Develop a Graph program using MaxCompute Studio - MaxCompute

MaxCompute Studio lets you write, debug, package, and run Graph programs directly from IntelliJ IDEA. This guide walks you through the full development cycle.

Prerequisites

Before you begin, ensure that you have:

Connected MaxCompute Studio to your project. For details, see Manage project connections.
Created a MaxCompute Java module. For details, see Create a MaxCompute Java module.

Write a Graph program

Create the Java classes

A Graph program requires two class types working together: a GraphLoader to load the graph from an input table, and a Vertex to define the computation each node performs. You must implement both.

Class type	Role	Configured by
GraphLoader	Reads the input table and constructs the graph	`setGraphLoaderClass` method of `GraphJob`
Vertex	Defines per-node computation; has `id`, `value`, `halted`, and `edges` properties	`setVertexClass` method of `GraphJob`

The following steps use BaseLoadingVertexResolver (GraphLoader) and SSSP (Vertex) as examples.

In the left-side navigation pane of the Project tab, choose src > main > java, right-click java, and choose New > MaxCompute Java.
Enter a Name and select GraphLoader or Vertex as the class type, then press Enter.
If no package exists yet, enter packagename.classname. The system creates the package automatically.
Implement the Graph logic in the editor. For complete BaseLoadingVertexResolver and SSSP implementations, see SSSP.

Configure Maven dependencies

Add the following dependencies to pom.xml:

<dependency>
    <groupId>com.aliyun.odps</groupId>
    <artifactId>odps-sdk-core</artifactId>
    <version>0.48.0-public</version>
</dependency>
<dependency>
    <groupId>com.aliyun.odps</groupId>
    <artifactId>odps-sdk-graph</artifactId>
    <version>0.48.0-public</version>
</dependency>

<!-- Required for local testing -->
<dependency>
    <groupId>com.aliyun.odps</groupId>
    <artifactId>odps-graph-local</artifactId>
    <version>0.48.0-public</version>
</dependency>

After saving, verify that all three dependencies load successfully in the Maven pane.

Debug the Graph program

Run the Graph program locally to verify it produces the expected results.

Right-click the Java file and select Run.
In the Run/Debug Configurations dialog box, set the following parameters:
Parameter Description
MaxCompute project The project in which to run the Graph program
Download Record limit Maximum number of records to download. Default value: 100
Click OK. The Graph program runs in local execution mode, reading data from the specified table in the warehouse directory. View the output in the console. For details about the warehouse directory, see Overview.

Parameter	Description
MaxCompute project	The project in which to run the Graph program
Download Record limit	Maximum number of records to download. Default value: 100

Package and upload the Graph program

Package the Graph program into a JAR file and upload it to your MaxCompute project as a resource. For details, see Package, upload, and register a Java program.

Run the Graph program

In the left-side navigation pane of IntelliJ IDEA, click Project Explorer.
Right-click the name of your MaxCompute project and select Open in Console.
In the Console section, run the following command:
```
jar -libjars xxx.jar -classpath /Users/home/xxx.jar com.aliyun.odps.graph.examples.PageRank pagerank_in pagerank_out;
```
Replace the placeholders based on your Graph program:
Placeholder Description
PageRank The main class defined in your Graph program
pagerank_in The input table containing the graph data
pagerank_out The output table for results
For additional command options, see Submit a MapReduce job.

Placeholder	Description
`PageRank`	The main class defined in your Graph program
`pagerank_in`	The input table containing the graph data
`pagerank_out`	The output table for results

What's next

Explore more Graph algorithm examples: SSSP
Learn about packaging and registering resources: Package, upload, and register a Java program