Merge pull request #111104 from likebupt/add-web-service-input-output

add web service module doc

Author: PMEds28

articles/machine-learning/algorithm-module-reference/media/module/web-service-input-output-example.png

@@ -7,9 +7,9 @@ ms.service: machine-learning
 ms.subservice: core
 ms.topic: reference
 
-author: peterclu
-ms.author: peterlu
-ms.date: 02/22/2020
+author: likebupt
+ms.author: keli19
+ms.date: 04/13/2020
 ---
 # Algorithm & module reference for Azure Machine Learning designer (preview)
 
@@ -55,6 +55,10 @@ For help with choosing algorithms, see
 | Anomaly Detection | Build anomaly detection models. | [PCA-Based Anomaly Detection](pca-based-anomaly-detection.md) <br/> [Train Anomaly Detection Model](train-anomaly-detection-model.md) |
 
 
+## Web Service
+
+Learn about the [web service modules](web-service-input-output.md) which are necessary for real-time inference in Azure Machine Learning designer.
+
 ## Error messages
 
 Learn about the [error messages and exception codes](designer-error-codes.md) you might encounter using modules in Azure Machine Learning designer.

articles/machine-learning/algorithm-module-reference/module-reference.md

@@ -152,6 +152,10 @@
     href: pca-based-anomaly-detection.md
   - name: Train Anomaly Detection Model
     href: train-anomaly-detection-model.md
+- name: Web Service
+  items: 
+  - name: Web Service Input/Output
+    href: web-service-input-output.md
 - name: Module errors & troubleshooting
   href: designer-error-codes.md
 

articles/machine-learning/algorithm-module-reference/toc.yml

@@ -0,0 +1,47 @@
+---
+title:  "Web Service Input/Output"
+description: Learn about the web service modules in Azure Machine Learning designer (preview)
+titleSuffix: Azure Machine Learning
+services: machine-learning
+ms.service: machine-learning
+ms.subservice: core
+ms.topic: reference
+
+author: likebupt
+ms.author: keli19
+ms.date: 04/13/2020
+---
+# Web Service Input/Output
+
+This article describes **Web Service Input** module and **Web Service Output** module in Azure Machine Learning designer (preview).
+
+**Web Service Input** module can only connect to input port with type **DataFrameDirectory**. And **Web Service Output** module can only be connected from output port with type **DataFrameDirectory**. The two modules can be found in the module tree, under **Web Service** category. 
+
+**Web Service Input** module is used to indicate where user data enters the pipeline and **Web Service Output** module is used to indicate where user data is returned in a Real-time inference Pipeline.
+
+## How to use Web Service Input/Output
+
+- When you create a real-time inference pipeline from your training pipeline, **Web Service Input** and **Web Service Output** module will be automatically added to show where user data enters the pipeline and where data is returned. 
+
+    Learn more about [create a real-time inference pipeline](https://docs.microsoft.com/azure/machine-learning/tutorial-designer-automobile-price-deploy#create-a-real-time-inference-pipeline).
+
+    > [!NOTE]
+    > Automatically generating real-time inference pipeline is a rule-based best-effort process, there is no guarantee for the correctness. You can manually add or remove **Web Service Input/Output** modules to satisfy your requirements. Make sure there is at least one **Web Service Input** module and one **Web Service Output** module in your real-time inference pipeline. If you have multiple **Web Service Input** or **Web Service Output** modules, make sure they have unique names, which you can input the name in the right panel of the module.
+
+- You can also manually create a real-time inference pipeline by adding **Web Service Input** and **Web Service Output** modules to your unsubmitted pipeline.
+
+    > [!NOTE]
+    >  The pipeline type will be determined at the first time you submit it. So be sure to add **Web Service Input** and **Web Service Output** module before you submit for the first time if you want to create a real-time inference pipeline.
+
+   Below example shows how to manually create real-time inference pipeline from **Execute Python Script** module. 
+
+   ![Example](media/module/web-service-input-output-example.png)
+   
+   After you submit the pipeline and the run completes successfully, you will be able to deploy the real-time endpoint.
+   
+   > [!NOTE]
+   >  In the above example, **Enter Data Manually** provides the data schema for web service input and is necessary for deploying the real-time endpoint. Generally, you should always connect a module or dataset to the port which **Web Service Input** is connected to provide the data schema.
+   
+## Next steps
+Learn more about [deploy the real-time endpoint](https://docs.microsoft.com/azure/machine-learning/tutorial-designer-automobile-price-deploy#deploy-the-real-time-endpoint).
+See the [set of modules available](module-reference.md) to Azure Machine Learning.