Database

In Content v3, we have introduced a robust storage layer based on SQLite, which offers a powerful and efficient method for managing content. This marks a significant enhancement over the previous file-based storage system, which was constrained by performance and scalability limitations.

In Content v2, the system read and parsed content during the Nitro runtime, creating a cache file for each content file to store the parsed data. This method introduced considerable overhead to the website's runtime.

• The I/O time in production was substantial, as the module had to load all cache files to search through the content.
• Additionally, the lack of optimization and compression for the content resulted in a large bundle size, particularly problematic in edge environments.

Content management in Content v3 involves three key steps, which are designed to streamline the process and enhance performance.

Generating Database Dump

For each collection in your project, the module reads the content from the defined source and parses it into an Abstract Syntax Tree (AST). It creates a specific table for each collection based on its schema. The parsed content is then inserted into the corresponding table, ensuring that the data structure aligns with the defined schema for optimal querying. Everything is then saved in a dump file.

Restoring Dump on Cold Start

During runtime, when the application executes the first query to retrieve content, the module reads the generated dump from the previous step and restores it into the target database. This process is fast and optimized for each deployment mode and platform.

The module employs a special integrity check mechanism to ensure that the database is updated with the latest content. This same integrity check also prevents duplicate imports, maintaining the integrity and accuracy of the data stored.

WASM SQLite in Browser

For client-side navigation, the module utilizes a similar approach. When the application executes the first query for content, it downloads the generated dump from the server and initializes a local SQLite database within the browser. From that point onward, all queries are executed locally without needing to call the server, significantly improving the responsiveness of the application and providing a seamless user experience.

This architecture not only enhances performance but also allows for offline capabilities, enabling users to access content even without an active internet connection. The combination of server-side and client-side processing ensures that Nuxt Content v3 is both powerful and flexible, catering to a wide range of use cases and environments.