Introduction to the newline guide to building and deploying a REST API using Deno.

Welcome to Build and Deploy a REST API with Deno!

In this lesson, we will install Deno and create our setup for the rest of the course.

How to Install Deno Locally and in Docker Containers

Introduction

Deno vs Node: A Deno Fresh Start to JavaScript Runtimes

We will look at the command line interface for running Deno

How to Use the Deno REPL Command Line Interface

In this lesson, we will look at general usage of Deno by running a file and also introduce the config file

Executing TypeScript Files With Deno With the Config File

Learn how to use the Deno standard library and other external modules with URL imports

Deno vs Golang's Standard Library

How to Start a Deno Server and Set Permissions

You will learn how to make a simple token authentication in Deno

How to Build a Token Authentication Server With Deno

Getting Started with Deno

Using Databases with Deno

How to Use PostgreSQL in Docker with Deno

How to Integrate MySQL or MariaDB With Deno

How to Add SQLite to Deno with deno-sqlite

Connect to a database

Introduction to database migrations - what, why and how

What is Database Migration? The Strategy to Migrate Data

Get started with database migrations using Nessie

How to Add Nessie to Deno to Migrate Databases

We will go through creating a migration using the Nessie CLI

How to Create a New Database Migration With Deno Nessie

We will look at how to write a migration script where we create a table

Edit a Deno Nessie Database Migration to Catch Errors

We will run our migration and learn how to roll back the migration

How to Run and Roll Back a Database Migration With Deno Nessie

We will create a seed file using the Nessie CLI

How to Create a Database Seed File With Deno Nessie

We will run our seed file and populate our database

How to Import JSON Data and Seed a Database With Deno Nessie

Database migrations

Introduction to the concept of REST APIs and why we use them

REST API Introduction

We will go through the thought process of planning an application.

Best Practices for a REST API: Endpoints and Database Schema

We will create the initial project files and set up our environment

How to Prepare Deno to Build a REST API

We will create the router and look at how to test our endpoints

How to Build REST API Endpoints in Deno

It's time to add Nessie! We will now initialize Nessie in our project and also spin up our development database

Integrate a PostgreSQL Database With Nessie Into a Deno API

Time to connect our models to our database!

Building a MySQL or PostgreSQL Database Pool in Deno

In this step, we will create our first data model and create a corresponding table using a migration

Migrating Data to PostgreSQL and Modeling Data in Deno

Creating the methods to access the database

Build API Calls That Generate SQL Queries in Deno

Creating the server endpoints for the Duck namespace

Create a Database Controller in Deno

We will look at how to set up authentication and further authorize the requests

How to Authenticate and Authorize REST API Requests in Deno

We will create two new tables using a migration and create the equivalent models

Create New Tables and Models With Deno Database Migration

We will create our Quack service and controller

Add List, Create, Read, Update, and Delete Requests in Deno

The last stretch! We will create our final namespace, Crumb, and connect it to our other models and our router

How to Build Many-to-Many Namespace Connections in Deno

Now that our application is completed, we will consider some improvements and discuss alternative approaches.

API Afterthoughts

REST API

We will introduce deployments and what we will learn in this module

Deployment Introduction

We will create a Docker image for our application

How to Create a Docker Image to Containerize an Application

You will learn how to create a basic CI file for GitHub and GitLab, and run tests on them

How to Create Continuous Integration Files for GitHub and GitLab

We will deploy our application to a server using CI, SSH and git.

How to Deploy a Deno API Server With the Digital Ocean API

We will deploy our application to Digital Ocean Apps. This is by far the easiest (but most costly) approach.

How to Deploy a Deno Docker Image to Digital Ocean

Similar to how we did it in lesson 1, we will be deploying using CI and SSH, but instead of git, we will be using Docker.

Deploy with Docker

Deployment

In this course, we will learn how to work with [Deno](https://deno.land) to create a REST API using TypeScript with the web framework [Oak](https://github.com/oakserver/oak) and the migration tool [Nessie](https://github.com/halvardssm/deno-nessie). 
Deno is a JavaScript runtime created by [Ryan Dahl](https://tinyclouds.org), the same creator for Node, as an attempt to simplify the JavaScript landscape by strictly following the WebAPI specs. 
If you are interested in learning more about the motivation behind Deno, you can watch this video where Ryan Dahl explains "[10 Things I Regret About Node.js](https://www.youtube.com/watch?v=M3BM9TB-8yA)"

We will explore how to create an application using the powers of TypeScript, and learn how to traverse the Deno ecosystem which is much like the native web ecosystem. 
There are 5 modules that will take you from installing Deno and running your first lines of code, to working with web frameworks and databases, and in the end, deploying your application to a web server using a deployment pipeline.

Throughout this course, you will experience all the different aspects of Deno which you will need to create your own application in the future. 
And from this course, you will get a good feeling for if Deno is ready to be your next default for new projects.

In this course, we will learn how to create and deploy a REST API with Deno from scratch.

Build and deploy a REST API with Deno

Integrating databases into your application

Importance of database migrations in production applications

Deploying Deno applications to production using CICD

In the previous steps, we created a basic Duck interface and connected the
database pool to our application. In this step, we will be adding a way for us
to store and fetch the data we need.

When working on large applications, I usually structure my namespaces as such:

There are other files and conventions, like guards to validate incoming
requests, but we will not discuss them in this guide. In our case we will merge
the 

As with everything else, this is a design decision for our application. If you
really wanted to, you could write everything in one file, but that wouldn't be
very nice to work with as the file would quickly become quite big and
convoluted. But remember, it always depends on your application, so in some
cases it might be better, cleaner, or more efficient to have everything in one
file. In short, it depends!

To get started on the service file, we will add some new imports to our

As we are also using the context of the Oak app to get our postgres client, we
should add a helper function to get this out of the context.

export function getClient(ctx: AppRouterContext): Promise<PoolClient> {
  return ctx.app.state.pool.connect();
}

, it is simply a clone of the state, while

 is the actual global state where you can access the 

As you can see, it would be quite tiresome to write this long chain every time
so with this function, we can quicly return a pool client for us to use.

Lastly before we begin the core of this lesson, we will create a new file

, and add the following imports to it. We will be needing
them for the functions we will create.

import { httpErrors } from "oak";
import { cuid } from "cuid";
import { hash as hashPassword } from "bcrypt";
import type { AppRouterContext } from "../../main.ts";
import { getClient } from "../utils.ts";
import { Duck, DuckKeys } from "./model.ts";

To create a Duck, we need to create a SQL query and also do some data
validation. A Duck also contains a password, and as every good programmer knows,
this shuld never be stored as clear text, but be hashed (not

).
We will discuss password hashing and authentication in two lessons. We also want
to generate an id for every new entry.

export async function createDuck(
  ctx: AppRouterContext,
  data: Partial<Duck>,
): Promise<Duck> {
  if (!data.username) {
    throw new httpErrors.BadRequest("Username is missing from data");
  }
  if (!data.password) {
    throw new httpErrors.BadRequest("Password is missing from data");
  }

  const id = cuid();
  const hashedPassword = await hashPassword(data.password);

  const client = await getClient(ctx);

  const result = await client.queryObject<Duck>({
    text:
      `INSERT INTO ${DuckKeys.TableName} (${DuckKeys.ColId}, ${DuckKeys.ColUsername}, ${DuckKeys.ColPassword}) VALUES ($1, $2, $3) RETURNING *`,
    args: [id, data.username, hashedPassword],
  });

  client.release();

  return result.rows[0] as Duck;
}

We start our function by verifying the incoming data, and as we only have two
fields to worry about, we can make it a bit simple for ourselves. We check if
the username and password is set, and if not, we throw an http error which will
be caught by our error middleware.

---
title: Build API Calls That Generate SQL Queries in Deno
description: Creating the methods to access the database
code: https://gitlab.com/fullstackio/books/newline-course-apps/halvard-morstad-build-and-deploy-a-rest-api-with-deno-app/-/tree/main/app_with_sqlite_db%20(updated)/module_4.7?ref_type=heads
---

## Duck Service

In the previous steps, we created a basic Duck interface and connected the
database pool to our application. In this step, we will be adding a way for us
to store and fetch the data we need.

When working on large applications, I usually structure my namespaces as such:

-   A `Model` to hold information about the database table and its structure
-   A `DAO`
    ([Data Access Object](https://en.wikipedia.org/wiki/Data_access_object)) to be
    the connection point between the application and the database. This is where
    the queries are placed.
-   A `Service` serving as an entrypoint to the namespace. Usually contains
    distinkt namespace logic, and works as a bridge between the `DAO` and other
    code.
-   A `Controller` to be the entrypoint connected to the endpoints. Uses other
    services to provide logic and data.

There are other files and conventions, like guards to validate incoming
requests, but we will not discuss them in this guide. In our case we will merge
the `DAO` and the `Service` into one file to keep our app simple.

> As with everything else, this is a design decision for our application. If you
> really wanted to, you could write everything in one file, but that wouldn't be
> very nice to work with as the file would quickly become quite big and
> convoluted. But remember, it always depends on your application, so in some
> cases it might be better, cleaner, or more efficient to have everything in one
> file. In short, it depends!

To get started on the service file, we will add some new imports to our
`import_map.json`.

```typescript { actualFilename=../../code/module_4.7/import_map.json endChar=370 endLine=11 startChar=0 startLine=1 statedFilename=import_map.json url=https&#x3A;//gitlab.com/fullstackio/books/build-a-rest-api-with-deno-and-postgresql/module_05/lesson_07/.../code/module_4.7/import_map.json }
{
  "imports": {
    "oak": "https://deno.land/x/oak@v11.1.0/mod.ts",
    "dotenv": "https://deno.land/std@0.155.0/dotenv/mod.ts",
    "postgres": "https://deno.land/x/postgres@v0.16.1/mod.ts",
    "nessie": "https://deno.land/x/nessie@2.0.7/mod.ts",
    "cuid": "https://deno.land/x/cuid@v1.0.0/index.js",
    "bcrypt": "https://deno.land/x/bcrypt@v0.4.0/mod.ts"
  }
}
```

As we are also using the context of the Oak app to get our postgres client, we
should add a helper function to get this out of the context.

Add this to our `src/utils.ts` file.

```typescript { actualFilename=../../code/module_4.7/src/utils.ts endLine=36 startLine=34 statedFilename=src/utils.ts url=https&#x3A;//gitlab.com/fullstackio/books/build-a-rest-api-with-deno-and-postgresql/module_05/lesson_07/.../code/module_4.7/src/utils.ts#L34 }
export function getClient(ctx: AppRouterContext): Promise<PoolClient> {
  return ctx.app.state.pool.connect();
}
```

> When you access `ctx.state`, it is simply a clone of the state, while
> `ctx.app.state` is the actual global state where you can access the `pool`.

As you can see, it would be quite tiresome to write this long chain every time
so with this function, we can quicly return a pool client for us to use.

Lastly before we begin the core of this lesson, we will create a new file
`src/duck/service.ts`, and add the following imports to it. We will be needing
them for the functions we will create.

```typescript { actualFilename=../../code/module_4.7/src/duck/service.ts endLine=6 startLine=1 statedFilename=src/duck/service.ts url=https&#x3A;//gitlab.com/fullstackio/books/build-a-rest-api-with-deno-and-postgresql/module_05/lesson_07/.../code/module_4.7/src/duck/service.ts }
import { httpErrors } from "oak";
import { cuid } from "cuid";
import { hash as hashPassword } from "bcrypt";
import type { AppRouterContext } from "../../main.ts";
import { getClient } from "../utils.ts";
import { Duck, DuckKeys } from "./model.ts";
```

### Create a Duck

To create a Duck, we need to create a SQL query and also do some data
validation. A Duck also contains a password, and as every good programmer knows,
this shuld never be stored as clear text, but be hashed (not
[encrypted](https://stackoverflow.com/questions/326699/difference-between-hashing-a-password-and-encrypting-it)).
We will discuss password hashing and authentication in two lessons. We also want
to generate an id for every new entry.

Here is our implementation for the `createDuck` function:

```typescript { actualFilename=../../code/module_4.7/src/duck/service.ts endLine=33 startLine=8 statedFilename=src/duck/service.ts url=https&#x3A;//gitlab.com/fullstackio/books/build-a-rest-api-with-deno-and-postgresql/module_05/lesson_07/.../code/module_4.7/src/duck/service.ts#L8 }
export async function createDuck(
  ctx: AppRouterContext,
  data: Partial<Duck>,
): Promise<Duck> {
  if (!data.username) {
    throw new httpErrors.BadRequest("Username is missing from data");
  }
  if (!data.password) {
    throw new httpErrors.BadRequest("Password is missing from data");
  }

  const id = cuid();
  const hashedPassword = await hashPassword(data.password);

  const client = await getClient(ctx);

  const result = await client.queryObject<Duck>({
    text:
      `INSERT INTO ${DuckKeys.TableName} (${DuckKeys.ColId}, ${DuckKeys.ColUsername}, ${DuckKeys.ColPassword}) VALUES ($1, $2, $3) RETURNING *`,
    args: [id, data.username, hashedPassword],
  });