Given a GraphQL schema file, this tool will generate interfaces and classes to enable strongly typed querying from C# to a GraphQL API.
Example, given the following GraphQL schema
schema {
query: Query
mutation: Mutation
}
scalar Date
type Query {
] actors: [Person]
directors: [Person]
movie(id: Int!): Movie
movies: [Movie]
}
type Movie {
id: Int
name: String
genre: Int
released: Date
actors: [Person]
writers: [Person]
director: Person
rating: Float
}
type Person {
id: Int
dob: Date
actorIn: [Movie]
writerOf: [Movie]
directorOf: [Movie]
died: Date
age: Int
name: String
}
type Mutation {
addPerson(dob: Date!, name: String!): Person
}
Running dotnet run -- schema.graphql -m Date=DateTime
will generate the following
public interface RootQuery
{
[GqlFieldName("actors")]
List<Person> Actors();
[GqlFieldName("actors")]
List<TReturn> Actors<TReturn>(Expression<Func<Person, TReturn>> selection);
[GqlFieldName("directors")]
List<Person> Directors();
[GqlFieldName("directors")]
List<TReturn> Directors<TReturn>(Expression<Func<Person, TReturn>> selection);
[GqlFieldName("movie")]
Movie Movie();
[GqlFieldName("movie")]
TReturn Movie<TReturn>(int id, Expression<Func<Movie, TReturn>> selection);
[GqlFieldName("movies")]
List<Movie> Movies();
[GqlFieldName("movies")]
List<TReturn> Movies<TReturn>(Expression<Func<Movie, TReturn>> selection);
}
public interface Movie
{
[GqlFieldName("id")]
int Id { get; }
[GqlFieldName("name")]
string Name { get; }
[GqlFieldName("genre")]
int Genre { get; }
[GqlFieldName("released")]
DateTime Released { get; }
[GqlFieldName("actors")]
List<Person> Actors();
[GqlFieldName("actors")]
List<TReturn> Actors<TReturn>(Expression<Func<Person, TReturn>> selection);
[GqlFieldName("writers")]
List<Person> Writers();
[GqlFieldName("writers")]
List<TReturn> Writers<TReturn>(Expression<Func<Person, TReturn>> selection);
[GqlFieldName("director")]
Person Director();
[GqlFieldName("director")]
TReturn Director<TReturn>(Expression<Func<Person, TReturn>> selection);
[GqlFieldName("rating")]
double Rating { get; }
}
public interface Person
{
[GqlFieldName("id")]
int Id { get; }
[GqlFieldName("dob")]
DateTime Dob { get; }
[GqlFieldName("actorIn")]
List<Movie> ActorIn();
[GqlFieldName("actorIn")]
List<TReturn> ActorIn<TReturn>(Expression<Func<Movie, TReturn>> selection);
[GqlFieldName("writerOf")]
List<Movie> WriterOf();
[GqlFieldName("writerOf")]
List<TReturn> WriterOf<TReturn>(Expression<Func<Movie, TReturn>> selection);
[GqlFieldName("directorOf")]
List<Movie> DirectorOf();
[GqlFieldName("directorOf")]
List<TReturn> DirectorOf<TReturn>(Expression<Func<Movie, TReturn>> selection);
[GqlFieldName("died")]
DateTime Died { get; }
[GqlFieldName("age")]
int Age { get; }
[GqlFieldName("name")]
string Name { get; }
}
public interface Mutation
{
[GqlFieldName("addPerson")]
TReturn AddPerson<TReturn>(DateTime dob, string name, Expression<Func<Person, TReturn>> selection);
}
It also generates a GraphQLClient
class that will work for an unauthenticated API. You can modify that class to implement the authentication you may need. GraphQLClient
exposes 2 methods, async Task<GqlResult<TQuery>> QueryAsync<TQuery>(Expression<Func<RootQuery, TQuery>> query)
for queries and async Task<GqlResult<TQuery>> MutateAsync<TQuery>(Expression<Func<Mutation, TQuery>> query)
for mutations.
Example usage
var client = new GraphQLClient();
var result = await client.QueryAsync(q => new {
Movie = q.Movie(2, m => new {
m.Id,
m.Name,
ReleaseDate = m.Released,
Director = m.Director(d => new {
d.Name
})
}),
Actors = q.Actors(a => new {
a.Name,
a.Dob
})
});
This looks similar to the GraphQL. And now result.Data
is a strongly type object with the expected types.
If a field in the schema is an object (can have a selection protected on to it) it will be exposed in the generated code as a method where you pass in any field arguments first and then the selection.
The GraphQL created by the above will look like this
{
query {
Movie: movie(id: 2) {
Id: id
Name: name
ReleaseDate: released
Director: director {
Name: name
}
}
Actors: actors {
Name: name
Dob: dob
}
}
}
Mutations look very similar to the above query, just as they do in GraphQL. The above schema had one mutation that can be called like so
// mutations
var mutationResult = await client.MutateAsync(m => new {
NewPerson = m.AddPerson(new DateTime(1801, 7, 3), "John Johny", p => new { p.Id })
});
m
is the Mutation
type interface and lets you call any of the mutations generated from the schema. Like a query, you create a new anaoymous object and call as many mutations as you wish. Each mutation method has all their arguments and the last one is a selection query on the mutation's return type.
The above has mutationResult
strongly typed. E.g. mutationResult.Data.NewPerson
will only have an Id
field.
Any input types are generated as actual class
es as the are used as arguments. The -m Date=DateTime
tells the tool that the scalar
type Date
should be the DateTime
type in C#. You can use this to support any other custom scalar
types. E.g. -m Date=DateTime,Point=PointF
, just comma seperate a GqlScalar=DotnetType
list.