Projects
Introduction
A project in PromptQL forms the backbone for building and managing your application. It contains a semantic collection of metadata objects that define the behavior, relationships, and permissions of your application. We sometimes refer to this as the supergraph. You can organize the files containing these metadata objects into subgraphs, with each subgraph representing a distinct data domain.
Projects are designed to support both local development and cloud deployment. During development, you work with a local
version of your project, which is linked to a cloud project through a
context file. This linkage enables seamless
development, testing, and deployment workflows. You can also define multiple contexts (e.g., staging) to manage
different environments, with unique configurations for environment variables and cloud resources.
Data domains
A data domain represents a distinct area of responsibility or focus within your project, typically aligned with a specific team or business function. These domains are managed as subgraphs, which are collections of metadata files that describe the relationships, permissions, and structure for the data within that domain.
Organizing your project into subgraphs provides several benefits:
- Team Ownership: Each team can focus on their own data domain without interfering with others, making collaboration simpler and reducing bottlenecks.
- Clear Boundaries: Subgraphs establish clear boundaries between domains, making it easier to define and enforce data access permissions and relationships.
- Scalability: By breaking your project into manageable pieces, you can scale your application incrementally as your organization and data requirements grow.
- Flexibility: Subgraphs can be independently updated and extended, allowing you to adapt your project to changing needs without disrupting the overall application.
This structure ensures that your project remains organized, collaborative, and adaptable, enabling you to build and maintain robust applications efficiently.
PromptQL also supports multi-repository setups, giving teams greater autonomy in their development process. With this setup:
- Each team can maintain their metadata in a separate repository while still contributing to the shared semantic metadata layer.
- Teams can develop, test, and deploy their subgraphs independently, enabling faster iteration and minimizing dependencies on other teams.
- The query engine integrates these subgraphs into a single metadata layer, ensuring that the overall application remains consistent and cohesive.
This approach is ideal for large organizations where multiple teams work on distinct data domains but need to collaborate through a unified API. For more information, check out this section of the docs.