9 releases

new 0.2.0 Apr 23, 2024
0.1.3 Dec 7, 2023
0.1.1 Nov 25, 2023
0.0.2 Nov 16, 2023
0.0.1-alpha.2 Jul 1, 2023

#196 in Database interfaces

Download history 35/week @ 2024-02-06 6/week @ 2024-02-13 37/week @ 2024-02-20 8/week @ 2024-02-27 1/week @ 2024-03-05 6/week @ 2024-03-12 4/week @ 2024-03-26 25/week @ 2024-04-02 11/week @ 2024-04-09 45/week @ 2024-04-16

85 downloads per month

Apache-2.0

2MB
1.5K SLoC

Atmosphere

🌍 Atmosphere

A lightweight sql framework for sustainable database reliant systems

SQLx Crate Book Docs

Overview

Atmosphere is a lightweight SQL framework designed for sustainable, database-reliant systems. It leverages Rust's powerful type and macro systems to derive SQL schemas from your rust struct definitions into an advanced trait system.

Key Features

  • SQL schema derivation from Rust structs.
  • Advanced trait system for query generation.
  • Automated database code testing with atmosphere::testing
  • ORM-like CRUD traits.
  • Code reusability across API layers using generics.
  • Compile-time introspection for type-safe schema generation.

Quickstart

use atmosphere::prelude::*;

#[derive(Schema)]
#[table(schema = "public", name = "user")]
struct User {
    #[sql(pk)]
    id: i32,
    name: String,
    #[sql(unique)]
    email: String,
}

#[derive(Schema)]
#[table(schema = "public", name = "post")]
struct Post {
    #[sql(pk)]
    id: i32,
    #[sql(fk -> User, rename = "author_id")]
    author: i32,
    #[sql(unique)]
    title: String,
}

#[tokio::main]
async fn main() -> sqlx::Result<()> {
    let pool = atmosphere::Pool::connect(&std::env::var("DATABASE_URL").unwrap()).await?;

    // CRUD operations

    let user = User { id: 0, name: "demo".to_owned(), email: "some@email.com".to_owned(), };

    user.save(&pool).await?;
    user.delete(&pool).await?;
    user.create(&pool).await?;

    // Field Queries

    assert_eq!(
        User::find(&0, &pool).await?,
        User::find_by_email("some@email.com", &pool).await?.unwrap()
    );

    // Relationships

    Post { id: 0, author: 0, title: "test".to_owned() }
        .save(&pool)
        .await?;

    Post::find_by_author(&0, &pool).await?;
    Post::delete_by_author(&0, &pool).await?;

    // Inter-Table Operations

    Post { id: 1, author: 0, title: "test1".to_owned() }
        .author(&pool).await?;

    user.posts(&pool).await?;
    user.delete_posts(&pool).await?;

    Ok(())
}

Atmosphere introspects the User and Post structs at compile time and generates const available type information about the schema into the Table trait.

Roadmap

Alpha Release

  • Advanced SQL Trait System (Table, Column, Relation ..)
  • Derive Macro (Schema)
  • SQL Field Attributes (#[sql(pk)], #[sql(fk -> Model)] and so on)
  • SQL Query Generation
  • Automated Integration Testing
  • Attribute Macro (#[table])

Beta Release

  • Transaction Support
  • Getting Database Agnostic
  • Hook into query execution using atmosphere::hooks
  • Errors using miette
  • Combined Primary and Foreign Keys

Stable Release

  • Postgres Composite Types
  • Support Custom Types
  • Runtime Inspection
  • Provide Application Utils
  • Stabilize Traits
  • Stabilize Query Generation
  • Table Lenses (subsets / views)
  • validator support
  • Auto Timestamping

Advanced

  • Virtual Columns using (#[virtual = "<sql>"])
  • Soft Delete Support
  • Attribute Macro (#[query])
  • Custom queries

Longterm

  • Generate GraphQL + HTTP Servers?
  • Generate Graphs

Functionalities

Given a struct Model that derives its atmosphere schema using #[derive(Schema)] and #[table]:

use atmosphere::prelude::*;

#[derive(Schema)]
#[table(schema = "public", name = "model")]
struct Model {
    #[sql(pk)]
    id: i32,
    a: String,
    #[sql(unique)]
    b: String,
}

Atmosphere is able to derive and generate the following queries:

CRUD

atmosphere::Create

  • Model::create

atmosphere::Read

  • Model::find
  • Model::find_all
  • Model::reload

atmosphere::Update

  • Model::update
  • Model::save

atmosphere::Delete

  • Model::delete
  • Model::delete_by

Field Queries

Each struct field that is marked with #[sql(unique)] becomes queryable.

In the above example b was marked as unique so atmosphere implements:

  • Model::find_by_b
  • Model::delete_by_b

Relationships & Inter-Table Queries

Given that a model contains fields are marked as a foreign key / point to another atmosphere::Table atmosphere – for example:

#[derive(Schema)]
#[table(schema = "public", name = "submodel")]
struct Submodel {
    #[sql(pk)]
    id: i32,
    #[sql(fk -> Model)]
    super: i32,
}

Atmosphere is able to generate utility queries to move across Table boundaries:

  • Model::submodels
  • Model::delete_submodels
  • Submodel::model
  • Submodel::find_by_model
  • Submodel::delete_by_model

Note that the function names contain model and submodel – they are derived from the respective struct names.

Contribution

We welcome contributions! Please see our contribution guidelines for more details.

License

Atmosphere is licensed under Apache 2.0.

Dependencies

~36–49MB
~900K SLoC