python-jsonapi-datastore

Client-side JSON API data handling made easy. This is a re-write of Lucas Hosseini's JavaScript jsonapi-datastore for Python. The serialization is not included yet. If you need it feel free to open a PR or ask for it.

Description

The JSONAPI standard is great for exchanging data (which is its purpose), but the format is not ideal to work directly within an application. python-jsonapi-datastore is a framework-agnostic library that takes away the burden of handling JSONAPI data on the client side.

What it does:

• read JSONAPI payloads,
• rebuild the underlying data graph,
• allows you to query models and access their relationships directly,
• create new models.

What it does not do:

• make requests to your API. You design your endpoints URLs, the way you handle authentication, caching, etc. is totally up to you.

Installing

Install python-jsonapi-datastore with pip as usual:

pip install jsonapi-datastore

This library has no 3rd-party dependencies.

Parsing data

Just call the .sync() method of your store.

store = JsonApiDataStore()
store.sync(data)

This parses the data and incorporates it in the store, taking care of already existing records (by updating them) and relationships.

Parsing with meta data

If you have meta data in your payload use the .syncWithMeta method of your store.

store = JsonApiDataStore()
store.syncWithMeta(data)

This does everything that .sync() does, but returns an object with data and meta split.

Retrieving models

Just call the .find(type, id) method of your store.

article = store.find('article', 123)

or call the .findAll(type) method of your store to get all the models of that type.

articles = store.findAll('article')

All the attributes and relationships are accessible through the model as object properties.

print(article.author.name)

In case a related resource has not been fetched yet (either as a primary resource or as an included resource), the corresponding property on the model will contain only the type and id (and the ._placeHolder property will be set to true). However, the models are updated in place, so you can fetch a related resource later, and your data will remain consistent.

Examples

# Create a store:
store = JsonApiDataStore()

# Then, given the following payload, containing two `articles`, with a related `user` who is the author of both:
payload = {
  'data': [{
    'type': 'article',
    'id': 1337,
    'attributes': {
      'title': 'Cool article'
    },
    'relationships': {
      'author': {
        'data': {
          'type': 'user',
          'id': 1
        }
      }
    }
  },
  {
    'type': 'article',
    'id': 300,
    'attributes': {
      'title': 'Even cooler article'
    },
    'relationships': {
      'author': {
        'data': {
          'type': 'user',
          'id': 1
        }
      }
    }
  }]
}

# we can sync it:
articles = store.sync(payload)

# which will return the list of synced articles.

# Later, we can retrieve one of those:
article = store.find('article', 1337)

# If the author resource has not been synced yet, we can only access its id and its type:
print(article.author)
# { id: 1, _type: 'article' }

# If we do sync the author resource later:
authorPayload = {
  'data': {
    'type': 'user',
    'id': 1,
    'attributes': {
      'name': 'Lucas'
    }
  }
}

store.sync(authorPayload)

# we can then access the author's name through our old `article` reference:
print(article.author.name)
# 'Lucas'

What's missing

Currently, the store does not handle links attributes or resource-level or relationship-level meta.

Contributing

Pull-requests welcome!