Topics

Topics provide a structured and organized approach to reviewing Change Requests, helping reviewers navigate complex changes by grouping them logically. Instead of listing changes by file or module, Baz uses AI to categorize changes by flow, feature, or topic, creating a concise "table of contents" that makes reviews more manageable.

Select a topic to start a review

How It Works

Baz's AI analyzes the code changes in a Change Request, using both the code diff and the user-provided description to generate relevant topics. Each topic represents a distinct aspect of the change, which may span across multiple files and modules. The goal is to create a minimal set of groups, where each group helps the reviewer understand how different parts of the change interact in the overall app flow.

This method provides several benefits:

• Structured Review: Changes are grouped in meaningful ways, making it easier for reviewers to focus on related changes together rather than jumping between files.

• Flow-Oriented: Topics are based on flows, features, or usage rather than technical details, helping reviewers understand the broader context of the change.

• Prioritized Focus: Each topic is ranked by importance within the Change Request, ensuring that reviewers can focus on the most significant aspects first.

AI Prompt for Grouping

To generate these topics, Baz uses an AI prompt with the following rules:

• Minimize Groups: The AI limits the number of topics to between 1-5 to avoid overwhelming the reviewer.

• Cross-File Grouping: A group can contain changes from multiple files and modules, ensuring that all related changes are considered together.

• Flow/Feature-Based Grouping: The AI guesses which flows, features, or topics a technical change serves, ensuring that even minor technical changes are placed into the right context.

• Ranked by Change Importance: Each topic is assigned an importance score (1 to 5), so reviewers can prioritize changes that are most relevant to the Change Request.