@nutshelllab/aurora-store v0.0.1

Goal

This repo was originaly made to provide a serverless aurora postgresql connector to be used in Keyro Login with a strong retro-compatibility with our dynamodb connector in mind. By implementing the same API, it offers the opportunity to easily switch between Aurora and Dynamo.

Usage

Installation

yarn add @nutshelllab/aurora-store

This module allows you to run your code on a local database. To do so, you will also need to have a proper postgresql installation running.

Configuration

Config is done through env variables so that you can use your serverless environnement to handle it as code :

Access to the database

Migrations

You can easily configure a migration source to handle your database schema. Migration are internally done through Knex, so that the configuration is identical. Here's an example with webpack context :

import { configure } from '@nutshelllab/aurora-store' //The module expose a simple configure method

const context = require.context('../../migrations', false, /\.js$/)

const makeMigrationSource = context => {
  return {
    getMigrations() {
      return Promise.resolve(context.keys().sort())
    },
    getMigrationName(migration) {
      return migration
    },
    getMigration(migration) {
      return context(migration)
    }
  }
}

configure({
  migrationSource: makeMigrationSource(context) // Configuration is mutating the state, keep it for initialization as much as possible.
})

Initialize a store

To initialize a store, just call default function of the module :

import createStore from '@nutshelllab/aurora-store'

//[...] configuration has been done at this point

const teamStore = createStore({ kind: 'teams', validate: makeTeam })
const result = await teamStore.find({ id: 'team-00007'})

Options you can pass to the default store builder function are :

Features

• findIn
• findAll
• find
• put
• update
• create
• findOrCreate
• remove
• truncate
• knexClient

findIn(field, values)

Returns all the records having a field's value intercecting with a predefined array of value.

store.findIn('id', ['001', '002', '003'])

findAll(filters)

Returns all the records matching a pattern

store.findAll({ contracted: true })

find(filters)

Returns exactly one record matching a pattern or throw an error with name attribute set to 'NOT_FOUND'.

store.find({ id: '001' })

put(data)

Write an object with the store. If it does not exist, it will be created.

store.put({ id: 'movie-0001', name: 'Black Widow' })

update(data)

Update the record matching with data id (default is 'id' field, see idField configuration to change it).

store.update({id: 'movie-0001', name: 'Black Windows' })

create(data)

Insert a new record with the given object values. AlreadyExisting will throw.

store.create({id: 'movie-0002', name: 'Black Windows' })

findOrCreate(filters, data)

Fetch a record with filters. A creation attempt is made it nothing is found. Bad filtering may result in throwing already exist error.

store.create({name: 'Black Windows'}, { id: 'newId-002', name: 'Black Windows', company: 'marvelcrosoft'})

remove(filters)

Destroy all records with filters forever.

store.remove({name: 'Black Windows'})

truncate(options)

Destroy all records of a table forever. You must pass force attribute to true to execute this as a way to make it clear that you know what you're doing.

store.truncate() // won't work
store.truncate({ force: true })

knexClient()

Obviously the point of using a relational database was to go beyond dynamo key-values restriction. This function gives you a Knex querybuilder. You can use that to manipulate any table in a sql-query manner :

store.knexClient().then(db =>
  db('movies').innerjoin('authors', 'authors.id', 'movies.authorId').where('authors.name', 'marvelcrosoft').select('movies.*')
)

See knex documentation for further features.

Combine with serverless-offline

With serverless-offline, this module can be used to provide a 100% local execution which may greatly increase your productivity. Keyro Login has been tested and successfully run on a localhost with a local duplicated database of our remote dev environement.

Contribute

1. Install postgresql
2. Create a pguser user with a pguser password
3. yarn
4. Check integration tests yarn test