Platform For AI:Call a service over a VPC endpoint

After a model is deployed as a service in EAS, a VPC endpoint is automatically generated for the service. You can perform the following steps to obtain the VPC endpoint of a service: On the Elastic Algorithm Service page, find the service that you want to call and click Invoke Intro in the Service Method column. In the Invoke Intro dialog box, click the VPC Invoke tab and check the values of the Access address and Token parameters. Based on the call information, you can perform a call test. The following code provides an example on a call test:

$ curl http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/heart_predict_online -H 'Authorization: ZGI1YWYxNmQwYjMzMDM1YTNlMmFmNmEzYjIzZTVlZGQ0NDJhYTRm****' -d '[{"sex":0,"cp":0,"fbs":0,"restecg":0,"exang":0,"slop":0,"thal":0,"age":0,"trestbps":0,"chol":0,"thalach":0,"oldpeak":0,"ca":0}]'

After the call test succeeds, you can call the service. To facilitate service calls, EAS allows you to use one of the following methods to call a service over a VPC endpoint:

• Use the official SDK for Python

  EAS encapsulates the call logic and provides the official SDK for Python.

• Use the official SDK for Java

  EAS encapsulates the call logic and provides the official SDK for Java.

• Use custom call logic

  We recommend that you use the preceding two official SDKs to call services. This reduces the amount of time consumed to define call logic. If you need to use SDKs in other languages or use custom call logic, you can follow the demo in the "Use custom call logic" section of this topic. To implement custom call logic, you must construct service requests based on specific frameworks. For more information, see Construct a request for a TensorFlow service.

To use the official SDK for Python to call the service, perform the following steps:

1. Install the SDK.

   pip install -U eas-prediction --user

   For more information about how to use the SDK for Python, visit GitHub.
2. Compile a call program.
   In this example, a program that uses strings as input and output is used. For information about sample programs with other input and output formats, such as TensorFlow and PyTorch programs, visit GitHub.

   #!/usr/bin/env python
   from eas_prediction import PredictClient
   from eas_prediction import StringRequest
   if __name__ == '__main__':
       # Set the input of the client = PredictClient() function to the VPC endpoint that is used to call the service. 
       client = PredictClient('http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com','heart_predict_online')
       # Use the token that is obtained on the VPC Invoke tab. For more information, see the "Call methods" section of this topic. 
       client.set_token('ZGI1YWYxNmQwYjMzMDM1YTNlMmFmNmEzYjIzZTVlZGQ0NDJhYTRm****')    
       client.init()
       # Construct the request based on the model that you want to use. In this example, the input and output are of the STRING type. 
       request = StringRequest('[{"sex":0,"cp":0,"fbs":0,"restecg":0,"exang":0,"slop":0,"thal":0,"age":0,"trestbps":0,"chol":0,"thalach":0,"oldpeak":0,"ca":0}]')    
       for x in range(0, 1):
           resp = client.predict(request)
           print(resp)

   The input of the client = PredictClient() function is the VPC endpoint that is used to call the service. For example, if the VPC endpoint is http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/heart_predict_online, the format that is used to call the PredictClient() function is client = PredictClient('http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com','heart_predict_online').
3. Run the call program.

   $ python heart_predict.py

   The heart_predict.py parameter specifies the name of the Python program. You can replace it with the actual program name. The following prediction result is returned after the call program is run:

   IT-C02YJ0V8JHD2:Desktop wowei$ python heart_predict.py
   [{"p_0":0.9941226679708888,"p_1":0.005877332029111252}]

To use the official SDK for Java to call the service, perform the following steps:

1. Add the required dependency.
   You must use Maven to manage projects when you define code that runs on a Java client. Therefore, you must add the client dependency eas-sdk to the pom.xml file. The latest version of the dependency is 2.0.1. The following code provides an example on how to add the dependency:

   <dependency>
     <groupId>com.aliyun.openservices.eas</groupId>
     <artifactId>eas-sdk</artifactId>
     <version>2.0.1</version>
   </dependency>

   For more information about how to use the SDK for Java, visit GitHub.
2. Compile a call program.
   In this example, a program that uses strings as input and output is used. For information about sample programs with other input and output formats, such as TensorFlow and PyTorch programs, visit GitHub.

   import com.aliyun.openservices.eas.predict.http.PredictClient;
   import com.aliyun.openservices.eas.predict.http.HttpConfig;
   public class Test_String {
       public static void main(String[] args) throws Exception{
           // Start and initialize the client. 
           PredictClient client = new PredictClient(new HttpConfig());
           // Use the token that is obtained on the VPC Invoke tab. For more information, see the "Call methods" section of this topic. 
           client.setToken("ZGI1YWYxNmQwYjMzMDM1YTNlMmFmNmEzYjIzZTVlZGQ0NDJhYTRm****");
           // Set the endpoint to the VPC endpoint that is used to call the service. 
           client.setEndpoint("http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com");
           // Set the service name. 
           client.setModelName("heart_predict_online");
           // Define the input string. Construct the request based on the model that you want to use. In this example, the input and output are of the STRING type. 
           String request = "[{\"sex\":0,\"cp\":0,\"fbs\":0,\"restecg\":0,\"exang\":0,\"slop\":0,\"thal\":0,\"age\":0,\"trestbps\":0,\"chol\":0,\"thalach\":0,\"oldpeak\":0,\"ca\":0}]";
           System.out.println(request);
           // Return a string by using EAS.  
           try {
               String response = client.predict(request);
               System.out.println(response);
           } catch(Exception e) {
               e.printStackTrace();
           }
           // Shut down the client. 
           client.shutdown();
           return;
       }
   }

   The input of the client.setEndpoint() and client.setModelName() functions is the VPC endpoint that is used to call the service. For example, if the VPC endpoint is http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/heart_predict_online, the format that is used to call the client.setEndpoint() function is client.setEndpoint("http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com"), and the format that is used to call the client.setModelName() function is client.setModelName("heart_predict_online").
3. Run the call program.
   The following result is returned after the call program is run:

   [{"p_0":0.9941226679708888,"p_1":0.005877332029111252}]

EAS supports custom call logic in Python, Java, and other languages. You must customize service requests based on specific frameworks. For more information, see Construct a request for a TensorFlow service. The following code provides an example on how to use custom call logic by sending HTTP requests:

import requests
# Use the VPC endpoint that is obtained on the VPC Invoke tab as the URL. For more information, see the "Call methods" section of this topic. 
url = 'http://166408185518****.vpc.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/heart_predict_online'
# Add the token that is obtained on the VPC Invoke tab to the header. For more information, see the "Call methods" section of this topic. 
headers = {"Authorization": 'ZGI1YWYxNmQwYjMzMDM1YTNlMmFmNmEzYjIzZTVlZGQ0NDJhYTRm****'}
# Construct the service request based on the data format required by the model that you want to use. In this example, the input and output are of the STRING type. 
data = '[{"sex":0,"cp":0,"fbs":0,"restecg":0,"exang":0,"slop":0,"thal":0,"age":0,"trestbps":0,"chol":0,"thalach":0,"oldpeak":0,"ca":0}]'
resp = requests.post(url, data=data, headers=headers)
print resp
print resp.content