Folders and files

Name		Name	Last commit message	Last commit date
Latest commit History 413 Commits
.github		.github
build		build
docs		docs
images		images
scripts		scripts
src		src
test		test
.gitignore		.gitignore
.jshintrc		.jshintrc
.npmignore		.npmignore
.travis.yml		.travis.yml
CONTRIBUTING.md		CONTRIBUTING.md
DEVELOPER.md		DEVELOPER.md
DOCUMENTATION.md		DOCUMENTATION.md
Gulpfile.js		Gulpfile.js
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md
bower.json		bower.json
jsdoc.json		jsdoc.json
package-lock.json		package-lock.json
package.json		package.json

Repository files navigation

Rekord

Rekord is an ORM - a way to define properties and relationships - that interacts with local storage, a RESTful service, and a real-time service.Rekord does this in a generic way soyou can use any libraries you wish or your own implementation - making it very easy to integrate into your existing projects.Rekord's design allows for offline capable and real-time applications if desired - offering complete control over how and where your data is stored.Rekord is the most powerful client-side Model/Active Record/ORM you'll find guaranteed - or your money back!

Download Stats

Features

RelationshipshasOne,belongsTo,hasMany,hasManyThrough,hasRemote,hasList, &hasReference
Polymorphic relationships forhasOne,belongsTo &hasMany
Validation (59 rules, 6 expressions, 14 transforms, and custom functions) throughrekord-validation
Migrations throughrekord-migrations
"Sessions" throughrekord-session
Batch REST execution for any/all types and operations
Inheritance (withextend option)
Horizontal scaling withsharding
Supports composite keys
Purging algorithms and "Contexts" to control memory/local storage usage
Specify default values
Handle collisions with a "revision" field
Handle propagating primary key changes returned by the server
Automatically refresh when application becomes online
Cache no data, all data, or only pending changes
Send only changed values to REST/real-time APIs or entire object
Convert values between client & server data types
Easily order by field, combination of fields, custom function, or expression
Use "Projections" to define subsets of data for efficient use
Control what information from relationships (if any) is stored locally or sent to the REST api
Add dynamic fields to model objects (setting & getting)
Data returned from REST calls or real-time events is intelligibly merged to avoid overwriting local unsaved changes
Addupdated_at andcreated_at timestamps and their automatic behavior with a single option
Configurable date/timestamp transformations
Add custom methods to the model objects
Asynchronous methods return Promises which can be chained together
Load bootstrapped data withmodel.boot( model or array of models )
Execute searches (fields are sent to REST API and an array of models is expected) withmodel.search( query, options, ... )
Execute paginated searches
Add global event listeners to the "database" or all model instances
Stores data locally throughRekord.store interface (ex:storkjs)
Stores data remotely throughRekord.rest interface (ex:angular,jquery,ajax,pouchdb,firebase,knexjs)
Real-time changes throughRekord.live interface (ex:pubsub,pouchdb,firebase)
Create a live filtered view of any collection
Create a live paginated view of any collection
All collections have the following notable operations: sort, page, filtered, where, subtract, intersect, complement, clear, removeWhere, min, max, first, last, sum, avg, count, pluck, reduce, random, chunk, reverse, & group
Model collections have the following notable operations: removeWhere, update, updateWhere, & saveWhere

FAQ (client-side usage)

Does Rekord directly interact with a database?
No, of course not. It interacts with a REST API.
Why do I need to use Rekord?
Working with relational data in javascript can be painful. Rekord eases that pain by allowing you to use plain looking objects that can have any type of relationship with other objects. Rekord takes into consideration things like foreign keys - where you need object A successfully remotely saved before you can save object B. These types of constraints are ugly and bothersome to handle yourself and easily result in bugs. If you're familiar with server-side ORMs, then Rekord should be really easy to pick up. You'll find all the same features and even more!
How are ID collisions avoided?
The key for a model can be given when creating a model - otherwise the key will be given a UUID. This is necessary to be offline capable, models need keys so related models can reference it. If thekeyChanges option is used the server can return a different key (like an auto-incrementing value) and the key changes will be propagated to all references to that model (foreign keys).
What do you mean by capable?
Caching data/changes locally and real-time behavior is optional - if you don't want either feature then you don't need to include an implementation.
Rekord can handle horizontal scaling via sharding?
Yes! You can say models of type X can exist on REST endpoints A, B, & C. You can provide a function which takes a model and returns the set of REST endpoints that need to be sent saves/removes. When you query on a sharded type it can contact all REST endpoints and combine the results.
Why do some functions in the API start with $?
TheRekord.Model andRekord.Search classes can have custom properties therefore to avoid collisions the functions and private variables start with$. If your design includes properties likestatus,operation,db,relations, etc it won't interfere with Rekord.