stargate-mongoose

stargate-mongoose is a Mongoose driver for Data API which runs on top of Apache Cassandra / DataStax Enterprise.

1. Quickstart
2. Architecture
3. Version compatibility
4. Connecting to AstraDB
5. Sample Applications
6. Features
7. NodeJS MongoDB Driver Overriding (experimental)
8. API Reference
9. Developer Guide

Quickstart

Prerequisites: node (>=14.0.0), npm/yarn, Docker (for testing the sample app locally using docker compose)

• Start Docker on your local machine.
• Clone this repository

git clone https://github.com/stargate/stargate-mongoose.git
cd stargate-mongoose

• Execute below script and wait for it to complete, which starts a simple Data API on local with a DSE 6.8 (DataStax Enterprise) as database backend.

For macOS/Linux

bin/start_data_api.sh

For Windows

bin\start_data_api.cmd

• Create a sample project called 'sample-app'

mkdir sample-app
cd sample-app

• Initialize and add required dependencies

npm init -y && npm install express mongoose stargate-mongoose --engine-strict

OR

yarn init -y && yarn add express mongoose stargate-mongoose

• Create a file called index.js under the 'sample-app' directory and copy below code into the file.

//imports
const express = require('express');
const mongoose = require('mongoose');
const stargate_mongoose = require('stargate-mongoose');
const Schema = mongoose.Schema;
const driver = stargate_mongoose.driver;

//override the default native driver
mongoose.setDriver(driver);

//Set up mongoose & end points definition
const Product = mongoose.model('Product', new Schema({ name: String, price: Number }));
mongoose.connect('http://localhost:8181/v1/inventory', {
    username: 'cassandra',
    password: 'cassandra'
});
Object.values(mongoose.connection.models).map(Model => Model.init());
const app = express();
app.get('/addproduct', (req, res) => {
    const newProduct = new Product(
        {
            name: 'product' + Math.floor(Math.random() * 99 + 1),
            price: '' + Math.floor(Math.random() * 900 + 100)
        });
    newProduct.save();
    res.send('Added a product!');
});
app.get('/getproducts', (req, res) => {
    Product.find()
        .then(products => res.json(products));
});

//Start server
const HOST = '0.0.0.0';
const PORT = 8097;
app.listen(PORT, HOST, () => {
    console.log(`Running on http://${HOST}:${PORT}`);
    console.log('http://localhost:' + PORT + '/addproduct');
    console.log('http://localhost:' + PORT + '/getproducts');
});

• Execute below to run the app & navigate to the urls listed on the console

node index.js

• Stop the Data API once the test is complete

docker compose -f bin/docker-compose.yml down -v

Architecture

High level architecture

Components

• Cassandra Cluster - Apache Cassandra / DataStax Enterprise Cluster as backend database.
• Stargate Coordinator Nodes - Stargate is an open source Data API Gateway for Cassandra. Coordinator is one of the primary components of Stargate which connects the API layer to the backend database. More details can be found here.
• Stargate Data API - Data API is an open source Data API that runs on top of Stargate's coordinator.
• JavaScript Clients that use Mongoose - Mongoose is an elegant mongodb object modeling library for node.js applications. By implementing a driver required by the Mongoose interface to connect to the Data API instead of native mongodb access layer, now a JavaScript client can store/retrieve documents on an Apache Cassandra/DSE backend.

The current implementation of the Data API uses DataStax Enterprise (DSE) as the backend database.

Version compatibility

Component/Library Name	Version
Mongoose	^7.5.0 || ^8.0.0
data-api	1.x
DataStax Enterprise	6.8.x

CI tests are run using the Stargate and Data API versions specified in the api-compatibility.versions file.

Connecting to AstraDB

Here's a quick way to connect to AstraDB using stargate-mongoose driver.

const mongoose = require("mongoose");
const { driver, createAstraUri } = require("stargate-mongoose");

const uri = createAstraUri(
  process.env.ASTRA_DB_API_ENDPOINT,
  process.env.ASTRA_DB_APPLICATION_TOKEN,
  process.env.ASTRA_DB_NAMESPACE // optional
);

mongoose.setDriver(driver);

await mongoose.connect(uri, {
  isAstra: true,
});

And the step-by-step instructions with a sample application can be found here in below guide.

https://docs.datastax.com/en/astra/astra-db-vector/api-reference/data-api-with-mongoosejs.html

Sample Applications

Sample applications developed using stargate-mongoose driver are available in below repository.

https://github.com/stargate/stargate-mongoose-sample-apps

Features

Connection APIs

Operation Name	Description
createDatabase	When flag createNamespaceOnConnect is set to true the keyspace passed on to the mongoose.connect function via the URL, is created automatically
dropDatabase	Drops the database
createCollection	mongoose.model('ModelName',modelSchema) creates a collection as required
dropCollection	model.dropCollection() drops the collection

Collection APIs

Operation Name	Description
countDocuments	Model.countDocuments(filter) returns the count of documents
deleteMany	Model.deleteMany(filter). This API will throw an error when more than 20 records are found to be deleted.
deleteOne	Model.deleteOne(filter, options) options - sort
find	Model.find(filter, projection, options) options - limit, pageState, skip, sort (skip works only with sorting)
findOne	Model.findOne(filter, options) options - sort Example: findOne({}, { sort: { username: -1 } })
findOneAndDelete	Model.findOneAndDelete(filter, options) options - sort
findOneAndReplace	Model.findOneAndReplace(filter, replacement, options) options upsert: (default false) true - if a document is not found for the given filter, a new document will be inserted with the values in the filter (eq condition) and the values in the $set and $setOnInsertoperators. false - new document will not be inserted when no match is found for the given filter -------- returnDocument: (default before) before - Return the document before the changes were applied after - Return the document after the changes are applied
findOneAndUpdate	Model.findOneAndUpdate(filter, update, options) options upsert: (default false) true - if a document is not found for the given filter, a new document will be inserted with the values in the filter (eq condition) and the values in the $set and $setOnInsertoperators. false - new document will not be inserted when no match is found for the given filter -------- returnDocument: (default before) before - Return the document before the changes were applied after - Return the document after the changes are applied
insertMany	Model.insertMany([{docs}], options) In a single call, only 20 records can be inserted. options - ordered
insertOne	Model.insertOne({doc})
updateMany	Model.updateMany(filter, update, options) options upsert: (default false) true - if a document is not found for the given filter, a new document will be inserted with the values in the filter (eq condition) and the values in the $set and $setOnInsertoperators. false - new document will not be inserted when no match is found for the given filter ** This API will throw an error when more than 20 records are found to be updated.
updateOne	Model.updateOne(filter, update, options) options upsert: (default false) true - if a document is not found for the given filter, a new document will be inserted with the values in the filter (eq condition) and the values in the $set and $setOnInsertoperators. false - new document will not be inserted when no match is found for the given filter -------- returnDocument: (default before) before - Return the document before the changes were applied after - Return the document after the changes are applied

Filter Clause

Operator	Description
literal comparison	Equal to. Example: { 'first_name' : 'jim' }
$eq	Example: { 'first_name' : { '$eq' : 'jim' } }
$gt	Example (age > 25): { 'age' : { '$gt' : 25 } }
$gte	Example (age >= 25): { 'age' : { '$gte' : 25 } }
$lt	Example (age < 25): { 'age' : { '$lt' : 25 } }
$lte	Example (age <= 25): { 'age' : { '$lte' : 25 } }
$ne	Example: { 'first_name' : { '$ne' : 'jim' } }
$in	Example: { '_id' : { '$in' : ['nyc', 'la'] } } $in is not supported in non _id columns at the moment
$nin	Example: { 'address.city' : { '$nin' : ['nyc', 'la'] } }
$not	Not supported. Example: { 'first_name' : { '$not' : { '$eq' : 'jim' }}}
$exists	Example: { 'address.city' : { '$exists' : true} }
$all	Array operation. Matches if all the elements of an array matches the given values. Example: { 'tags' : { '$all' : [ 'home', 'school' ] } }
$elemMatch	Not supported. Matches if the elements of an array in a document matches the given conditions. Example: {'goals': { '$elemMatch': { '$gte': 2, '$lt': 10 }}}
$size	Array Operation. Example: { 'tags' : { '$size' : 1 } }
$and (implicit)	Logical expression. Example : { '$and' : [ {first_name : 'jim'}, {'age' : {'$gt' : 25 } } ] }
$and (explicit)	Example : { '$and' : [ {first_name : 'jim'}, {'age' : {'$gt' : 25 } } ] }
$or	Example: { '$or' : [ {first_name : 'jim'}, {'age' : {'$gt' : 25 } } ] }

Projection Clause

Operator	Description
$elemMatch (projection)	Not supported
$slice	Array related operation. Example: { 'tags' : { '$slice': 1 }} returns only the first element from the array field called tags.
$ (projection)	Example: Model.find({}, { username : 1, _id : 0}) - This returns username in the response and the _id field

Sort Clause

Operator	Description
Single Field Sort	Supported
Multi Field Sort	Not supported

Update Clause

Operator	Description
$inc	Example: { '$inc': { 'points' : 5 } }
$min	Example: { 'col': { '$min' : 5 } } if the columns value is greater than 5, it will be updated with 5
$max	Example: { 'col': { '$max' : 50 } } if the columns value is lesser than 50, it will be updated with 50
$rename	Example: { $rename: { '$max' : 50 } } if the columns value is lesser than 50, it will be updated with 50
$set	Example: {'update' : {'$set': {'location': 'New York'} }}
$setOnInsert	Example: {'update' : {'$set': {'location': 'New York'}, '$setOnInsert': {'country': 'USA'} }}
$unset	Example: {'update' : {'$unset': [address.location] }}
$addToSet	Example: {'$addToSet' : {'points': 10}}. This will add 10 to an array called points in the documents, without duplicates (i.e. ll skip if 10 is already present in the array)
$pop	Example: {'$pop' : {'points': 1 }}. This removes the last 1 item from an array called points. -1 will remove the first 1 item.
$pull	Not supported
$push	Example. '$push': {'tags': 'work'}. This pushes an element called work to the array tags
$pullAll	Not supported

Index Operations

Index operations are not supported. There is one caveat for ttl indexes: When adding a document, you can add a ttl option (determined in seconds) that will behave in the similar way to a ttl index. For example, with the collection's client:

import { Client } from 'stargate-mongoose';
// connect to Data API
const client = await Client.connect(process.env.DATA_API_URI);
// get a collection
const collection = client.db().collection('docs');
// insert and expire this document in 10 seconds
await collection.insertOne({ hello: 'world' }, { ttl: 10 });

Aggregation Operations

Aggregation operations are not supported.

Transaction Operations

Transaction operations are not supported.

NodeJS MongoDB Driver Overriding (experimental)

If you have an application that uses the NodeJS MongoDB driver, or a dependency that uses the NodeJS MongoDB driver, it is possible to override its use with the collections package of stargate-mongoose. This makes your application use Data API documents instead of MongoDB documents. Doing so requires code changes in your application that address the features section of this README, and a change in how you set up your client connection.

If your application uses mongodb you can override its usage like so:

In your app's mongodb package.json entry:

"mongodb": "[email protected]",

Then, re-install your dependencies

npm i

Finally, modify your connection so that your driver connects to Data API

import { MongoClient } from 'stargate-mongoose';

// connect to Data API
const client = await MongoClient.connect(process.env.DATA_API_URI);

If you have an application dependency that uses mongodb, you can override its usage like so (this example uses mongoose):

Add an override to your app's package.json (requires NPM 8.3+), also, add `stargate-mongoose as a dependency:

"dependencies": {
    "stargate-mongoose": "^0.2.0-ALPHA-3"
},
"overrides": {
    "mongoose": {
        "mongodb":  "[email protected]"
    }
},

Then, re-install your dependencies

npm i

Finally, modify your dependencies connection so that your driver connects to Data API

import mongoose from 'mongoose';

// connect to Data API
await mongoose.connect(process.env.DATA_API_URI);