Language support

Code explanation is available for the following languages:

• C
• C++
• Go
• Java
• JavaScript
• Python
• TypeScript

When you reference a method, or use CodeLens on a method, watsonx Code Assistant supports certain code languages. Referencing a full file works for all languages. For more information, see Language support when you work with methods.

Using a chat command to explain code

You can use the /explain command in chat to explain code for a referenced class, file, function, or method in the active workspace.

Use this syntax:

/explain <code reference> [additional instructions]

• For <code reference>, type the @ symbol to see a list of classes, files, functions, and methods from your workspace. Use one class, file, function, or method reference at a time.

• The [additional instructions] are optional. Add instructions if you want specific details.

When you reference a file, the size limit is 50 KB. If you reach this limit, split the file into individual functions and reference each function. Or, split the file at 49 KB, taking care of function boundaries, and reference the file at each split. With either approach, you need to merge the results.

Using the CodeLens in the editor to explain code

In the IDE editor, the CodeLens shows a line of generative AI options that precedes code blocks and snippets.

Known issue If you are using the watsonx Code Assistant for Enterprise Java Applications extension, upon startup it might take a few seconds for the enhanced Java capabilities to be available from the CodeLens.

1. Click the Explain option that immediately precedes a code block to generate an explanation.

   In the following code example, the Explain | Document | Unit Test options immediately precede the protected void keywords.

   

2. The watsonx Code Assistant chat window opens, displays the /explain @<*item name*> command, runs the command, and displays the explanation.

Using the Explorer context menu to explain code

To generate an explanation from a context menu:

1. In the Explorer view, expand your application to the code that you want to generate an explanation for.

2. Right-click the code, click watsonx Code Assistant, then click Explain.

3. The watsonx Code Assistant chat displays the /explain command for the code that you selected. The following syntax is used in the command:

   /explain <code syntax>

4. Watsonx Code Assistant processes the request and in the chat displays the explanation for the code that you selected.

Explaining Java applications

Standard plan Application explanation is only available with the watsonx Code Assistant for Enterprise Java Applications extension.

Application explanation is only supported for Java applications that contain one or more classes that extend javax.servlet.http.HttpServlet. These classes need to implement any one of the following methods: doGet, doPost, or doPut.

For large applications, the explanation considers the 40 entry points with the highest cyclomatic complexity. If the application exceeds 40 entry points, the explanation contains the section The following methods were not considered: that provides details about which entry points were not considered.

To request and view an explanation for an application:

1. In your IDE, right-click any item in the hierarchy in the directory of the application that you want to explain, and then click Explain Application.

   In the Eclipse IDE, if you have a multimodule application with a peer multimodule root, the application explanation applies only to the selected module.

2. Watsonx Code Assistant scans the application to generate an overview and a list of main services with a description of functions for each method.

3. Click Save to retain a copy, or the explanation is discarded.

4. Click Open explanation and review, and then you can click Save application explanation to store in a local file.