A brief introduction to the Java spi mechanism

Recently I have seen that some of the company's frameworks and some of the open source frameworks I have seen before use java's spi mechanism for service discovery and access.

So briefly summarize the ideas of java spi mechanism.

There are often many different implementation solutions for each abstract module in our system, such as the solution for the log module, xmlThe solution for the parsing module, the jdbc module, etc. In object-oriented design, we generally recommend that modules be programmed based on interface, and implementation classes should not be hard-coded between modules. Once a specific implementation class is involved in the code, it violates the principle of pluggability. If an implementation needs to be replaced, the code needs to be modified.

In order to realize that the module can be assembled without dynamically specifying it in the program, a service discovery mechanism is needed. Java SPI provides such a mechanism: a mechanism to find service implementations for a certain interface. It is somewhat similar to the idea of ​​​​IOC, which is to move the control of assembly outside the program. This mechanism is especially important in modular design.

The specific agreement of java spi is as follows:

When the service provider provides an implementation of the service interface, in the META-INF/services of the jar package A file named after the service interface is also created in the / directory. This file contains the specific implementation class that implements the service interface. When an external program assembles this module, it can find the specific implementation class name through the configuration file in the jar package META-INF/services/, and load the instantiation to complete the module injection.

Based on such a convention, the implementation class of the service interface can be found very well without the need to formulate it in the code.

jdk provides a tool class for service implementation lookup: java.util.ServiceLoader

Example

1.common-logging

The earliest log facade interface provided by apache. Only interface, no implementation. The specific solution is implemented by each provider. It is found that the log provider scans the META-INF/services/org.apache.commons.logging.LogFactory configuration file and reads the file. Find the content of the log mentioning the industrial and commercial implementation class. As long as our log implementation includes this file and formulates the implementation class of LogFactory factory interface in the file.

2.jdbc

Before jdbc4.0, developers still needed to be based on Class.forName("xxx") to load driver, jdbc4 also uses the spi mechanism to discover the driver provider, you can use the META-INF/services/java.sql.Driver file There is a way to specify the implementation class to expose the driver provider.

3. Write a simple example by yourself

Suppose there is a contentsearch system, which is divided into two modules: display and search. Display and search are based on interface programming. The implementation of search may be based on file system search, or it may be based on database search. The example code is as follows

Search.java: Search interface

package search;  
  
import java.util.List;  
  
import definition.Doc;  
  
public interface Search {  
    List<Doc> search(String keyword);  
}

FileSearch.java: File system search implementation

package search;  
  
import java.util.List;  
  
import definition.Doc;  
  
public class FileSearch implements Search {  
  
    @Override  
    public List<Doc> search(String keyword) {  
        System.out.println("now use file system search. keyword:" + keyword);  
        return null;  
    }  
  
}

DatabaseSearch.java

package search;  
  
import java.util.List;  
  
import definition.Doc;  
  
public class DatabaseSearch implements Search {  
  
    @Override  
    public List<Doc> search(String keyword) {  
        System.out.println("now use database search. keyword:" + keyword);  
        return null;  
    }  
  
}

SearchTest.java

package search;  
  
import java.util.Iterator;  
import java.util.ServiceLoader;  
  
public class SearchTest {  
  
    public static void main(String[] args) {  
        ServiceLoader<Search> s = ServiceLoader.load(Search.class);  
        Iterator<Search> searchs = s.iterator();  
        if (searchs.hasNext()) {  
            Search curSearch = searchs.next();  
            curSearch.search("test");  
        }  
    }  
}

Finally created in the META-INF/services/search.Search file.

When the search.Search file content is "search.FileSearch", the program output is:

now use file system search. keyword:test

When the search.Search file content is "search.DatabaseSearch", the program output is:

now use database search. keyword:test

It can be seen that there is no code related to the specific implementation in SearchTest, but is based on the spi mechanism to find the implementation of the service

The above is the detailed content of A brief introduction to the Java spi mechanism. For more information, please follow other related articles on the PHP Chinese website!