Accelerate Development with Amazon CodeCatalyst and Generative AI

Amazon CodeCatalyst integrates with Amazon Q Developer Agent to provide generative AI features that help team members complete tasks faster and focus on the most important parts of their work. Amazon Q Developer is an AI-powered generative conversational assistant that can help you understand, build, extend, and operate AWS applications. This tutorial explores how to use these features to streamline development workflows.

Before you begin, ensure you have the following: An AWS Builder ID or SSO identity to log in to CodeCatalyst. Generative AI features enabled in your space. Contributor or project administrator role in the project. At least one source repository configured for your existing project (unless creating a project with generative AI). Note that projects configured with the Jira Software extension cannot be used when assigning transactions to create initial solutions.

Collaborate with Amazon Q to create new projects or add components to existing ones. Provide project requirements in a chat-like interface, and Amazon Q will recommend blueprints, outlining any unmet requirements. Custom blueprints are also considered. You can then proceed with Amazon Q's suggestions, creating necessary resources like source repositories with code that meets your requirements. Amazon Q also creates transactions for unmet requirements. To create a project, navigate to your space in the CodeCatalyst console and select 'Create with Amazon Q'. Provide a short description of your project. Review Amazon Q's suggestions and choose to configure the blueprint or skip configuration. Enter a name for the project and its associated resources. Select 'Create project' to create the project using the blueprint. Amazon Q can also create transactions for unmet requirements, which can then be assigned to Amazon Q. The process for adding blueprints to existing projects is similar, starting with selecting 'Add with Amazon Q' in your project.

Pull requests are essential for reviewing and merging code changes. To help reviewers understand the changes, use the 'Write a description for me' feature in Amazon Q to create a summary of the changes included in the pull request. Amazon Q analyzes the differences between the source and target branches, summarizing the changes and their intent. This feature is not available for Git submodules or linked repositories. To test this feature, create a branch, make a simple code change, and then create a pull request. In the pull request description, select 'Write a description for me' to have Amazon Q generate a summary. Review and accept the suggested text, modifying it as needed.

When reviewing pull requests, users often leave multiple comments. To easily identify common themes and ensure all comments have been reviewed, use the 'Create comment summary' feature. Amazon Q analyzes all comments left on code changes in the pull request and creates a summary. Note that comment summaries are temporary and do not include comments on the entire pull request, only those on code differences in the revisions. This feature is not available for comments on code changes in Git submodules or linked repositories. To create a summary, navigate to the pull request and select 'Create comment summary'.

Transactions are used to track and manage work, but sometimes issues persist due to unclear ownership or the need for code research. Assign transactions to Amazon Q, which analyzes the title and description to create a draft solution. This helps focus resources on urgent issues while Amazon Q addresses others. Amazon Q is effective for simple transactions. Use clear and concise language to describe the desired action. When assigning a transaction to Amazon Q, you must confirm whether you want it to confirm each step, allow it to update workflow files, and allow it to suggest tasks. You must also specify the source repository to work in. After making these selections, Amazon Q will analyze the transaction and create a potential solution. It will create a branch, commit the code, and create a pull request to merge the branch with the default branch. Once complete, Amazon Q will move the transaction to 'In review'.

For complex or lengthy work, have Amazon Q analyze the transaction and suggest a breakdown into logical tasks. This allows for easier assignment of work and faster completion. To use this feature, create a transaction and select 'Suggest tasks'. Choose the source repository containing the code and start the task suggestion process. Review the suggested tasks, add or modify them as needed, and then create the tasks. These tasks can then be assigned to users, including Amazon Q.

After completing the tutorial, clean up any resources that are no longer needed. Unassign Amazon Q from any transactions it is no longer processing. Move all completed transactions to 'Done'. If the project is no longer needed, delete it.