Mongoose

DaVinci, per se is database agnostic and doesn't have an opinion on how your API persists the data.

However, it provides a set of utilities that allow you to use MongoDB in conjunction with it.

More specifically, it provides:

• Functions to generate Mongoose models starting from a decorated typescript class.
• Functions to register some write/read hooks.

Installation

npm i --save @davinci/mongoose mongoose

Define the Schema

The schema is, once again, defined using decorators applied to methods of a class. It means that the same schema can be reused for different things, for example as mongoose and openapi schemas.

import { Schema } from 'mongoose';
import { mgoose } from '@davinci/mongoose';

class CustomerPhone {
    @mgoose.prop()
    number: string;

    @mgoose.prop()
    isPrimary: boolean;
}

// we can define compound indexes using the index decorator
// alternatively, they can be specified at the field level
@mgoose.index({ firstname: 1, lastname: 1 }, { unique: true })
export default class Customer {
    @mgoose.prop({ required: true })
    firstname: string;

    @mgoose.prop({ required: true })
    lastname: string;

    @mgoose.prop({ enum: ['member', 'admin'] })
    role: string;

    @mgoose.prop({ type: [CustomerPhone] })
    phones: CustomerPhone[];

    @mgoose.prop({ type: Schema.Types.ObjectId })
    @mgoose.populate({
        name: 'account',
        opts: { ref: 'Account', foreignField: '_id', justOne: true }
    })
    accountId: string;
}

Create the Model

Using the generateModel function of the @davinci/mongoose package, we can generate a Mongoose Model by passing as argument the schema defined above.

import { mgoose } from '@davinci/mongoose';
import CustomerSchema from './CustomerSchema';

const { generateModel } = mgoose;

const Customer = generateModel<CustomerSchema>(CustomerSchema, 'customer', 'customers');

export default Customer;

Hooks

@davinci/mongoose provides six different hooks: beforeRead, afterRead, beforeWrite, afterWrite, beforeDelete and afterDelete.\ Under the hood, they use Mongoose Middlewares

• beforeRead / afterRead\ It gets triggered before/after executing any find/fetch operation. Under the hood it register the following Mongoose Middlewares: find, findOne, findOneAndDelete, findOneAndRemove, findOneAndUpdate, deleteMany, update, updateOne, updateMany
• beforeWrite / afterWrite\ It gets triggered before/after executing any save/persist operation. Under the hood it register the following Mongoose Middlewares: findOneAndUpdate, save, update, updateMany
• beforeDelete / afterDelete\ It gets triggered before/after executing any delete operation. Under the hood it register the following Mongoose Middlewares: deleteMany, findOneAndDelete, findOneAndRemove

import { Document, model } from 'mongoose';
import { mgoose } from '@davinci/mongoose';
import { httpErrors } from '@davinci/core';
import CustomerSchema from './CustomerSchema';
import { afterDelete } from './hooks';

const { generateSchema, beforeRead, beforeWrite, afterDelete } = mgoose;

const schema = generateSchema(CustomerSchema);

beforeRead<Context>(schema, ({ query, context }) => {
    // inject accountId before persisting into DB
    if (!context) return;

    const currentQuery = query.getQuery();
    query.setQuery({ ...currentQuery, accountId: context.accountId });
});

beforeWrite<Context, CustomerSchema>(schema, ({ query, doc, context }) => {
    // inject accountId before persisting into DB
    if (!context) return;

    // required check for atomic operations
    if (doc) {
        doc.accountId = context.accountId;
    } else {
        // @ts-ignore
        query.setUpdate({
            ...query.getUpdate(),
            accountId: context.accountId
        });
    }
});

afterDelete<Context>(schema, ({ doc }) => {
    if (doc) {
        // perform some cleanup
    }
});

const Customer = model<CustomerSchema & Document>('Customer', schema, 'customers');

export default Customer;