tmp__nrmlzr

0.0.2 • Public • Published

normalizr build status npm version npm downloads

Normalizes deeply nested JSON API responses according to a schema for Flux and Redux apps.
Kudos to Jing Chen for suggesting this approach.

Installation

npm install --save normalizr

Sample App

Flux

See flux-react-router-example.

Redux

See redux/examples/real-world.

The Problem

  • You have a JSON API that returns deeply nested objects;
  • You want to port your app to Flux or Redux;
  • You noticed it's hard for Stores (or Reducers) to consume data from nested API responses.

Normalizr takes JSON and a schema and replaces nested entities with their IDs, gathering all entities in dictionaries.

For example,

[{
  id: 1,
  title: 'Some Article',
  author: {
    id: 1,
    name: 'Dan'
  }
}, {
  id: 2,
  title: 'Other Article',
  author: {
    id: 1,
    name: 'Dan'
  }
}]

can be normalized to

{
  result: [1, 2],
  entities: {
    articles: {
      1: {
        id: 1,
        title: 'Some Article',
        author: 1
      },
      2: {
        id: 2,
        title: 'Other Article',
        author: 1
      }
    },
    users: {
      1: {
        id: 1,
        name: 'Dan'
      }
    }
  }
}

Note the flat structure (all nesting is gone).

Features

  • Entities can be nested inside other entities, objects and arrays;
  • Combine entity schemas to express any kind of API response;
  • Entities with same IDs are automatically merged (with a warning if they differ);
  • Allows using a custom ID attribute (e.g. slug).

Usage

import { normalize, Schema, arrayOf } from 'normalizr';

First, define a schema for our entities:

const article = new Schema('articles');
const user = new Schema('users');

Then we define nesting rules:

article.define({
  author: user,
  contributors: arrayOf(user)
});

Now we can use this schema in our API response handlers:

const ServerActionCreators = {
 
  // These are two different XHR endpoints with different response schemas.
  // We can use the schema objects defined earlier to express both of them:
 
  receiveOneArticle(response) {
 
    // Here, the response is an object containing data about one article.
    // Passing the article schema as second parameter to normalize() lets it
    // correctly traverse the response tree and gather all entities:
 
    // BEFORE:
    // {
    //   id: 1,
    //   title: 'Some Article',
    //   author: {
    //     id: 7,
    //     name: 'Dan'
    //   },
    //   contributors: [{
    //     id: 10,
    //     name: 'Abe'
    //   }, {
    //     id: 15,
    //     name: 'Fred'
    //   }]
    // }
    //
    // AFTER:
    // {
    //   result: 1,                    // <--- Note object is referenced by ID
    //   entities: {
    //     articles: {
    //       1: {
    //         author: 7,              // <--- Same happens for references to
    //         contributors: [10, 15]  // <--- other entities in the schema
    //         ...}
    //     },
    //     users: {
    //       7: { ... },
    //       10: { ... },
    //       15: { ... }
    //     }
    //   }
    // }
 
    response = normalize(response, article);
 
    AppDispatcher.handleServerAction({
      type: ActionTypes.RECEIVE_ONE_ARTICLE,
      response
    });
  },
 
  receiveAllArticles(response) {
 
    // Here, the response is an object with the key 'articles' referencing
    // an array of article objects. Passing { articles: arrayOf(article) } as
    // second parameter to normalize() lets it correctly traverse the response
    // tree and gather all entities:
 
    // BEFORE:
    // {
    //   articles: [{
    //     id: 1,
    //     title: 'Some Article',
    //     author: {
    //       id: 7,
    //       name: 'Dan'
    //     },
    //     ...
    //   },
    //   ...
    //   ]
    // }
    //
    // AFTER:
    // {
    //   result: {
    //    articles: [1, 2, ...]     // <--- Note how object array turned into ID array
    //   },
    //   entities: {
    //     articles: {
    //       1: { author: 7, ... }, // <--- Same happens for references to other entities in the schema
    //       2: { ... },
    //       ...
    //     },
    //     users: {
    //       7: { ... },
    //       ..
    //     }
    //   }
    // }
 
    response = normalize(response, {
      articles: arrayOf(article)
    });
 
    AppDispatcher.handleServerAction({
      type: ActionTypes.RECEIVE_ALL_ARTICLES,
      response
    });
  }
}

Finally, different Stores can tune in to listen to all API responses and grab entity lists from action.response.entities:

AppDispatcher.register((payload) => {
  const { action } = payload;
 
  if (action.response && action.response.entities && action.response.entities.users) {
    mergeUsers(action.response.entities.users);
    UserStore.emitChange();
    break;
  }
});

API Reference

new Schema(key, [options])

Schema lets you define a type of entity returned by your API.
This should correspond to model in your server code.

The key parameter lets you specify the name of the dictionary for this kind of entity.

const article = new Schema('articles');
 
// You can use a custom id attribute
const article = new Schema('articles', { idAttribute: 'slug' });
 
// Or you can specify a function to infer it
function generateSlug(entity) { /* ... */ }
const article = new Schema('articles', { idAttribute: generateSlug });
 
// You can also specify meta properties to be used for customizing the output in assignEntity (see below)
const article = new Schema('articles', { idAttribute: 'slug', meta: { removeProps: ['publisher'] }});

Schema.prototype.define(nestedSchema)

Lets you specify relationships between different entities.

const article = new Schema('articles');
const user = new Schema('users');
 
article.define({
  author: user
});

Schema.prototype.getKey()

Returns the key of the schema.

const article = new Schema('articles');
 
article.getKey();
// articles

Schema.prototype.getIdAttribute()

Returns the idAttribute of the schema.

const article = new Schema('articles');
const slugArticle = new Schema('articles', { idAttribute: 'slug' });
 
article.getIdAttribute();
// id
slugArticle.getIdAttribute();
// slug

Schema.prototype.mappedBy(foreignKey)

This method permits to link a Schema property with a foreign property in another one Schema.

const project = new Schema('projects');
const user = new Schema('users');
 
project.define({
  collaborators: user.mappedBy('projects'),
});
 
user.define({
  projects: user.mappedBy('collaborators'),
});

This way, the relations will always be consistent. If you get this response for a user :

{
  id: 1,
  projects: [1, 2],
}

Normalizr will give you this :

{
  result: [1],
  entities: {
    users: {
      1: { id: 1, projects: [1, 2] }
    },
    projects: {
      1: { id: 1, collaborators: [1] },
      2: { id: 1, collaborators: [1] }
    }
  }
}

Using this property is completely optional and you can even chose to use it only on one side of the relationship.

arrayOf(schema, [options])

Describes an array of the schema passed as argument.

const article = new Schema('articles');
const user = new Schema('users');
 
article.define({
  author: user,
  contributors: arrayOf(user)
});

If the array contains entities with different schemas, you can use the schemaAttribute option to specify which schema to use for each entity:

const article = new Schema('articles');
const image = new Schema('images');
const video = new Schema('videos');
const asset = {
  images: image,
  videos: video
};
 
// You can specify the name of the attribute that determines the schema
article.define({
  assets: arrayOf(asset, { schemaAttribute: 'type' })
});
 
// Or you can specify a function to infer it
function inferSchema(entity) { /* ... */ }
article.define({
  assets: arrayOf(asset, { schemaAttribute: inferSchema })
});

valuesOf(schema, [options])

Describes a map whose values follow the schema passed as argument.

const article = new Schema('articles');
const user = new Schema('users');
 
article.define({
  collaboratorsByRole: valuesOf(user)
});

If the map contains entities with different schemas, you can use the schemaAttribute option to specify which schema to use for each entity:

const article = new Schema('articles');
const user = new Schema('users');
const group = new Schema('groups');
const collaborator = {
  users: user,
  groups: group
};
 
// You can specify the name of the attribute that determines the schema
article.define({
  collaboratorsByRole: valuesOf(collaborator, { schemaAttribute: 'type' })
});
 
// Or you can specify a function to infer it
function inferSchema(entity) { /* ... */ }
article.define({
  collaboratorsByRole: valuesOf(collaborator, { schemaAttribute: inferSchema })
});

unionOf(schemaMap, [options])

Describe a schema which is a union of multiple schemas. This is useful if you need the polymorphic behavior provided by arrayOf or valuesOf but for non-collection fields.

Use the required schemaAttribute option to specify which schema to use for each entity.

const group = new Schema('groups');
const user = new Schema('users');
 
// a member can be either a user or a group
const member = {
  users: user,
  groups: group
};
 
// You can specify the name of the attribute that determines the schema
group.define({
  owner: unionOf(member, { schemaAttribute: 'type' })
});
 
// Or you can specify a function to infer it
function inferSchema(entity) { /* ... */ }
group.define({
  creator: unionOf(member, { schemaAttribute: inferSchema })
});

A unionOf schema can also be combined with arrayOf and valueOf with the same behavior as each supplied with the schemaAttribute option.

const group = new Schema('groups');
const user = new Schema('users');
 
const member = unionOf({
  users: user,
  groups: group
}, { schemaAttribute: 'type' });
 
group.define({
  owner: member,
  members: arrayOf(member),
  relationships: valuesOf(member)
});

normalize(obj, schema, [options])

Normalizes object according to schema.
Passed schema should be a nested object reflecting the structure of API response.

You may optionally specify any of the following options:

  • assignEntity (function): This is useful if your backend emits additional fields, such as separate ID fields, you'd like to delete in the normalized entity. See the tests and the discussion for a usage example.

  • mergeIntoEntity (function): You can use this to resolve conflicts when merging entities with the same key. See the test and the discussion for a usage example.

const article = new Schema('articles');
const user = new Schema('users');
 
article.define({
  author: user,
  contributors: arrayOf(user),
  meta: {
    likes: arrayOf({
      user: user
    })
  }
});
 
// ...
 
// Normalize one article object
const json = { id: 1, author: ... };
const normalized = normalize(json, article);
 
// Normalize an array of article objects
const arr = [{ id: 1, author: ... }, ...]
const normalized = normalize(arr, arrayOf(article));
 
// Normalize an array of article objects, referenced by an object key:
const wrappedArr = { articles: [{ id: 1, author: ... }, ...] }
const normalized = normalize(wrappedArr, {
  articles: arrayOf(article)
});

Explanation by Example

Say, you have /articles API with the following schema:

articles: article*

article: {
  author: user,
  likers: user*
  primary_collection: collection?
  collections: collection*
}

collection: {
  curator: user
}

Without normalizr, your Stores would need to know too much about API response schema.
For example, UserStore would include a lot of boilerplate to extract fresh user info when articles are fetched:

// Without normalizr, you'd have to do this in every store:
 
AppDispatcher.register((payload) => {
  const { action } = payload;
 
  switch (action.type) {
  case ActionTypes.RECEIVE_USERS:
    mergeUsers(action.rawUsers);
    break;
 
  case ActionTypes.RECEIVE_ARTICLES:
    action.rawArticles.forEach(rawArticle => {
      mergeUsers([rawArticle.user]);
      mergeUsers(rawArticle.likers);
 
      mergeUsers([rawArticle.primaryCollection.curator]);
      rawArticle.collections.forEach(rawCollection => {
        mergeUsers(rawCollection.curator);
      });
    });
 
    UserStore.emitChange();
    break;
  }
});

Normalizr solves the problem by converting API responses to a flat form where nested entities are replaced with IDs:

{
  result: [12, 10, 3, ...],
  entities: {
    articles: {
      12: {
        authorId: 3,
        likers: [2, 1, 4],
        primaryCollection: 12,
        collections: [12, 11]
      },
      ...
    },
    users: {
      3: {
        name: 'Dan'
      },
      2: ...,
      4: ....
    },
    collections: {
      12: {
        curator: 2,
        name: 'Stuff'
      },
      ...
    }
  }
}

Then UserStore code can be rewritten as:

// With normalizr, users are always in action.response.entities.users
 
AppDispatcher.register((payload) => {
  const { action } = payload;
 
  if (action.response && action.response.entities && action.response.entities.users) {
    mergeUsers(action.response.entities.users);
    UserStore.emitChange();
    break;
  }
});

Dependencies

  • Some methods from lodash, such as isObject, isEqual and mapValues

Browser Support

Modern browsers with ES5 environments are supported.
The minimal supported IE version is IE 9.

Running Tests

git clone https://github.com/gaearon/normalizr.git
cd normalizr
npm install
npm test # run tests once
npm run test:watch # run test watcher

Dependents (0)

Package Sidebar

Install

npm i tmp__nrmlzr

Weekly Downloads

1

Version

0.0.2

License

MIT

Last publish

Collaborators

  • frankwallis