fastgraphql 0.0.4

FastGraphQL is intended to help developer create code driven GraphQL APIs

FastGraphQL

FastGraphQL is intended to help developer create code driven GraphQL APIs.

Disclaimer

This is still a work in progress

Motivation

So far most of the projects that uses GraphQL need to duplicate many definitions to be able to have a consistent GraphQL API schema alongside well-defined models that governs the development and the application.

FastGraphQL tries to shortcut the path between python models and GraphQL schema using Pydantic models. This ensures not only a single source of truth when comes to type, inputs, query and mutation definition reflected in classes and methods, but also the ability to use Pydantic to validate models.

Installation

pip install fastgraphql

Usage

GraphQL Types and Inputs

Using annotation driven definitions and Pydantic, defining GraphQL types and inputs can be done by simple annotating Pydantic models with FastGraphQL.graphql_type() of FastGraphQL.graphql_input()

from datetime import datetime
from typing import Optional
from pydantic import BaseModel
from fastgraphql import FastGraphQL

fast_graphql = FastGraphQL()


@fast_graphql.type()
class Model(BaseModel):
    t_int: int
    t_opt_int: Optional[int]
    t_str: str
    t_opt_str: Optional[str]
    t_float: float
    t_opt_float: Optional[float]
    t_datatime: datetime
    t_opt_datatime: Optional[datetime]
    t_boolean: bool
    t_opt_boolean: Optional[bool]


@fast_graphql.input()
class Input(BaseModel):
    t_int: int


print(fast_graphql.render())

The above code example generates a schema as follows:

scalar DateTime

type Model {
    t_int: Int!
    t_opt_int: Int
    t_str: String!
    t_opt_str: String
    t_float: Float!
    t_opt_float: Float
    t_datatime: DateTime!
    t_opt_datatime: DateTime
    t_boolean: Boolean!
    t_opt_boolean: Boolean
}

input Input {
    t_int: Int!
}

Query and Mutation

Following the same approach with annotation driven defitions, query and mutations can easily be defined using FastGraphQL.graphql_query and FastGraphQL.mutation.

Note that all function arguments annotated with FastGraphQL.graphql_query_field are considered to be input arguments for the GraphQL API and simples types and Pydantic models can be used and arguments and also as return type and they don't need to be explicitly annotated.

from fastgraphql import FastGraphQL
from pydantic import BaseModel

fast_graphql = FastGraphQL()


class Model(BaseModel):
    param: str


@fast_graphql.query()
def my_first_query(
        model: Model = fast_graphql.parameter(),
        param: str = fast_graphql.parameter()
) -> str:
    ...

@fast_graphql.mutation()
def my_first_mutation(
        model: Model = fast_graphql.parameter(),
        param: str = fast_graphql.parameter()
) -> str:
    ...


print(fast_graphql.render())

The above code example generates a schema as follows:

input Model {
    param: String!
}
type Query {
    my_first_query(model: Model!, param: String!): String!
}

type Query {
    my_first_mutation(model: Model!, param: String!): String!
}

Dependecy Injection

Query and Mutation can have dependencies injected using FastGraphQL.depende(...) as showed bellow:`

from fastgraphql import FastGraphQL
from pydantic import BaseModel

fast_graphql = FastGraphQL()


class Model(BaseModel):
    param: str


def create_dependency() -> str:
    return ""


@fast_graphql.query()
def my_first_query(
        model: Model = fast_graphql.parameter(),
        dependecy: str = fast_graphql.depends(create_dependency)
) -> str:
    ...

In this example the parameter dependecy will be injected once the query is called.

Integrations

Ariadne

The developed GraphQL API can be easily integration with Ariadne.

pip install fastgraphql[ariadne]

The method make_executable_schema in the module fastgraphql.ariadne can create the ariadne's executable schema to integrate to other frameworks like FastAPI. See https://ariadnegraphql.org/docs/starlette-integration

FastAPI

The developed GraphQL API can be easily served using Ariadne.
and FastAPI.

pip install fastgraphql[ariadne,fastapi]

Note that Ariadne is needed to serve GraphQL APIs through FastAPI because no other GraphQL framework are yet integrated

To create the router that server the GraphQL API, make_ariadne_fastapi_router from the module fastgraphql.fastapi should be used. For example:

from fastgraphql import FastGraphQL
from fastapi import FastAPI
from fastgraphql.fastapi import make_ariadne_fastapi_router

app = FastAPI()
fast_graphql = FastGraphQL()

...

app.include_router(
    make_ariadne_fastapi_router(fast_graphql=fast_graphql)
)
    

SQLAlchemy

To integrate SQLAlchemy models to the GraphQL API first all dependency for SQLAlchemy should be installed using:

pip install fastgraphql[sqlalchemy]

SQLAlchemy, then, models can be incorporated to the GraphQL API by first telling FastGraphQLwhat which is the base class to be considered:

from fastgraphql import FastGraphQL
from sqlalchemy.ext.declarative import declarative_base

fast_graphql = FastGraphQL()
...
Base = declarative_base(...)
...

fast_graphql.set_sqlalchemy_base(Base)

and after that, any SQLAlchemy model can be using as types and inputs. The models can also be used as query's and mutation's inputs and outpus. For example:

from fastgraphql import FastGraphQL
from sqlalchemy import Column, Integer
from sqlalchemy.ext.declarative import declarative_base

fast_graphql = FastGraphQL()
...
Base = declarative_base(...)
...
fast_graphql.set_sqlalchemy_base(Base)

@fast_graphql.type()
class MyModel(Base):
    id = Column(Integer, primary_key=True)

@fast_graphql.mutation()
def mutation(input: int) -> MyModel:
    ...

Acknowledgment

Thanks FastAPI for inspiration