typegql 4.0.1

A Python GraphQL library that makes use of type hinting and concurrency support with the new async/await syntax.

TypeGQL

A Python GraphQL library that makes use of type hinting and concurrency support with the new async/await syntax. With the help of type hints and dataclass it’s easy to build a GraphQL schema using nothing but Python objects and primitives

Consider the following:

from graphql import (
    GraphQLSchema, GraphQLObjectType, GraphQLField, GraphQLString)

schema = GraphQLSchema(
    query=GraphQLObjectType(
        name='RootQueryType',
        fields={
            'hello': GraphQLField(
                GraphQLString,
                resolve=lambda obj, info: 'world')
        }))

Versus:

from dataclasses import dataclass
from typegql import Schema

@dataclass(init=False)
class RootQuery:
    hello: str

def resolve_hello(info):
    return 'world

 schema = Schema(query=RootQuery)

Clearly the second one looks more “Pythonic” and it’s easier to maintain for complex structures

Installation

pip install typegql

Usage

The following demonstrates how to use typegql for implementing a GraphQL API for a library of books. The example can be found in typegql/core/examples and you can run it with Sanic by executing python <path_to_example>/server.py

Define your query

from dataclasses import dataclass
from typing import List
from typegql import Connection
from typegql.examples.library.types import Author, Category
from typegql.examples.library.types import Book
from typegql.examples.library import db


@dataclass(init=False, repr=False)
class Query:
    books: List[Book]
    authors: List[Author]
    categories: List[Category]

    books_connection: Connection[Book]

    async def resolve_authors(self, info, **kwargs):
        return db.get('authors')

    async def resolve_books(self, info, **kwargs):
        return db.get('books')

    async def resolve_categories(self, info, **kwargs):
        return db.get('categories')

   async def resolve_books_connection(self, info, **kwargs):
        data = db.get('books')
        return {
            'edges': [{
                'node': node
            } for node in data]}

Define your types

from dataclasses import dataclass, field
from datetime import datetime
from decimal import Decimal
from enum import Enum
from typing import List

from examples.library import db


class Gender(Enum):
    MALE = 'male'
    FEMALE = 'female'


@dataclass
class GeoLocation:
    latitude: Decimal
    longitude: Decimal


@dataclass
class Author:
    """Person that is usually a writer"""

    id: ID = field(metadata={'readonly': True})
    name: str
    gender: Optional[Gender] = None
    geo: Optional[GeoLocation] = None
    books: Optional[List[Book]] = None


@dataclass
class Category:
    id: ID = field(metadata={'readonly': True})
    name: str


@dataclass
class Book:
    """A book... for reading :|"""

    id: ID = field(metadata={'readonly': True})
    author_id: ID
    title: str
    author: Optional[Author] = field(default=None, metadata={'description': 'The author of this book'})
    categories: Optional[List[Category]] = None
    published: Optional[datetime] = None
    tags: Optional[List[str]] = None

    def __post_init__(self):
        self.published = datetime.strptime(self.published, '%Y-%m-%d %H:%M:%S')

    async def resolve_author(self, info):
        data = filter(lambda x: x['id'] == self.author_id, db.get('authors'))
        data = next(data)
        author = Author(**data)
        author.gender = Gender(author.gender)
        if 'geo' in data:
            author.geo = GeoLocation(**data.get('geo'))
        return author

    async def resolve_categories(self, selections, name=None):
        data = filter(lambda x: x['id'] in self.categories, db.get('categories'))
        for d in data:  # showcasing async generator
            yield Category(**d)

    def resolve_tags(self, selections):
        return ['testing', 'purpose']

Run your query

from typegql import Schema
from examples.library.query import Query


schema = Schema(Query)
query = '''
query BooksConnection {
  books_connection {
    edges {
      node {
        id
        title
        published
        author {
          id
          name
        }
      }
    }
  }
}
'''

async def run():
    result = await schema.run(query)

Client

TypeGQL supports DSL client for working with a GraphQL API. The client automatically converts snake to camelcase. Set camelcase=False if this is not desired.

pip install typegql[client]

For example:

from typegql.client import Client

async with Client(url, camelcase=True) as client:
    await client.introspection()
    dsl = client.dsl
    query = dsl.Query.books_connection.select(dsl.BooksConnection.total_count)
    doc = dsl.query(query)

    status, result = await client.execute(doc)

Change Log

4.0.1 [2020-03-17]

• fixes enum return value

4.0.0 [2020-03-16]

• BREAKING CHANGES:

  • Schema now accepts a more granular list of custom graphql types. Signature is:

def __init__(self,
         query: Type,
         mutation: Optional[Type] = None,
         subscription: Optional[Type] = None,
         scalars: Optional[GraphQLScalarMap] = None,
         enums: Optional[GraphQLEnumMap] = None,
         interfaces: Optional[GraphQLInterfaceMap] = None,
         query_types: Optional[GraphQLObjectTypeMap] = None,
         mutation_types: Optional[GraphQLInputObjectTypeMap] = None,
         camelcase=True):

3.1.0 [2020-01-29]

• fixes an issue with camelcase parameters when a load method is provided

3.0.9 [2020-01-13]

• fix client execution function

• update dsl selections to work with FrozenList

3.0.8 [2020-01-07]

• added ability to ignore attributes in building the schema by using field(metadata={‘skip’: True}). This can be useful when you don’t want to expose certain fields in GraphQL, like a user’s password for example.

3.0.5 [2019-12-16]

• added support for subscriptions

3.0.4 [2019-12-04]

• updates graphql-core-next to grapqhl-core >= 3

• use Sequence instead of List where possible

3.0.3 [2019-11-29]

• fixed a bug where a custom connection arguments don’t include the relay pagination arguments as well

3.0.2 [2019-11-26]

• PEP 561 compliant

3.0.1 [2019-11-26]

• BREAKING: Removed Graph as a baseclass

• now makes use of dataclasses.dataclass and dataclasess.fields for building the Schema

• bug fixing and improvements

2.0.9 [2019-10-29]

• changed the name of an input object from ObjectMuation to ObjectInput

2.0.8 [2019-10-15]

• allows forward reference between graph types (ie: Book has an author and an Author has books).

  this only works with python 3.7(using from __future__ import annotations, or python 3.8

2.0.6 [2019-06-24]

• updates uvloop dependency

2.0.5 [2019-04-24]

• fixed a bug when sending introspection schema

2.0.4 [2019-04-24]

• updates assert for introspection add message with status and result

• adds support for enum objects in resolve_field_velue_or_error

2.0.3 [2019-02-08]

• changes Connection, Edge, Node and PageInfo to interfaces IConnection, IEdge, etc.

• implements default Connection and PageInfo objects

• removes has_next, has_previous from PageInfo

2.0.1 [2019-01-19]

• all properties that don’t have a Field instance assigned to them will be ignored by the Schema

• updates docs & example to reflect 2.0 changes

• fixed a bug when using a List argument in mutations

1.0.7 [2018-12-09]

• bug fixing

• adds support for camelcase in Client

1.0.1 [2018-11-19]

• adds support for client DSL

Initial

• added graphql-core-next as a baseline for all GraphQL operations

TODO

• testing

• travis

• more testing

• please help with testing :|