Project-MONAI
diff --git a/‎docs/source/developing_with_sdk/core_concepts.md‎
Lines changed: 1 addition & 1 deletion b/‎docs/source/developing_with_sdk/core_concepts.md‎
Lines changed: 1 addition & 1 deletion
diff --git a/‎docs/source/developing_with_sdk/creating_application_class.md‎
Lines changed: 22 additions & 26 deletions b/‎docs/source/developing_with_sdk/creating_application_class.md‎
Lines changed: 22 additions & 26 deletions
diff --git a/‎docs/source/developing_with_sdk/creating_operator_classes.md‎
Lines changed: 17 additions & 34 deletions b/‎docs/source/developing_with_sdk/creating_operator_classes.md‎
Lines changed: 17 additions & 34 deletions
diff --git a/‎docs/source/developing_with_sdk/deploying_and_hosting_map.md‎
Lines changed: 3 additions & 3 deletions b/‎docs/source/developing_with_sdk/deploying_and_hosting_map.md‎
Lines changed: 3 additions & 3 deletions
diff --git a/‎docs/source/developing_with_sdk/designing_workflow.md‎
Lines changed: 5 additions & 1 deletion b/‎docs/source/developing_with_sdk/designing_workflow.md‎
Lines changed: 5 additions & 1 deletion
@@ -29,7 +29,7 @@ classDiagram
     Graph "1" --> "many" Operator : contains
 ```
 
-[Application](/modules/_autosummary/monai.deploy.core.Application) represents a workflow as a [Graph](/modules/_autosummary/monai.deploy.core.graphs.Graph) and the graph handles [Operator](/modules/_autosummary/monai.deploy.core.Operator)s which are computational tasks.
+[Application](/modules/_autosummary/monai.deploy.core.Application) represents a workflow as a [Graph](/modules/graphs) and the graph handles [Operator](/modules/_autosummary/monai.deploy.core.Operator)s which are computational tasks.
 
 To develop and deploy your MONAI App, you can follow the steps below (click a node to see the detail):
 
 
@@ -16,9 +16,6 @@ caption: |
 from monai.deploy.core import Application, env, resource
 
 
-@resource(cpu=1, gpu=1, memory="2Gi")
-# pip_packages can be a string that is a path(str) to requirements.txt file or a list of packages.
-@env(pip_packages=["scikit-image >= 0.17.2"])
 class App(Application):
     """This is a very basic application.
 
@@ -37,41 +34,34 @@ class App(Application):
         pass
 
 if __name__ == "__main__":
-    App(do_run=True)
+    App().run()
 
 ```
 
-### Decorators
-
-The resource requirements (such as `cpu`, `memory`, and `gpu`) for the application can be specified by using [@resource](/modules/_autosummary/monai.deploy.core.resource) decorator. This information is used only when the packaged app (Docker image) is executed.
-
-[@env](/modules/_autosummary/monai.deploy.core.env) accepts `pip_packages` parameter as a string that is a path to requirements.txt file or a list of packages to install. If `pip_packages` is specified, the definition will be aggregated with the package dependency list of other operators. The aggregated requirement definitions are stored as a "[requirements.txt](https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file)" file and it would be installed in [packaging time](/developing_with_sdk/executing_packaged_app_locally).
-
-
 ### compose() method
 
 In `compose()` method, operators are instantiated and connected through <a href="../modules/_autosummary/monai.deploy.core.Application.html#monai.deploy.core.Application.add_flow">self.add_flow()</a>.
 
-> add_flow(source_op, destination_op, io_map=None)
+> add_flow(source_op, destination_op, port_pairs )
 
-`io_map` is a dictionary of mapping from the source operator's label to the destination operator's label(s) and its type is `Dict[str, str|Set[str]]`.
+`port_pairs` is a sequence of string tuples mapping from the source operator's named output(s) to the destination operator's named input(s) and its type is `Set[Tuple[str, str]]`.
 
-We can skip specifying `io_map` if both the number of `source_op`'s outputs and the number of `destination_op`'s inputs are one.
-For example, if Operators named `task1` and `task2` has only one input and output (with the label `image`), `self.add_flow(task1, task2)` is same with `self.add_flow(task1, task2, {"image": "image"})` or `self.add_flow(task1, task2, {"image": {"image"}})`.
+We can skip specifying `Set[Tuple[str, str]]` if both the number of `source_op`'s outputs and the number of `destination_op`'s inputs are one.
+For example, if Operators named `task1` and `task2` has only one input and output (with the label `image`), `self.add_flow(task1, task2)` is same with `self.add_flow(task1, task2, {("image", "image")})` or `self.add_flow(task1, task2, {("image", "image")})`.
 
 ```python
     def compose(self):
         task1 = Task1()
         task2 = Task2()
 
         self.add_flow(task1, task2)
-        # self.add_flow(task1, task2, {"image": "image"})
-        # self.add_flow(task1, task2, {"image": {"image"}})
+        # self.add_flow(task1, task2, {("image", "image")})
+        # self.add_flow(task1, task2, {("image", "image")})
 ```
 
 > add_operator(operator)
 
-If an operator in the workflow graph is both a root node and a leaf node, you can execute <a href="../modules/_autosummary/monai.deploy.core.Application.html#monai.deploy.core.Application.add_flow">self.add_operator()</a> for adding the operator to the workflow graph of the application.
+If an operator in the workflow graph is both a root node and a leaf node, you can execute <a href="../modules/_autosummary/monai.deploy.core.Application.html#monai.deploy.core.Application.add_operator">self.add_operator()</a> for adding the operator to the workflow graph of the application.
 
 ```python
     def compose(self):
@@ -83,7 +73,7 @@ If an operator in the workflow graph is both a root node and a leaf node, you ca
 
 ```python
 if __name__ == "__main__":
-    App(do_run=True)
+    App().run()
 ```
 
 The above lines in `app.py` are needed to execute the application code by using `python` interpreter.
@@ -101,9 +91,15 @@ caption: |
 from app import App
 
 if __name__ == "__main__":
-    App(do_run=True)
+    App().run()
 ```
 
+## Package Dependency and Resource Requirements
+
+Unlike in previous versions where Python decorators are used to define the resource requirements (such as `cpu`, `memory`, and `gpu`) for the application, a YAML file is required to store such information with sections and attributes as defined in the [MONAI Application Package Specification](https://github.com/Project-MONAI/monai-deploy/blob/main/guidelines/monai-application-package.md). This file is only needed when the application is packaged into a MONAI Application Package container image. When the MAP is run, the executor is expected to parse the resource requirements embedded in the MAP to ensure they are met in the host system.
+
+Similarly, instead of using Python decorators, package dependencies of the application and all of its operators need to be aggregated and stored as a "[requirements.txt](https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file)" file, to be installed at [packaging time](/developing_with_sdk/packaging_app).
+
 ## Creating a Reusable Application
 
 Like <a href="./creating_operator_classes.html#creating-a-reusable-operator">Operator class</a>, an Application class can be implemented in a way that the common Application class can be reusable.
@@ -174,12 +170,12 @@ The above workflow can be expressed like below
         notifier = Notifier()
         writer = Writer()
 
-        self.add_flow(reader1, processor1, {"image": {"image1", "image2"},
-                                            "metadata": "metadata"})
-        self.add_flow(reader2, processor2, {"roi": "roi"})
-        self.add_flow(processor1, processor2, {"image": "image"})
-        self.add_flow(processor1, writer, {"image": "image"})
+        self.add_flow(reader1, processor1, {("image", "image1"), ("image", "image2"),
+                                            ("metadata", "metadata")})
+        self.add_flow(reader2, processor2, {("roi", "roi")})
+        self.add_flow(processor1, processor2, {("image", "image")})
+        self.add_flow(processor1, writer, {("image", "image")})
         self.add_flow(processor2, notifier)
         self.add_flow(processor2, processor3)
-        self.add_flow(processor3, writer, {"seg_image": "seg_image"})
+        self.add_flow(processor3, writer, {("seg_image", "seg_image")})
 ```
@@ -10,26 +10,24 @@ lineno-start: 1
 caption: |
     An Operator class definition example
 ---
-from typing import Any, Dict
+from pathlib import Path
+from monai.deploy.core import (ExecutionContext, Image, InputContext,
+                               Operator, OutputContext, OperatorSpec)
 
-import monai.deploy.core as md  # 'md' stands for MONAI Deploy (or can use 'core' instead)
-from monai.deploy.core import (DataPath, ExecutionContext, Image, InputContext,
-                               IOType, Operator, OutputContext)
 
-
-@md.input("image", DataPath, IOType.DISK)
-@md.input("mask", Image, IOType.IN_MEMORY)
-@md.output("image", Image, IOType.IN_MEMORY)
-@md.output("metadata", Dict[str, Any], IOType.IN_MEMORY)
-@md.env(pip_packages=["scikit-image>=0.17.2"])
 class MyOperator(Operator):
     """Sample Operator implementation."""
 
+    def setup(self, spec: OperatorSpec):
+        spec.input("image_path")
+        spec.output("image")
+        spec.output("metadata")
+
     def compute(self, op_input: InputContext, op_output: OutputContext, context: ExecutionContext):
         from skimage import filters, io
 
         # Get input image
-        image_path = op_input.get("image").path
+        image_path = Path(op_input.receive("image_path"))  # omitting validation for brevity
         if image_path.is_dir():
             image_path = next(image_path.glob("*.*"))  # take the first file
         input_image = io.imread(image_path)
@@ -44,24 +42,13 @@ class MyOperator(Operator):
         metadata = {"shape": input_image.shape, "dtype": input_image.dtype}
 
         # Set output
-        op_output.set(Image(output_image), "image")
-        op_output.set(metadata, "metadata")
+        op_output.emit(Image(output_image), "image")
+        op_output.emit(metadata, "metadata")
 ```
 
-### Decorators
-
-The input and output properties of the operator are specified by using [@input](/modules/_autosummary/monai.deploy.core.input) and [@output](/modules/_autosummary/monai.deploy.core.output) decorators.
+### setup() method
 
-[@input](/modules/_autosummary/monai.deploy.core.input) and [@output](/modules/_autosummary/monai.deploy.core.output) decorators accept (**`<Label>`**, [`<Data Type>`](/modules/domain_objects), [`<Storage Type>`](/modules/_autosummary/monai.deploy.core.IOType)) as parameters.
-
-If no `@input` or `@output` decorator is specified, the following properties are used by default:
-
-```python
-@md.input("", DataPath, IOType.DISK)  # if no @input decorator is specified.
-@md.output("", DataPath, IOType.DISK)  # if no @output decorator is specified.
-```
-
-[@env](/modules/_autosummary/monai.deploy.core.env) accepts `pip_packages` parameter as a string that is a path to requirements.txt file or a list of packages to install. If `pip_packages` is specified, the definition will be aggregated with the package dependency list of other operators and the application. The aggregated requirement definitions are stored as a "[requirements.txt](https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file)" file and it would be installed in [packaging time](/developing_with_sdk/executing_packaged_app_locally).
+In prior releases, the input and output properties of the operator are specified by using `@input` and `@output` decorators, but starting with release v0.6, the **<a href="../modules/_autosummary/monai.deploy.core.Operator.html#monai.deploy.core.Operator.setup">setup()</a>** method is used.
 
 ### compute() method
 
@@ -73,15 +60,11 @@ Please check the description of **<a href="../modules/_autosummary/monai.deploy.
 2. Application's input/output
 3. [Model](/modules/_autosummary/monai.deploy.core.models.Model)'s name/path/predictor
 
-Note that, if the operator is a leaf operator in the workflow graph and the operator output's `(<data type>, <storage type>) == (DataPath, DISK)`, you cannot call `op_output.set()` method.
-Instead, you can use the destination path available by `op_output.get().path` to store output data and the
-following logic is expected:
+Note that, if the operator is a leaf operator in the workflow graph and the operator output is file(s) written to the file system, the path needs to be set using the operator's constructor or a named input for path object or string. The leaf operator can also have in-memory output(s) without requiring other operator(s) as receiver, if the output is configured correctly like in the [GaussionOperator in Simple Image Processing App](https://github.com/Project-MONAI/monai-deploy-app-sdk/blob/main/examples/apps/simple_imaging_app/gaussian_operator.py)
 
-```python
-output_folder = op_output.get().path                 # get the output folder path
-output_path = output_folder / "final_output.png"     # get the output file path
-imsave(output_path, data_out)                        # save the output data
-```
+## Package dependencies
+
+The dependencies of the operator need to be captured in a "[requirements.txt](https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file)", instead of using the `@env` decorator as in earlier releases. The aggregated requirement definitions for an application are then store in a consolidated "[requirements.txt](https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file)" file, to be installed at [packaging time](/developing_with_sdk/packaging_app).
 
 ## Creating a Reusable Operator
 
 
@@ -1,6 +1,6 @@
 # Deploying and Hosting MONAI App Package
 
 The MONAI Application Package, MAP, built and packaged using MONAI Deploy App SDK, can be deployed in multiple ways, each with different levels of integration with a hosting platform.
-- A MAP is an OCI compliant container, albeit requiring <a href="https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/overview.html"> Nvidia Container Toolkit </a> as it is based on the <a href="https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch">Nvidia PyTorch Python container image </a>. The hosting platforms, via their own tooling, can inspect the MAP metadata and ensure dependencies are satisfied on launching the MAP container, for example, an MAP can used in a Argo workflow.
-- A MAP runs "natively" on a platform, where the platform specific adaptor or shim embedded in the MAP can understand the MAP and the platform APIs thus manages the life cycle of the application per platform requests. At the onset, the only native execution planned is on the <a href="https://github.com/Project-MONAI/monai-deploy/releases">MONAI Deploy</a>. The support for other platforms is on the horizon.
-- As envisioned for the long term, standardization of deployment package specification will be needed, so that compliant application packages shall work on all compliant platforms. One of the initiatives in this area is the <a href="https://oam.dev/">Open Appication Model</a>, though the App SDK project is not yet actively investigating the integration.
+- A MAP is an OCI compliant container, albeit requiring <a href="https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/overview.html"> Nvidia Container Toolkit </a> as it is by default based on <a href="https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara-holoscan/containers/holoscan">Nvidia container image with GPU support</a>. The hosting platforms, via their own tooling, can inspect the MAP metadata and ensure dependencies are satisfied on launching the MAP container, for example, an MAP can be used in a [Argo workflow](https://argoproj.github.io/argo-workflows/), in [Docker Compose](https://docs.docker.com/compose/), or simply run using [Docker](https://www.docker.com/).
+- A MAP runs "natively" on a platform, where the platform specific adaptor or utilities embedded in the MAP can understand the MAP and the platform APIs thus manage the life cycle of the application per platform requests. At the onset, the only native execution planned is on the <a href="https://github.com/Project-MONAI/monai-deploy/releases">MONAI Deploy</a>. The support for other platforms is on the horizon.
+- As envisioned for the long term, standardization of deployment package specification will be needed, so that compliant application packages shall work on all compliant platforms. One of the initiatives in this area is the <a href="https://oam.dev/">Open Application Model</a>, where an MAP can be used as the Application component.
@@ -84,6 +84,9 @@ Those domain object classes are controllable by the SDK so can be optimized furt
 
 The full list of Domain Object classes are available [here](/modules/domain_objects).
 
+:::{note}
+**`The functionality of mapping DataPath to the input and output of root and leaf operator(s) is absent starting with Release V0.6 of this App SDK, due to the move to rely on Holoscan SDK. It is planned to be re-introduced at a later time. For the time being, the application's input and output folders are passed to the root and leaf operators' constructor, as needed.`**
+
 Among those classes, [**DataPath**](/modules/_autosummary/monai.deploy.core.domain.DataPath) data type is special.
 
 - If an operator in the workflow graph is a root node (a node with no incoming edges) and its input's `(<data type>, <storage type>) == (DataPath, DISK)`, the input path given by the user [during the execution](/developing_with_sdk/executing_app_locally) would be mapped into the input of the operator.
@@ -102,9 +105,10 @@ In `A linear workflow` example above, if the workflow is processing the image da
   - **Output** (`output3`): a file path ([`DataPath`](/modules/_autosummary/monai.deploy.core.domain.DataPath), [`DISK`](/modules/_autosummary/monai.deploy.core.IOType))
 
 Note that `input1` and `output3` are [DataPath](/modules/_autosummary/monai.deploy.core.domain.DataPath) type with [IOType.DISK](/modules/_autosummary/monai.deploy.core.IOType). Those paths are mapped into input and output paths given by the user during the execution.
+:::
 
 :::{note}
-The above workflow graph is the same as a [Simple Image Processing App](/getting_started/tutorials/01_simple_app). Please look at the tutorial to see how such an application can be made with MONAI Deploy App SDK.
+The above workflow graph is the same as a [Simple Image Processing App](/getting_started/tutorials/simple_app). Please look at the tutorial to see how such an application can be made with MONAI Deploy App SDK.
 :::
 
 ## Complex Workflows