img", "seg"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["img", "seg"], channel_dim=-1),
             ScaleIntensityd(keys="img"),
         ]
     )

diff --git a/acceleration/distributed_training/unet_evaluation_horovod.py b/acceleration/distributed_training/unet_evaluation_horovod.py
@@ -56,7 +56,15 @@
 from monai.data import DataLoader, Dataset, create_test_image_3d, decollate_batch
 from monai.inferers import sliding_window_inference
 from monai.metrics import DiceMetric
-from monai.transforms import Activations, AsChannelFirstd, AsDiscrete, Compose, LoadImaged, ScaleIntensityd, EnsureType
+from monai.transforms import (
+    Activations,
+    EnsureChannelFirstd,
+    AsDiscrete,
+    Compose,
+    LoadImaged,
+    ScaleIntensityd,
+    EnsureType,
+)
 
 
 def evaluate(args):
@@ -86,7 +94,7 @@ def evaluate(args):
     val_transforms = Compose(
         [
             LoadImaged(keys=["img", "seg"]),
-            AsChannelFirstd(keys=["img", "seg"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["img", "seg"], channel_dim=-1),
             ScaleIntensityd(keys="img"),
         ]
     )

diff --git a/acceleration/distributed_training/unet_evaluation_workflows.py b/acceleration/distributed_training/unet_evaluation_workflows.py
@@ -73,7 +73,7 @@
 from monai.inferers import SlidingWindowInferer
 from monai.transforms import (
     Activationsd,
-    AsChannelFirstd,
+    EnsureChannelFirstd,
     AsDiscreted,
     Compose,
     KeepLargestConnectedComponentd,
@@ -110,7 +110,7 @@ def evaluate(args):
     val_transforms = Compose(
         [
             LoadImaged(keys=["image", "label"]),
-            AsChannelFirstd(keys=["image", "label"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["image", "label"], channel_dim=-1),
             ScaleIntensityd(keys="image"),
         ]
     )

diff --git a/acceleration/distributed_training/unet_training_ddp.py b/acceleration/distributed_training/unet_training_ddp.py
@@ -61,7 +61,7 @@
 import monai
 from monai.data import DataLoader, Dataset, create_test_image_3d, DistributedSampler
 from monai.transforms import (
-    AsChannelFirstd,
+    EnsureChannelFirstd,
     Compose,
     LoadImaged,
     RandCropByPosNegLabeld,
@@ -99,7 +99,7 @@ def train(args):
     train_transforms = Compose(
         [
             LoadImaged(keys=["img", "seg"]),
-            AsChannelFirstd(keys=["img", "seg"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["img", "seg"], channel_dim=-1),
             ScaleIntensityd(keys="img"),
             RandCropByPosNegLabeld(
                 keys=["img", "seg"], label_key="seg", spatial_size=[96, 96, 96], pos=1, neg=1, num_samples=4

diff --git a/acceleration/distributed_training/unet_training_horovod.py b/acceleration/distributed_training/unet_training_horovod.py
@@ -60,7 +60,7 @@
 import monai
 from monai.data import DataLoader, Dataset, create_test_image_3d
 from monai.transforms import (
-    AsChannelFirstd,
+    EnsureChannelFirstd,
     Compose,
     LoadImaged,
     RandCropByPosNegLabeld,
@@ -99,7 +99,7 @@ def train(args):
     train_transforms = Compose(
         [
             LoadImaged(keys=["img", "seg"]),
-            AsChannelFirstd(keys=["img", "seg"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["img", "seg"], channel_dim=-1),
             ScaleIntensityd(keys="img"),
             RandCropByPosNegLabeld(
                 keys=["img", "seg"], label_key="seg", spatial_size=[96, 96, 96], pos=1, neg=1, num_samples=4

diff --git a/acceleration/distributed_training/unet_training_smartcache.py b/acceleration/distributed_training/unet_training_smartcache.py
@@ -67,7 +67,7 @@
 import monai
 from monai.data import DataLoader, SmartCacheDataset, create_test_image_3d, partition_dataset
 from monai.transforms import (
-    AsChannelFirstd,
+    EnsureChannelFirstd,
     Compose,
     LoadImaged,
     RandCropByPosNegLabeld,
@@ -105,7 +105,7 @@ def train(args):
     train_transforms = Compose(
         [
             LoadImaged(keys=["img", "seg"]),
-            AsChannelFirstd(keys=["img", "seg"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["img", "seg"], channel_dim=-1),
             ScaleIntensityd(keys="img"),
             RandCropByPosNegLabeld(
                 keys=["img", "seg"], label_key="seg", spatial_size=[96, 96, 96], pos=1, neg=1, num_samples=4

diff --git a/acceleration/distributed_training/unet_training_workflows.py b/acceleration/distributed_training/unet_training_workflows.py
@@ -73,7 +73,7 @@
 from monai.inferers import SimpleInferer
 from monai.transforms import (
     Activationsd,
-    AsChannelFirstd,
+    EnsureChannelFirstd,
     AsDiscreted,
     Compose,
     KeepLargestConnectedComponentd,
@@ -111,7 +111,7 @@ def train(args):
     train_transforms = Compose(
         [
             LoadImaged(keys=["image", "label"]),
-            AsChannelFirstd(keys=["image", "label"], channel_dim=-1),
+            EnsureChannelFirstd(keys=["image", "label"], channel_dim=-1),
             ScaleIntensityd(keys="image"),
             RandCropByPosNegLabeld(
                 keys=["image", "label"], label_key="label", spatial_size=[96, 96, 96], pos=1, neg=1, num_samples=4

diff --git a/acceleration/threadbuffer_performance.ipynb b/acceleration/threadbuffer_performance.ipynb
@@ -63,7 +63,7 @@
     "from monai.data import DataLoader, Dataset, ThreadBuffer, create_test_image_2d\n",
     "from monai.losses import Dice\n",
     "from monai.networks.nets import UNet\n",
-    "from monai.transforms import AddChanneld, Compose, MapTransform\n",
+    "from monai.transforms import EnsureChannelFirstd, Compose, MapTransform\n",
     "\n",
     "monai.utils.set_determinism(seed=0)\n",
     "\n",
@@ -99,7 +99,7 @@
     "trans = Compose(\n",
     "    [\n",
     "        RandomGenerator(keys=(\"im\", \"seg\")),\n",
-    "        AddChanneld(keys=(\"im\", \"seg\")),\n",
+    "        EnsureChannelFirstd(keys=(\"im\", \"seg\"), channel_dim=\"no_channel\"),\n",
     "    ]\n",
     ")\n",
     "\n",

diff --git a/automl/DiNTS/search_dints.py b/automl/DiNTS/search_dints.py
@@ -81,9 +81,6 @@
     Transform,
     AsDiscrete,
     AsDiscreted,
-    AddChannel,
-    AddChanneld,
-    AsChannelFirstd,
     CastToTyped,
     Compose,
     ConcatItemsd,

diff --git a/automl/DiNTS/train_dints.py b/automl/DiNTS/train_dints.py
@@ -77,9 +77,6 @@
     Transform,
     AsDiscrete,
     AsDiscreted,
-    AddChannel,
-    AddChanneld,
-    AsChannelFirstd,
     CastToTyped,
     Compose,
     ConcatItemsd,

diff --git a/deepgrow/ignite/inference.ipynb b/deepgrow/ignite/inference.ipynb
@@ -76,10 +76,10 @@
     "    SpatialCropGuidanced,\n",
     ")\n",
     "from monai.transforms import (\n",
-    "    AsChannelFirstd,\n",
+    "    EnsureChannelFirstd,\n",
     "    Spacingd,\n",
     "    LoadImaged,\n",
-    "    AddChanneld,\n",
+    "    Transposed,\n",
     "    NormalizeIntensityd,\n",
     "    EnsureTyped,\n",
     "    ToNumpyd,\n",
@@ -449,8 +449,8 @@
     "slice_idx = original_slice_idx = data[\"foreground\"][0][2]\n",
     "\n",
     "pre_transforms = [\n",
-    "    LoadImaged(keys=\"image\"),\n",
-    "    AsChannelFirstd(keys=\"image\"),\n",
+    "    LoadImaged(keys=\"image\", image_only=False),\n",
+    "    Transposed(keys=\"image\", indices=[2, 0, 1]),\n",
     "    Spacingd(keys=\"image\", pixdim=pixdim, mode=\"bilinear\"),\n",
     "    AddGuidanceFromPointsd(\n",
     "        ref_image=\"image\",\n",
@@ -460,7 +460,7 @@
     "        spatial_dims=dimensions,\n",
     "    ),\n",
     "    Fetch2DSliced(keys=\"image\", guidance=\"guidance\"),\n",
-    "    AddChanneld(keys=\"image\"),\n",
+    "    EnsureChannelFirstd(keys=\"image\", channel_dim=\"no_channel\"),\n",
     "    SpatialCropGuidanced(keys=\"image\", guidance=\"guidance\", spatial_size=roi_size),\n",
     "    Resized(keys=\"image\", spatial_size=roi_size, mode=\"area\"),\n",
     "    ResizeGuidanced(guidance=\"guidance\", ref_image=\"image\"),\n",
@@ -500,7 +500,7 @@
     "        slice_idx = 0\n",
     "    if tname == \"LoadImaged\":\n",
     "        original_image = data[\"image\"]\n",
-    "    if tname == \"AddChanneld\":\n",
+    "    if tname == \"EnsureChannelFirstd\":\n",
     "        original_image_slice = data[\"image\"]"
    ]
   },

diff --git a/deepgrow/ignite/inference_3d.ipynb b/deepgrow/ignite/inference_3d.ipynb
@@ -61,10 +61,10 @@
     "    SpatialCropGuidanced,\n",
     ")\n",
     "from monai.transforms import (\n",
-    "    AsChannelFirstd,\n",
+    "    EnsureChannelFirstd,\n",
     "    Spacingd,\n",
     "    LoadImaged,\n",
-    "    AddChanneld,\n",
+    "    Transposed,\n",
     "    NormalizeIntensityd,\n",
     "    EnsureTyped,\n",
     "    ToNumpyd,\n",
@@ -177,8 +177,8 @@
     "slice_idx = original_slice_idx = data[\"foreground\"][0][2]\n",
     "\n",
     "pre_transforms = [\n",
-    "    LoadImaged(keys=\"image\"),\n",
-    "    AsChannelFirstd(keys=\"image\"),\n",
+    "    LoadImaged(keys=\"image\", image_only=False),\n",
+    "    Transposed(keys=\"image\", indices=[2, 0, 1]),\n",
     "    Spacingd(keys=\"image\", pixdim=pixdim, mode=\"bilinear\"),\n",
     "    AddGuidanceFromPointsd(\n",
     "        ref_image=\"image\",\n",
@@ -187,7 +187,7 @@
     "        background=\"background\",\n",
     "        spatial_dims=dimensions,\n",
     "    ),\n",
-    "    AddChanneld(keys=\"image\"),\n",
+    "    EnsureChannelFirstd(keys=\"image\", channel_dim=\"no_channel\"),\n",
     "    SpatialCropGuidanced(keys=\"image\", guidance=\"guidance\", spatial_size=roi_size),\n",
     "    Resized(keys=\"image\", spatial_size=model_size, mode=\"area\"),\n",
     "    ResizeGuidanced(guidance=\"guidance\", ref_image=\"image\"),\n",
@@ -216,15 +216,15 @@
     "        image[:, :, slice_idx]\n",
     "        if tname in (\"LoadImaged\")\n",
     "        else image[slice_idx]\n",
-    "        if tname in (\"AsChannelFirstd\", \"Spacingd\", \"AddGuidanceFromPointsd\")\n",
+    "        if tname in (\"Transposed\", \"Spacingd\", \"AddGuidanceFromPointsd\")\n",
     "        else image[0][slice_idx]\n",
     "    )\n",
     "    label = None\n",
     "\n",
     "    show_image(image, label, guidance, slice_idx)\n",
     "    if tname == \"LoadImaged\":\n",
     "        original_image = data[\"image\"]\n",
-    "    if tname == \"AddChanneld\":\n",
+    "    if tname == \"EnsureChannelFirstd\":\n",
     "        original_image_slice = data[\"image\"]\n",
     "    if tname == \"SpatialCropGuidanced\":\n",
     "        spatial_image = data[\"image\"]\n",
@@ -367,7 +367,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.13"
+   "version": "3.8.10"
   },
   "vscode": {
    "interpreter": {

diff --git a/deepgrow/ignite/train.py b/deepgrow/ignite/train.py
@@ -47,7 +47,7 @@
 from monai.networks.nets import BasicUNet, UNet
 from monai.transforms import (
     Activationsd,
-    AddChanneld,
+    EnsureChannelFirstd,
     AsDiscreted,
     Compose,
     EnsureTyped,
@@ -91,7 +91,7 @@ def get_network(network, channels, dimensions):
 def get_pre_transforms(roi_size, model_size, dimensions):
     t = [
         LoadImaged(keys=("image", "label")),
-        AddChanneld(keys=("image", "label")),
+        EnsureChannelFirstd(keys=("image", "label"), channel_dim="no_channel"),
         SpatialCropForegroundd(keys=("image", "label"), source_key="label", spatial_size=roi_size),
         Resized(keys=("image", "label"), spatial_size=model_size, mode=("area", "nearest")),
         NormalizeIntensityd(keys="image", subtrahend=208.0, divisor=388.0),

diff --git a/deployment/Triton/models/mednist_class/1/model.py b/deployment/Triton/models/mednist_class/1/model.py
@@ -53,7 +53,7 @@
 from monai.transforms import Compose
 from monai.transforms import (
     Activations,
-    AddChannel,
+    EnsureChannelFirst,
     AsDiscrete,
     CropForeground,
     CastToType,
@@ -121,8 +121,8 @@ def initialize(self, args):
             [
                 LoadImage(reader="PILReader", image_only=True, dtype=np.float32),
                 ScaleIntensity(),
-                AddChannel(),
-                AddChannel(),
+                EnsureChannelFirst(channel_dim="no_channel"),
+                EnsureChannelFirst(channel_dim="no_channel"),
                 ToTensor(),
                 Lambda(func=lambda x: x.to(device=self.inference_device)),
             ]

diff --git a/deployment/Triton/models/monai_covid/1/model.py b/deployment/Triton/models/monai_covid/1/model.py
@@ -49,7 +49,7 @@
 from monai.transforms import Compose
 from monai.transforms import (
     Activations,
-    AddChannel,
+    EnsureChannelFirst,
     AsDiscrete,
     CropForeground,
     LoadImage,
@@ -110,11 +110,11 @@ def initialize(self, args):
         self.pre_transforms = Compose(
             [
                 LoadImage(reader="NibabelReader", image_only=True, dtype=np.float32),
-                AddChannel(),
+                EnsureChannelFirst(channel_dim="no_channel"),
                 ScaleIntensityRange(a_min=-1000, a_max=500, b_min=0.0, b_max=1.0, clip=True),
                 CropForeground(margin=5),
                 Resize([192, 192, 64], mode="area"),
-                AddChannel(),
+                EnsureChannelFirst(channel_dim="no_channel"),
                 ToTensor(),
                 Lambda(func=lambda x: x.to(device=self.inference_device)),
             ]

diff --git a/deployment/bentoml/mednist_classifier_bentoml.ipynb b/deployment/bentoml/mednist_classifier_bentoml.ipynb
@@ -379,7 +379,7 @@
     "import torch\n",
     "\n",
     "from monai.transforms import (\n",
-    "    AddChannel,\n",
+    "    EnsureChannelFirst,\n",
     "    Compose,\n",
     "    Transform,\n",
     "    ScaleIntensity,\n",
@@ -414,7 +414,7 @@
     "class MedNISTClassifier(bentoml.BentoService):\n",
     "    @cached_property\n",
     "    def transform(self):\n",
-    "        return Compose([LoadStreamPIL(\"L\"), AddChannel(), ScaleIntensity(), EnsureType()])\n",
+    "        return Compose([LoadStreamPIL(\"L\"), EnsureChannelFirst(channel_dim=\"no_channel\"), ScaleIntensity(), EnsureType()])\n",
     "\n",
     "    @bentoml.api(input=FileInput(), output=JsonOutput(), batch=True)\n",
     "    def predict(self, file_streams: List[BinaryIO]) -> List[str]:\n",

diff --git a/deployment/ray/mednist_classifier_ray.ipynb b/deployment/ray/mednist_classifier_ray.ipynb
@@ -106,7 +106,7 @@
     "from monai.apps import download_url\n",
     "from monai.config import print_config\n",
     "from monai.transforms import (\n",
-    "    AddChannel,\n",
+    "    EnsureChannelFirst,\n",
     "    Compose,\n",
     "    ScaleIntensity,\n",
     "    EnsureType,\n",
@@ -148,7 +148,7 @@
     "class MedNISTClassifier:\n",
     "    def __init__(self):\n",
     "        # create the transform for normalizing the image data\n",
-    "        self.transform = Compose([AddChannel(), ScaleIntensity(), EnsureType()])\n",
+    "        self.transform = Compose([EnsureChannelFirst(channel_dim=\"no_channel\"), ScaleIntensity(), EnsureType()])\n",
     "        # load the network on the CPU for simplicity and in eval mode\n",
     "        self.net = torch.jit.load(\"../bentoml/classifier.zip\", map_location=\"cpu\").eval()\n",
     "\n",
@@ -358,7 +358,7 @@
     "\n",
     "from monai.config import print_config\n",
     "from monai.transforms import (\n",
-    "    AddChannel,\n",
+    "    EnsureChannelFirst,\n",
     "    Compose,\n",
     "    ScaleIntensity,\n",
     "    EnsureType,\n",
@@ -371,7 +371,7 @@
     "@serve.deployment\n",
     "class MedNISTClassifier:\n",
     "    def __init__(self):\n",
-    "        self.transform = Compose([AddChannel(), ScaleIntensity(), EnsureType()])\n",
+    "        self.transform = Compose([EnsureChannelFirst(channel_dim=\"no_channel\"), ScaleIntensity(), EnsureType()])\n",
     "        self.net = torch.jit.load(\"../bentoml/classifier.zip\", map_location=\"cpu\").eval()\n",
     "\n",
     "    async def __call__(self, request):\n",