Mongql

A package to convert your mongoose schema to graphql schema

Features

Create graphql schema (typedef and resolvers) from mongoose schema
Stitch already created typedef and resolvers
Easily configurable (any of the typedef and resolvers can be turned off)
Output the generated SDL
Auto addition of graphql validators with mongoose
Auto generation of fragments and operations
Automatic integration with custom various scalar types from graphql-scalars
Helper methods for dealing with graphql ast nodes
Auto generation of fragments and operations
Auto generation of basic crud operations

Motivation

Creating a graphql SDL is not a difficult task, but things get really cumbersome after a while, especially since a lot of the typedefs and resolvers are being repeated.
Automating the schema generation helps to avoid errors regarding forgetting to define something in the schema thats been added to the resolver or vice versa.
Creating resolvers for subtypes in a PITA, especially if all of them just refers to the same named key in parent
Generating fragments and operation manually is quite difficult and error prone, typing duplicate stuffs and figuring out the interrelationship between each fragment.

Usage

Without initial typedef and resolvers

// User.schema.js
const mongoose = require('mongoose');

const UserSchema = mongoose.Schema({
  name: {
    type: String,
    mongql: {
      nullable: {
        object: [true]
      } // field level config
    }
  }
});

UserSchema.mongql = {
  resource: 'user'
}; // schema level config

module.exports = UserSchema;

// index.js
const {
  makeExecutableSchema
} = require('@graphql-tools/schema');
const {
  ApolloServer
} = require('apollo-server-express');
const Mongql = require('MonGql');

const UserSchema = require('./User.schema.js');

(async function() {
  const mongql = new Mongql({
    Schemas: [UserSchema], // Global level config
  });

  // Calling the generate method generates the typedefs and resolvers
  const {
    TransformedResolvers,
    TransformedTypedefs, // Contains both arr and obj representation
  } = await mongql.generate();

  const GRAPHQL_SERVER = new ApolloServer({
    schema: makeExecutableSchema({
      typeDefs: TransformedTypedefs.arr,
      resolvers: TransformedResolvers.arr,
    }),
    context: () => ({
      user: req.user
    })
  });
})();

With initial typedef and resolvers

// user.typedef.js
module.exports = gql `
  type BasicUserType{
    name:String!
  }
`;

// user.resolver.js
module.exports = {
  Mutation: {
    updateUserSettings: ...
  }
}

const UserAST = require('./user.typedef.js');
const UserResolver = require('./user.resolver.js');

const PreTransformedTypeDefsASTs = {
  user: UserAST // This has to match with the resource name added in the mongoose schema
}

const PreTransformedResolvers = {
  user: UserResolver
}

const mongql = new Mongql({
  Schemas: [UserSchema, SettingsSchema],
  Typedefs: {
    init: PreTransformedTypeDefsASTs
  },
  Resolvers: {
    init: PreTransformedResolvers
  }
});

Output SDL and AST

  const mongql = new Mongql({
    Schemas: [],
    output: {
      SDL: path.resolve(__dirname, "./SDL"),
      AST: path.resolve(__dirname, "./AST")
    }
  });

  await mongql.generate()

Using Schema config typedefs and resolvers

  const UserSchema = new mongoose.model({
    name: String
  });
  UserSchema.mongql = {
    TypeDefs: `type userinfo{
      name: String!
    }`,
    Resolvers: {
      Query: {
        getUserInfo() {}
      }
    }
  }
  const mongql = new Mongql({
    Schemas: [UserSchema],
  });

  await mongql.generate()

FieldSchema Configs

  const NestedSchema = new mongoose.model({
    nested: Boolean
  });

  NestedSchema.mongql = {
    // FieldSchema configs
  }

  const UserSchema = new mongoose.model({
    name: String,
    nested: NestedSchema
  });

  UserSchema.mongql = {
    TypeDefs: `type userinfo{
      name: String!
    }`,
    Resolvers: {
      Query: {
        getUserInfo() {}
      }
    }
  }
  const mongql = new Mongql({
    Schemas: [UserSchema],
  });

  await mongql.generate()

Fine grain Mutation configuration

const mongql = new Mongql({
  Schemas: [UserSchema, SettingsSchema],
  generate: {
    mutation: false, // will not generate any mutation typedef and resolver,
    mutation: {
      create: false, // Will not generate any create mutation typedef and resolver,
      update: {
        multi: false // Will not generate any update multi mutation typedef and resolver
      },
      single: false // Will not generate any single mutation typedef and resolver
    }
  }
});

Fine grain Query configuration

const mongql = new Mongql({
  Schemas: [UserSchema, SettingsSchema],
  generate: {
    query: false,
    query: {
      all: false
    },
    query: {
      paginated: {
        self: false
      }
    },
    query: {
      filtered: {
        others: {
          whole: false
        }
      }
    },
    query: {
      self: false, // remove all self related typedefs and resolvers,
      self: {
        whole: false // remove all selfwhole related typedefs and resolvers,
      },
      count: false, // remove all count related typedefs and resolvers,
    }
  }
});

Fine grain Type configuration

const mongql = new Mongql({
  Schemas: [UserSchema, SettingsSchema],
  generate: {
    input: {
      update: false
    },
    interface: false,
    enum: false,
    union: false,
    object: {
      self: false
    }
  }
});

generating Models

const {
  makeExecutableSchema
} = require('@graphql-tools/schema');
const Mongql = require('mongql');
const {
  ApolloServer
} = require('apollo-server');

(async function() {
  const mongql = new Mongql({
    Schemas: [
      /* Your schema array here */
    ],
  });
  const {
    TransformedTypedefs,
    TransformedResolvers
  } = await mongql.generate();
  const server = new ApolloServer({
    schema: makeExecutableSchema({
      typeDefs: TransformedTypedefs.arr,
      resolvers: TransformedResolvers.arr,
    }),
    context: mongql.generateModels()
  });
  await server.listen();
})();

Using local folders

const Mongql = require('mongql');
const {
  ApolloServer
} = require('apollo-server');

(async function() {
  const mongql = new Mongql({
    Schemas: path.resolve(__dirname, './schemas'),
    output: {
      dir: __dirname + '\\SDL'
    },
    Typedefs: {
      init: path.resolve(__dirname, './typedefs')
    },
    Resolvers: {
      init: path.resolve(__dirname, './resolvers')
    }
  });
  const server = new ApolloServer({
    schema: makeExecutableSchema({
      typeDefs: TransformedTypedefs.arr,
      resolvers: TransformedResolvers.arr,
    }),
    context: mongql.generateModels()
  });
  await server.listen();
})();

Controlling nullability

const UserSchema = new mongoose.model({
  name: {
    type: String,
    mongql {
      nullable: {
        object: [true]
      } // name: String
    }
  },
  age: {
    type: [Number],
    mongql: {
      nullable: {
        input: [false, true]
      } // age: [Int!]
    }
  }
});

Mongql contains 4 levels of configs

Constructor/global level config: passed to the ctor during Mongql instantiation
Schema level config: Attached to the schema via mongql key
Field level config: Attached to the field via mongql key
FieldSchema level config: Contains both field and schema configs

Precedence of same config option is global < Schema < FieldSchema < field. That is for the same option the one with the highest precedence will be used.

Concept

During the generation of schema, a few concepts are followed

Generation of Query

Each Resource query object type contains four parts
1. Range(Input):
  1. All: Gets all the resource
  2. Paginated : Used to get resource through pagination inpu
  3. Filtered : Used to get resource through filter input
  4. ID: Used to get a resource by id
2. Auth:
  1. Self: Used to indicate logged in users resource
  2. Others: Used to indicate other users resource (when current user is authenticated)
  3. Mixed: Used to incicate others users resource (when user is unauthenticated)
3. Resource: Name of the resource (capitalized & pluralized form)
4. Part(Output):
  1. Whole: Get the whole data along with sub/extra types
  2. Count: get the count of resource

Generated Query Examples: getSelfSettingsWhole, getOthersSettingsCount ;

NOTE: Count part is not generate in paginated and id range as for paginated query the user already knows the count and id returns just one

Generation of Mutation

Each resource mutation object type contains 2 parts
1. Action: One of create|update|delete
2. Target: resource for targeting single resource, resources for targeting multiple resources
Specific named functions pre|post(action) attached to the model's static is called pre and post of each action

Generated Mutation Examples: createSetting, updateSettings

Generation of Types

Each resource types contains the following parts
1. For each schema (base and nested), based on the permitted auth, object will be created, and based on generate config interface, input and union will be created

Generation of Fragments

Each object type gets converted into 3 differnet kinds of fragment
1. RefsNone: A fragment which excludes all the Refs Object type
2. ScalarsOnly: A fragment which includes its and all included object types scalar fields
3. ObjectsNone: A fragment which includes only scalar fields
All the custom fragments are generated from all the schemas

Generation of Operations

All the generated operations have various kinds based on the generated fragments of that schema

API

All of the methods and configs have been commented along with their types

TODO

Add more well rounded tests
Provide ES modules to make the library tree-shakabl

PRS are more than welcome and highly appreciated!!!!

Devorein/mongql