graffy

Graffy

Graffy is a Javascript library to power efficient and intuitive real-time APIs. It's built on a sound set-theoretic model of queries and graphs.

Why Graffy?

Graffy provides live queries, which give clients a real-time view of the data they need. Graffy supports complex queries with nested graph traversals and pagination, while exposing a simple and intuitive API for building clients and servers.

Graffy was inspired by (and borrows from) Facebook's GraphQL and Netflix's Falcor. Compared to GraphQL, Graffy offers a more familiar data model, true live queries and more efficient caching. Compared to Falcor, it provides cursor-based pagination and real-time subscriptions.

Unlike GraphQL resolvers and Falcor data providers, Graffy providers can be composed like Express/Koa middleware. This allows authentication, validation, custom caches and resource limiting to be implemented in a straightforward manner.

Graffy providers can also perform efficient bulk reads from underlying data stores (for example by constructing optimized SQL queries). This is particularly hard to do with GraphQL (see dataloader) and Falcor.

Documentation

Development status

Graffy is under heavy development and not ready for production use.

Module	Status	Provides
core		-
client		EventStream/HTTP client
server		EventStream/HTTP server
cache		In-memory cache
react		Container and hooks API
gql	⌛ backlog	GraphQL subset DSL for queries
schema	⌛ backlog	Validation, introspection API
viewer	⌛ backlog	Schema introspection client
auth	⌛ backlog	Authentication and authorization
limit	⌛ backlog	Resource consumption accounting

Capabilities

	Graffy	GraphQL	Falcor	Description
Narrow queries	✅	✅	✅	Queries specify required fields; Allows API evolution
Deep queries	✅	✅	✅	Queries can expand nested resources; Reduces round-trips
Live queries	✅	❌	❌	Push changes to query results in real time
Pagination cursors	✅	✅	❌	Enables efficient pagination on the server
Parameters	✅	✅	❌	Custom filtering criteria, etc.
Caching pages	✅	❌	✅	Cache result of paginated queries
Atomic writes	✅	❌	✅	Writes that trigger accurate cache invalidation
Non-data endpoints	✅	✅	❌	Mutations, subscriptions, cross-resource search