GraphQL server framework in Rust
This framework lets you write type-safe, efficient GraphQL servers in Rust. We make heavy use of macros to cut down on boilerplate and the trait system to allow maximum flexibility.
This project is at 'proof of concept' stage. It can only handle minimal examples and some key components are missing. However, I believe the results are already promising - Rust and GraphQL are a great match!
In the future we should use Rusts emerging async IO systems to make extremely performant servers.
Example
Use the schema macro to specify the schema for your server (using IDL):
schema! {
schema {
query: Query,
}
type Query {
hero(episode: Episode): Character,
human(id : ID!): Human,
}
enum Episode {
NEWHOPE,
EMPIRE,
JEDI,
}
interface Character {
id: ID!,
name: String!,
friends: [Character],
appearsIn: [Episode]!,
}
type Human implements Character {
id: ID!,
name: String!,
friends: [Character],
appearsIn: [Episode]!,
homePlanet: String,
}
}You can see the output for this example use of the schema macro at schema.out.
The macro generates concrete and abstract versions of each item. The library user
must specify implementations for functions (e.g., hero in the above schema).
You can then use the generated types - enums are Rust enums, types are Rust
structs, etc.:
TODO these are equivalent to resolvers in the JS frameworks
struct MyServer;
impl Root for MyServer {
type Query = DbQuery;
fn query(&self) -> QlResult<DbQuery> {
Ok(DbQuery)
}
}
ImplRoot!(MyServer);
struct DbQuery;
impl AbstractQuery for DbQuery {
fn hero(&self, episode: Option<Episode>) -> QlResult<Option<Character>> {
match episode {
Some(Episode::JEDI) => {
// In real life, this would query the DB or execute business logic.
Ok(Some(Character {
id: Id("0".to_owned()),
name: "Luke".to_owned(),
friends: Some(vec![]),
appearsIn: vec![],
}))
}
_ => unimplemented!(),
}
}
fn human(&self, _id: Id) -> QlResult<Option<Human>> {
...
}
}If you don't want to use the generated representation for a certain item, you
can provide your own (perhaps using a HashMap of data, rather than fields).
You then implement the abstract view of the item (e.g., AbstractHuman for
Human) and override the relevant associated type (e.g., type Human = MyHuman;
in the implementations of Root and AbstractQuery, and anywhere else the type
is used):
struct MyHuman {
id: usize,
db_table: DbTablePtr,
}
ImplHuman!(MyHuman);
impl AbstractHuman for MyHuman {
fn resolve_field(&self, field: &query::Field) -> QlResult<result::Value> {
...
}
}TODO show main