@hydre/rgraph

@hydre/rgraph

A high performance Node.js RedisGraph client.

Redisgraph >2.10

• Light & Fast
• Easy to use
• Functionnal nature
• Procedures caching
• Parameterized queries (since 4.3.0)
• Multi graph
• BigInt support
• Tagged templates literals
• Es modules
• Side effects free
• Loved by ladies
• On the blockchain
• Less than 0.1% gluten
• I'm out off buzzwords

Rgraph use tagged templates in order to seamlessly serialize any inputs while keeping a nice flow which looks like simple interpolation.

Install

npm i @hydre/rgraph

Node >16

node index.js

TL;DR

import Rgraph from '@hydre/rgraph'
import Redis  from 'ioredis'
import Events from 'events'

const client = new Redis()
const Graph  = Rgraph(client)
const foo    = Graph('foo')
const user   = { uuid: 'xxxx-xxxx-xxxx', name: 'Tony' }

await Events.once(client, 'ready')
await foo.run`MERGE (tony:User ${ user }) RETURN tony`
await foo.delete()

Debug

allow debug logs with the DEBUG='rgraph*' env variable

Usage

Get yourself some tea and meat by providing your redis client

import Rgraph from '@hydre/rgraph'
import Redis  from 'ioredis'

const redis_client = new Redis()
const use_graph    = Rgraph(redis_client)

Wow now you can get as much graphs as you want 💃 ! go ahead don't be afraid it's redis, not neo4j 🦐

const my_first_graph = use_graph('myFirstGraph')
const foo            = use_graph('anotherGraph')
const thanos         = use_graph('hail-hydra')
const bar            = use_graph('barGraph')

Run any Cypher query with a tag template or delete your poor graph

await foo.run`MATCH (n) RETURN n`
await thanos.delete()

Let me show you the result structure for each type, with the visible properties you also have access to some internals infos like ids and labels etc..

import { Internals } from '@hydre/rgraph'

[
  // each object is a sequence of the results
  {
    // For a scalar
    ['label in the RETURN statement']: 'value', // the returned value

    // For a node
    ['label in the RETURN statement']: {
      ['each']: ...,
      ['node']: ...,
      ['properties']: ...,
      [Internals.ID]: 0, // the node internal id
      [Internals.NODE_LABELS]: [], // the node labels
    },

    // For an edge
    ['label in the RETURN statement']: {
      ['each']: ...,
      ['node']: ...,
      ['properties']: ...,
      [Internals.ID]: 0, // the edge internal id
      [Internals.EDGE_LABEL]: 'label', // the edge label
      [Internals.SOURCE_NODE_ID]: 0, // the source node id
      [Internals.DESTINATION_NODE_ID]: 0 // the destination id
    },

    // For a path
    ['label in the RETURN statement']: {
      nodes: [], // an array of nodes (same representation as above but without the return label)
      edges: [] // an array of edges
    }
  }
]

Operators

The library provide some additional operators

import { raw } from '@hydre/rgraph/operators'

const maybe = false ? 'AND 1 = 2' : ''
const conditionnal = `WHERE 1 = 1 ${ raw(maybe) }`

Welcome to the era of fast graphs..

await myFirstGraph.run/* cypher */`
MERGE (foo:User ${ user })-[:Knows]->(thanos { name: 'Thanos', age: ${5 + 5}, a: ${true}, c: ${51.000000000016} })
WITH foo, thanos
MATCH path = ()-[]-()
RETURN path`

Comments

new in 4.2.0

When writting long queries you might want to help your team to understand it by adding comments. Every trimmed new line starting with // will be ignored

await graph.run/* cypher */`
  MATCH (u:User)
  WHERE (
    u.name = 'pepeg' AND
    u.age > 30
  )
  // hey i'm a comment
  RETURN u AS goog_pepeg
`

FAQ

Can i use nested objects ?

No, a Graph database use Nodes, not documents. A node is a hash of key value pairs and edges (relations) to other nodes

Commonjs support

I don't plan to support commonjs as i don't plan to ask tesla to use fuel. If you're still using commonjs, it will be better if you shutdown your computer right now

Edit me