Store Data Structures

What it does

This skill provides guidance on structuring data within Zustand stores for optimal performance in LobeHub AI agents. It emphasizes separating data structures for list views (using arrays) and detail views (using maps), and utilizing types from @lobechat/types instead of database types. The skill defines distinct "Detail" and "List" types, with Detail types including comprehensive entity information and List types potentially incorporating computed UI fields.

When to use it

• When designing data structures for list pages within your LobeHub AI agent.
• When creating detail views that require access to complete entity information.
• When optimizing the performance of your agent by using appropriate data structures for caching and display.
• When needing to distinguish between lightweight list items and full detail entities.

Key capabilities

• Separation of List and Detail Structures: Uses arrays for lists and maps (Record<string, Detail>) for details.
• Type Definitions: Provides example type definitions for entities like AgentEvalBenchmark (Detail) and AgentEvalBenchmarkListItem (List).
• Usage of @lobechat/types: Specifies the use of types from @lobechat/types instead of database-specific types.
• Distinction between List and Detail Types: Allows for computed UI fields in list types.

Example prompts

• "Show me how to structure a data type for displaying a list of benchmarks."
• "What's the best way to cache detail pages using Zustand stores?"
• "Can you give me an example of a lightweight item type for a document list?"

Tips & gotchas

• Avoid mixing List and Detail types: They serve different purposes.
• Don’t use database types directly: Always use the types defined in @lobechat/types.
• Be mindful of "heavy" fields: Exclude them from list item types to improve performance.