Normalize JSON API data for redux applications
JavaScript
Latest commit 6e55ae7 Feb 9, 2017 @yury-dymov version bump

README.md

json-api-normalizer

Utility to normalize JSON API data for redux applications

npm version Downloads Build Status Coverage Status

Description

json-api-normalizer helps awesome JSON API and redux work together. Unlike normalizr json-api-normalizer supports JSON API specification, which means that you don't have to care about schemes. It also converts collections into maps, which is a lot more suitable for redux.

Demo - https://yury-dymov.github.io/json-api-react-redux-example/

Demo sources - https://github.com/yury-dymov/json-api-react-redux-example

Works great together with redux-object, which helps to fetch and denormalize data from the store.

Example

import normalize from 'json-api-normalizer';

const json = {
  data: [{
    "type": "post-block",
    "relationships": {
      "question": {
        "data": {
          "type": "question",
          "id": "295"
        }
      }
    },
    "id": "2620",
    "attributes": {
      "text": "I am great!",
      "id": 2620
    }
  }],
  included: [{
    "type": "question",
    "id": "295",
    "attributes": {
      "text": "How are you?",
      id: 295
    }
  }]    
};

console.log(normalize(json));
/* Output:
{
  question: {
    "295": {
      attributes: {
        id: 295,
        text: "How are you?"
      }
    }
  },
  postBlock: {
    "2620": {
      attributes: {
        id: 2620,
        text: "I am great!"
      },
      relationships: {
        question: {
          type: "question",
          id: "295"
        }      
      }
    }
  }
}
*/

Data Representation Feature

One-to-Many relationship is merged into one map like this

const json = {
  data: [{
    type: "post-block",
    id: "1",
    attributes: { ... },
    relationships: {
      posts: {
        data: [{
          type: "post",
          id: "1"
        }, {
          type: "post",
          id: "2"          
        }]
      }
    },
    ...
  }]
}

console.log(normalize(json));
/* Output:
{
  postBlock: {
    "1": {
      ...,
      relationships: {
        posts: {
          type: "post",
          id: "1,2"           <----- this!
        }
      }
    }
  }
}
*/

This approach makes redux state updates a lot easier as during merge new string with IDs will replace the old one. Otherwise, it would require some extra efforts to delete outdated relationships from the store.

Options

Endpoint And Metadata

While using redux, it is supposed that cache is incrementally updated during the application lifecycle. However, you might face an issue if two different requests are working with the same data objects, and after normalization, it is not clear how to distinguish, which data objects are related to which request. json-api-normalizer can handle such situations by saving the API response structure as metadata, so you can easily get only data corresponding to the certain request.

console.log(normalize(json, { endpoint: '/post-block/2620' }));
/* Output:
{
  question: {
    ...
  },
  postBlock: {
    ...
  },
  meta: {
    "/post-block/2620": {
      data: [{
        type: "postBlock",
        id: "2620",
        relationships: {
          "question": {
            type: "question",
            id: "295"
          }      
      }]
    }
  }
}
*/

Endpoint And Query Options

By default request query options are ignored as it is supposed that data is incrementally updated. You can override this behavior by setting filterEndpoint option value to false.

const d1 = normalize(json, { endpoint: '/post-block/2620?page[cursor]=0' });
const d2 = normalize(json, { endpoint: '/post-block/2620?page[cursor]=20' });
console.log(Object.assign({}, d1, d2));
/* Output:
{
  question: {
    ...
  },
  postBlock: {
    ...
  },
  meta: {
    "/post-block/2620": {
      ...
    }
  }
}
*/

const d1 = normalize(json, { endpoint: '/post-block/2620?page[cursor]=0', filterEndpoint: false });
const d2 = normalize(json, { endpoint: '/post-block/2620?page[cursor]=20', filterEndpoint: false });
console.log(Object.assign({}, d1, d2));
/* Output:
{
  question: {
    ...
  },
  postBlock: {
    ...
  },
  meta: {
    "/post-block/2620: {
      "?page[cursor]=0": {
        ...
      },
      "?page[cursor]=20": {
        ...
      }
    }
  }
}
*/

Pagination And Links

If JSON API returns links section and you define the endpoint, then links are also stored in metadata.

const json = {
  data: [{
    ...
  }],
  included: [{
    ...
  }],
  links: {
    first: "http://example.com/api/v1/post-block/2620?page[cursor]=0",
    next: "http://example.com/api/v1/post-block/2620?page[cursor]=20"      
  }   
};

console.log(normalize(json, { endpoint: '/post-block/2620?page[cursor]=0'}));
/* Output:
{
  question: {
    ...
  },
  postBlock: {
    ...
  },
  meta: {
    "/post-block/2620": {
      data: [{
        ...
      }],
      links: {
        first: "http://example.com/api/v1/post-block/2620?page[cursor]=0",
        next: "http://example.com/api/v1/post-block/2620?page[cursor]=20"            
      }
    }
  }
}
*/

Lazy Loading

If you want to lazy load nested objects, json-api-normalizer will store links for that

const json = {
  data: [{
    attributes: {
      ...
    },
    id: "29",
    relationships: {
      "movie": {
        "links": {
          "self": "http://...",
          "related": "http://..."
        }
      },        
    },
    type: "question"
  }]
};

console.log(normalize(json));
/* Output:
{
  question: {
    "29": {
      attributes: {
        ...
      },
      relationships: {
        movie: {
          links: {
            "self": "http://...",
            "related": "http://..."
          }
        }        
      }
    }
  }
}
*/

Camelize Keys

By default all object keys and type names are camelized, however, you can disable this with camelizeKeys option.

const json = {
  data: [{
    type: 'post-block',
    id: '1',
    attributes: {
      'camel-me': 1,
      id: 1
    }
  }]
}

console.log(normalize(json));
/* Output:
{
  postBlock: {
    "1": {
      attributes: {
        id: 1,
        camelMe: 1
      }
    }
  }
}
*/

console.log(normalize(json, { camelizeKeys: false }));
/* Output:
{
  "post-block": {
    "1": {
      attributes: {
        id: 1,
        "camel-me": 1
      }
    }
  }
}
*/

Copyright

MIT (c) Yury Dymov