IntegrationOS Storage

A versatile wrapper around multiple storage solutions, designed for single-tenant management of clients in the IntegrationOS project.

Purpose

IntegrationOS Storage provides a unified interface for managing different storage backends, enabling single-tenant configurations for clients. It supports seamless integration of new databases, making it adaptable to various storage needs.

Running the Storage Service

To run the storage service, use the following command:

$ cargo watch -x run -q | bunyan

By default, the service runs on port 5005, but this can be configured through environment variables.

Integrating a New Database

To add support for a new database, follow these steps:

1. Add the database type to the supported list:

#[derive(Debug, Clone, PartialEq, Eq, EnumString, AsRefStr)]
#[strum(serialize_all = "lowercase")]
pub enum DatabaseConnectionType {
    PostgreSql,
}

2. Create the necessary configuration and add it to the configuration loader:

#[derive(Envconfig, Clone, Serialize, Deserialize, PartialEq)]
pub struct DatabaseConnectionConfig {
    #[envconfig(from = "WORKER_THREADS")]
    pub worker_threads: Option<usize>,
    #[envconfig(from = "INTERNAL_SERVER_ADDRESS", default = "0.0.0.0:5005")]
    pub address: SocketAddr,
    #[envconfig(from = "ENVIRONMENT", default = "development")]
    pub environment: Environment,
    #[envconfig(nested = true)]
    pub postgres_config: PostgresConfig,
    #[envconfig(from = "DATABASE_CONNECTION_TYPE", default = "postgres")]
    pub database_connection_type: DatabaseConnectionType,
    #[envconfig(from = "CONNECTION_ID")]
    pub connection_id: String
}

3. Implement the Storage trait:

#[async_trait]
pub trait Storage: Send + Sync {
    async fn execute_raw(
        &self,
        query: &str,
    ) -> Result<Vec<HashMap<String, Value>>, IntegrationOSError>;

    async fn probe(&self) -> Result<bool, IntegrationOSError>;
}

Be mindful that implementing this trait usually requires creating serializers for your specific storage type.

4. Implement the Initializer trait:

#[async_trait]
pub trait Initializer {
    async fn init(config: &DatabaseConnectionConfig) -> Result<Server, anyhow::Error>;
}

After completing these steps, the compiler will guide you through the necessary changes to ensure the code compiles correctly. Remember to add the new tests to verify the functionality of the new storage type.

Running the Tests

To run the test suite for the storage service, execute:

cargo nextest run --all-features

This command will run all tests associated with the storage functionality, ensuring correct behavior across various scenarios.