document-ts: A TypeScript repository from duluca

A lightweight TypeScript MongoDB ODM with standout convenience features like CollectionFactory and findWithPagination

Read the excerpt from Angular for Enterprise on Understanding DocumentTS on the Wiki.

Looking to containerize MongoDB? Checkout excellalabs/mongo for a fully featured Mongo container (with Auth & SSL) inherited from the official Mongo Docker image and instructions on how to deploy it on AWS.

Major Features

DocumentTS is an ODM (Object Document Mapper) for MongoDB.

connect()

MongoDB async connection harness

It can be a challenge to ensure that database connectivity exists, when writing a fully async web application. connect() makes it easy to connect to a MongoDB instance and makes it safe to be called simultaneously from multiple threads starting up simultaneously.
Document and IDocument

Base Class and Interface to help define your own models
CollectionFactory

Define collections, organize indexes, and aggregate queries alongside collection implementation. Below are the convenience features of a DocumentTS collection
- get collection returns the native MongoDB collection, so you can directly operate on it
```
get collection(): ICollectionProvider<TDocument>
```
- aggregate allows you to run a MongoDB aggregation pipeline
```
aggregate(pipeline: object[]): AggregationCursor<TDocument>
```
- findOne and findOneAndUpdate simplifies the operation of commonly used database functionality, automatically hydrating the models it returns
```
async findOne(filter: FilterQuery<TDocument>, options?: FindOneOptions)
async findOneAndUpdate(
  filter: FilterQuery<TDocument>,
  update: TDocument | UpdateQuery<TDocument>,
  options?: FindOneAndReplaceOption
 ): Promise<TDocument | null>
```
- findWithPagination is by far the best feature of DocumentTS, allowing you to filter, sort, and paginate large collections of data. This function is geared towards use with data tables, so you specify searchable properties, turn off hydration, and use a debug feature to fine-tune your queries.
```
async findWithPagination<TReturnType extends IDbRecord>(
  queryParams: Partial<IQueryParameters> & object,
  aggregationCursorFunc?: Func<AggregationCursor<TReturnType>>,
  query?: string | object,
  searchableProperties?: string[],
  hydrate = true,
  debugQuery = false
): Promise<IPaginationResult<TReturnType>>
```

Quick Start

Supports MongoDB v4+, Mongo Driver 3.3+ and TypeScript 3.7+

Add DocumentTS to your project with npm install document-ts mongodb
Connect to your Mongo database using connect()
Connect will retry connecting to the database 10 times every 2 seconds
- Set connectionRetryWait (in seconds) and connectionRetryMax to modify this behavior
Specify isProd and certFileUri to connect using an SSL certificate

import { connect } from 'document-ts'

async function start() {
  // If isProd is set to true and a .pem file is provided, SSL will be used to connect: i.e. connect(config.mongoUri, isProd, 'server/compose-ca.pem')
  await connect(process.env.MONGO_URI)
}

start()

If you use connect() then you don't have to worry about having your Database Instance initialized during an asynchronous start-up sequence. getDbInstance gives you access to the native MongoDB driver to perform custom functions like creating indexes.

import { getDbInstance } from 'document-ts'

// assuming this is called within an async function
await dbInstance.collection('users').createIndexes([
  {
    key: {
      displayName: 1,
    },
  },
  {
    key: {
      email: 1,
    },
    unique: true,
  },
])

Define the interface for your first model

See tests\user.ts for sample Model implementation

import { IDocument } from 'document-ts'

export interface IUser extends IDocument {
  email: string;
  firstName: string;
  lastName: string;
  role: string;
}

Define the class for your model

See tests\user.ts for sample Model implementation

import { Document } from 'document-ts'

export class User extends Document<IUser> implements IUser {
  static collectionName = 'users'

  private password: string

  public email: string
  public firstName: string
  public lastName: string
  public role: string

  constructor(user?: IUser) {
    super(User.collectionName, user)
  }
  ...
}

Implement getCalculatedPropertiesToInclude() which will ensure that your get properties that are "calculate" on the fly will be serialized when sending the model down to the client, but it will not be saved in the database.

  getCalculatedPropertiesToInclude(): string[]{
      return ['fullName']
  }

Implement getPropertiesToExclude() which will ensure that certain properties like passwords will not be serialized when sending the model down to the client, but it will still be saved in the database.

  getPropertiesToExclude(): string[]{
      return ['password']
  }

Implement the CollectionFactory class, so that you can run Mongo queries without having to call getDbInstance or specify the collection and TypeScript type name every time you run a query. CollectionFactory provides convenience functions like find, findOne, findOneAndUpdate, findWithPagination, and similar, while also handling hydration tasks, such as serializing getters and child documents.

import { CollectionFactory } from 'document-ts'

class UserCollectionFactory extends CollectionFactory<User> {
  constructor(docType: typeof User) {
    super(User.collectionName, docType, ['firstName', 'lastName', 'email'])
  }
}

export let UserCollection = new UserCollectionFactory(User)

CollectionFactory is powerful and flexible. In your custom class, you can implement MongoDB aggregate queries to run advanced join-like queries, geo queries, and whatever MongoDB supports. findWithPagination itself is very powerful and will enable you to implement paginated dashboards with ease.
findWithPagination leverage query parameters for pagination and configuration

export interface IQueryParameters {
  filter?: string
  skip?: number
  limit?: number
  sortKeyOrList?: string | Object[] | Object
  projectionKeyOrList?: string | Object[] | Object
}

Optionally implement toJSON() to customize serialization/hydration behavior or extend ISerializable

  toJSON() {
    let keys = Object.keys(this).concat(['fullAddress', 'localAddress'])
    return Serialize(SerializationStrategy.JSON, this, keys)
  }

Optionally implement toBSON() to customize database serialization behavior or extend ISerializable

  toBSON() {
    let keys = Object.keys(this).concat(['fullAddress', 'localAddress'])
    return Serialize(SerializationStrategy.BSON, this, keys)
  }

To debug the default serialization behavior, implement

  toJSON() {
    // drop a breakpoint here or console.log(this)
    return super.toJSON()
  }

  toBSON() {
    return super.toBSON()
  }

See the Lemon Mart Server sample project for usage - https://github.com/duluca/lemon-mart-server

Goals

Reliable
- Rely on the rock-solid Native Node.js MongoDB drivers
- Don't inject custom code into DB calls without explicit intent by the developer
- Don't hide new MongoDB features, so you don't have to wait for DocumentTS to be updated before you can use them
Optional
- Stays out of the way, so developers can slowly transition
- If performance becomes a concern easily switch to native MongoDB calls for the best performance
Async
- Ensure developers can write simpler and more reliable code by surfacing promises and async/await features
Convenient
- Developers define their own models through simple Interfaces
- Choose fields that you want to automatically hydrate, such as child or related objects
- Serialize calculated fields with every request
- Protect certain fields (like passwords) from serialization, so they aren't accidentally sent across the wire
Promote Good Patterns
- Suggest/enable easy to understand and implement patterns for developers, so their code can scale and remain organized
Prevent Bloat
- Leverage TypeScript types, interfaces, generics and inheritance to achieve development-time certainty of proper database access
- Keep the code smart, readable and lean
- Be very selective about any new features

What It Isn't

Not a full-fledged ODM or ORM replacement and doesn't aspire to be one like Mongoose or Camo. Databases are HARD. MongoDB took many years to mature, and Microsoft has been trying for a really long time to build a reliable ORM with Entity Framework, Mongoose and many other ODMs are ridden with bugs (no offense) when you push them beyond the basics. It takes great resources to deliver a solid data access experience, so with DocumentTS you can develop directly against MongoDB while enjoying some conveniences as you choose.

Inspirations

Although DocumentTS doesn't aspire to replace Mongoose or Camo, it most definitely is inspired by them in the way they've solved certain problems such as hydration. Check out the source code for those projects here:

Mongoose
Camo

Building This Project