Your First Application

Contents

  1. Guestbook Application
  2. Quickstart Github Codespaces
  3. Installing JDK
  4. Installing a Build Tool
  5. Creating a new application
  6. Anatomy of a Kit application
  7. The Source Directory
  8. The Env Directory
  9. The Test Directory
  10. The Resources Directory
  11. Starting Our Server
  12. System
  13. Kit Modules
  14. Accessing The Database
  15. Exposing Database Queries in the Router Component
  16. Creating a controller for the guestbook
  17. Creating Pages and Handling Form Input
  18. Adding some tests
  19. Packaging the application

Guestbook Application

This tutorial will guide you through building a simple guestbook application using Kit. The guestbook allows users to leave a message and to view a list of messages left by others. The application will demonstrate the basics of HTML templating, database access, and project architecture.

If you don't have a preferred Clojure editor already, then it's recommended that you use Calva to follow along with this tutorial.

Quickstart Github Codespaces

You can follow the tutorial using browser based environment by starting a dev container from the following repo https://github.com/kit-clj/playground/.

Installing JDK

Clojure runs on the JVM and requires a copy of JDK to be installed. If you don't have JDK already on your system then OpenJDK is recommended and can be downloaded here. Note that Kit requires JDK 11 or greater to work with the default settings. Alternatively, follow the instructions for installing packages on your system.

Installing a Build Tool

For building and running a project, Kit supports Clojure Deps and CLI. Note that Kit requires tools.build version 1.10.3.933 or later.

To install Clojure CLI, follow the steps below depending on your operating system.

MacOS

brew install clojure/tools/clojure

Linux

curl -L -O https://github.com/clojure/brew-install/releases/latest/download/posix-install.sh
chmod +x posix-install.sh
sudo ./posix-install.sh

For both macOS and Linux, you will need clj-new installed as follows:

clojure -Ttools install com.github.seancorfield/clj-new '{:git/tag "v1.2.404"}' :as clj-new

For information on customization options, for example on how to change your install location, see the official docs here.

Creating a new application

Once you have the Clojure CLI installed, you can run the following commands in your terminal to initialize your application:

clojure -Tclj-new create :template io.github.kit-clj :name kit/guestbook
cd guestbook

The above will create a new project, named kit/guestbook, based on the kit-clj template.

Anatomy of a Kit application

The newly created application has the following structure:

├── Dockerfile
├── README.md
├── build.clj
├── deps.edn
├── env
│   ├── dev
│   │   ├── clj
│   │   │   ├── user.clj
│   │   │   └── kit
│   │   │       └── guestbook
│   │   │           ├── dev_middleware.clj
│   │   │           └── env.clj
│   │   └── resources
│   │       └── logback.xml
│   └── prod
│       ├── clj
│       │   └── kit
│       │       └── guestbook
│       │           └── env.clj
│       └── resources
│           └── logback.xml
├── kit.edn
├── kit.git-config.edn
├── project.clj
├── resources
│   └── system.edn
├── src
│   └── clj
│       └── kit
│           └── guestbook
│               ├── config.clj
│               ├── core.clj
│               └── web
│                   ├── controllers
│                   │   └── health.clj
│                   ├── handler.clj
│                   ├── middleware
│                   │   ├── core.clj
│                   │   ├── exception.clj
│                   │   └── formats.clj
│                   └── routes
│                       ├── api.clj
│                       └── utils.clj
└── test
    └── clj
        └── kit
            └── guestbook
                └── test_utils.clj

Let's take a look at what the files in the root folder of the application do:

  • deps.edn - used to manage the project configuration and dependencies by deps
  • build.clj - used to manage the build process by Clojure CLI tools
  • Dockerfile - used to facilitate Docker container deployments
  • README.md - where documentation for the application is conventionally put
  • resources/system.edn - used for system configuration
  • .gitignore - a list of assets, such as build-generated files, to exclude from Git

The Source Directory

All our code lives under the src/clj folder. Since our application is called kit/guestbook, this is the root namespace for the project. Let's take a look at all the namespaces that have been created for us.

guestbook

  • config.clj - this is the place where your system.edn is read in to create an integrant configuration map
  • core.clj - this is the entry point for the application that contains the logic for starting and stopping the server

guestbook.web

The web namespace is used to define the edges of your application that deal with server communication, such as receiving HTTP requests and returning responses.

  • handler.clj - defines the entry points for routing and request handling.

guestbook.web.controllers

The controllers namespace is where the controllers are located. By default, a healthcheck controller is created for you. When you add more controllers you should create namespaces for them here.

  • healthcheck.clj - default controller that returns basic statistics about your server

guestbook.web.middleware

The middleware namespace consists of functions that implement cross-cutting functionality such as session management, coercion, etc. These functions can be wrapped around groups of routes to provide common functionality.

  • core.clj - an aggregate of default middlewares and environment-specific middleware
  • exception.clj - logic for classifying exceptions within controllers and returning appropriate HTTP responses
  • formats.clj - handles coercion of requests data to Clojure data structures, and response data back to strings

guestbook.web.routes

The routes namespace is where the HTTP routes are defined. By default /api routes are created for you. When you add more routes you should create namespaces for them here.

  • api.clj - a namespace that routes (default = /api) with Swagger UI
  • utils.clj - general purpose helper functions for getting data from requests

The Env Directory

Environment-specific code and resources are located under the env/dev, env/test, and the env/prod paths. The dev configuration will be used during development and testing, test during testing, while the prod configuration will be used when the application is packaged for production.

The Dev Directory

Any source code that's meant to be used during development time is placed in the dev/clj. The following namespaces will be generated by default when the project is created:

  • user.clj - a utility namespace for any code you wish to run during REPL development. You start and stop your server from here during development.
  • guestbook/env.clj - contains the development configuration defaults
  • guestbook/dev_middleware.clj - contains middleware used for development that should not be compiled in production

Resources that are used during development are placed in dev/resources. By default this folder will contain logback configuration tuned for development:

  • logback.xml file used to configure the development logging profile

Similarly, testing configuration is placed in the test/resources:

  • logback.xml file used to configure the test logging profile

The Prod Directory

This directory is a counterpart of the dev directory, and contains versions of namespaces and resources that will be used when the application is built for production. The prod/clj folder will contain the following namespace when the project is created:

  • guestbook/env.clj namespace with the production configuration

Meanwhile, prod/resources will contain logback configuration tuned for production use:

  • logback.xml - default production logging configuration

The Test Directory

Here is where we put tests for our application. Some test utilities have been provided.

The Resources Directory

This is where we put all the resources that will be packaged with our application. Anything in the public directory under resources will be served to the clients by the server.

Starting Our Server

Your REPL is your best friend in Clojure. Let's start our local development REPL by running

clj -M:dev

You can alternatively use clj -M:dev:cider if you intend to connect to the REPL from Emacs and CIDER (this also works with VS Code and Calva). Or do clj -M:dev:nrepl if you want to start nREPL but don't need the CIDER middleware. See the Guestbook example README for more info on connecting the REPL with a Clojure editor.

Once you are in the REPL, you can start the system by running a command provided in env/dev/user.clj

(go) ;; To start the system

(halt) ;; To stop the system

(reset) ;; To refresh the system after making code changes

To confirm that your server is running, visit http://localhost:3000/api/health.

System

System resources such as HTTP server ports, database connections are defined in the resources/system.edn file. For example, this key defines HTTP server configuration such as the host, port, and HTTP handler:

:server/http
 {:port #long #or [#env PORT 3000]
  :host #or [#env HTTP_HOST "0.0.0.0"]
  :handler #ig/ref :handler/ring}

Now that we've looked at the structure of the default project, let's see how we can add some additional functionality via modules.

Kit Modules

Kit modules consist of templates that can be used to inject code and resources into a Kit project. The modules are defined in the kit.edn file. By default, the configuration will point to the official module repository:

:modules   {:root         "modules"
            :repositories [{:url  "git@github.com:kit-clj/modules.git"
                            :tag  "master"
                            :name "kit-modules"}]}

Since our application needs to serve some HTML content, let's add the official HTML module. In your REPL, you can execute the following:

;; This will download the official Kit modules from git
(kit/sync-modules)

;; This will list the available modules
(kit/list-modules)
;; =>
;; :kit/html - adds support for HTML templating using Selmer
;; :kit/sql - adds support for SQL. Available profiles [ :postgres :sqlite ]. Default profile :sqlite
;; :kit/cljs - adds support for cljs using shadow-cljs
;; nil

;; To be able to serve HTML pages, install the :html module
(kit/install-module :kit/html)
;; =>
;; updating file: resources/system.edn
;; injecting
;; path: [:reitit.routes/pages]
;; value: {:base-path "", :env #ig/ref :system/env}
;; updating file: deps.edn
;; injecting
;; path: [:deps selmer/selmer]
;; value: #:mvn{:version "1.12.44"}
;; injecting
;; path: [:deps ring/ring-defaults]
;; value: #:mvn{:version "0.3.3"}
;; injecting
;; path: [:deps luminus/ring-ttl-session]
;; value: #:mvn{:version "0.3.3"}
;; updating file: src/clj/kit/guestbook/core.clj
;; applying
;; action: :append-requires
;; value: [[kit.guestbook.web.routes.pages]]
;; html installed successfully!
;; restart required!
;; nil

We can see from the output of the kit/install-module that we need to restart our REPL. Let's do that. Once we are up again, we can test if our module is installed correctly by starting up the server with (go) and navigating to localhost:3000.

HTML templates

The module generated the following files under the resources/html directory:

  • home.html - home page
  • error.html - error page template

This directory is reserved for HTML templates that represent the application pages.

The module also generated the namespace kit.guestbook.web.pages.layout which helps you render HTML pages using Selmer templating engine

Routing

The module also helped generate some routes for us under kit.guestbook.web.routes.pages. See routing documentation for more details.

Adding a database

Similarly to the way we installed the HTML module, we can add a SQL module with SQLite called :kit/sql. The default profile includes SQLite out of the box, but if we wanted to be explicit we could also write (kit/install-module :kit/sql {:feature-flag :sqlite})

(kit/install-module :kit/sql)
;; updating file: resources/system.edn
;; injecting
;;  path: [:db.sql/connection]
;;  value: #profile {:dev {:jdbc-url "jdbc:sqlite:_dev.db"}, :test {:jdbc-url "jdbc:sqlite:_test.db"}, :prod {:jdbc-url #env JDBC_URL}}
;; injecting
;;  path: [:db.sql/query-fn]
;;  value: {:conn #ig/ref :db.sql/connection, :options {}, :filename "sql/queries.sql"}
;; injecting
;;  path: [:db.sql/migrations]
;;  value: {:store :database, :db {:datasource #ig/ref :db.sql/connection}, :migrate-on-init? true}
;; updating file: deps.edn
;; injecting
;;  path: [:deps io.github.kit-clj/kit-sql]
;;  value: #:mvn{:version "0.1.0"}
;; injecting
;;  path: [:deps org.xerial/sqlite-jdbc]
;;  value: #:mvn{:version "3.34.0"}
;; updating file: src/clj/kit/guestbook/core.clj
;; applying
;;  action: :append-requires
;;  value: [[kit.edge.db.sql]]
;; sql installed successfully!
;; restart required!

Restart again and create your first database migration.

(migratus.core/create
  (:db.sql/migrations state/system)
  "add-guestbook-table")

This will generate two files under your resources/migrations directory. They will look something like this, but with a different prefix:

20211109173842-add-guestbook-table.up.sql
20211109173842-add-guestbook-table.down.sql

Kit uses Migratus for migrations. Migrations are managed using up and down SQL files. The files are conventionally versioned using the date and will be applied in order of their creation.

Let's add some content to create our messages table under the <date>-add-guestbook-table.up.sql file

CREATE TABLE guestbook
(id INTEGER PRIMARY KEY AUTOINCREMENT,
 name VARCHAR(30),
 message VARCHAR(200),
 timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP);

The guestbook table will store all the fields describing the message, such as the name of the commenter, the content of the message, and a timestamp. Next, let's replace the contents of the <date>-add-guestbook-table.down.sql file accordingly:

DROP TABLE IF EXISTS guestbook;

Migrations will be run automatically using the configuration found in system.edn:

:db.sql/migrations {:store :database,
                     :db {:datasource #ig/ref :db.sql/connection},
                     :migrate-on-init? true}

Accessing The Database

The SQL queries are found in the resources/sql folder.

  • queries.sql - defines the SQL queries and their associated function names

Let's take a look at the queries.sql template file. Its contents should look as follows:

-- :name create-user! :! :n
-- :doc creates a new user record
INSERT INTO users
(id, first_name, last_name, email, pass)
VALUES (:id, :first_name, :last_name, :email, :pass)

-- :name update-user! :! :n
-- :doc update an existing user record
UPDATE users
SET first_name = :first_name, last_name = :last_name, email = :email
WHERE id = :id

-- :name get-user :? :1
-- :doc retrieve a user given the id.
SELECT * FROM users
WHERE id = :id

The file initially contains some placeholder queries to help remind you of basic SQL syntax. As we can see, each function is defined using the comment that starts with -- :name followed by the name of the function. The next comment provides the doc string for the function and finally we have the body that's plain SQL. For full documentation of this syntax you can view the HugSQL documentation. The parameters are denoted using : notation. Let's replace the existing queries with some of our own:

-- :name save-message! :! :n
-- :doc creates a new message
INSERT INTO guestbook
(name, message)
VALUES (:name, :message)

-- :name get-messages :? :*
-- :doc selects all available messages
SELECT * FROM guestbook

Now that our model is all set up, let's reload the application, and test our queries in the REPL:

(reset)

(def query-fn (:db.sql/query-fn state/system))

(query-fn :save-message! {:name      "m1"
                          :message   "hello world"})
;; => 1

(query-fn :get-messages {})
;; => [{:id 1, :name "m1", :message "hello world", :timestamp 1636480432353}]

In this example, the newly defined query-fn function allows you to execute the SQL functions you defined in queries.sql. It achieves this using the :db.sql/query-fn component that comes with kit-sql (a dependency of kit/sqlite you installed).

As you can see, query-fn takes two arguments: name of the SQL query function to call, and a map of parameters required by that function.

For more information on how components like :db.sql/query-fn work, see Accessing Components.

Exposing Database Queries in the Router Component

Now that we've added the queries, we'll need to update resources/system.edn to make these queries available in the page router component. To do this, add a :query-fn key in the component definition as follows:

:reitit.routes/pages
{:base-path "",
 :query-fn #ig/ref :db.sql/query-fn
 :env      #ig/ref :system/env}

The key references the :db.sql/query-fn component which is responsible for instantiating query functions using the template found in the resources/sql/queries.sql file:

:db.sql/query-fn
{:conn #ig/ref :db.sql/connection,
 :options {},
 :filename "sql/queries.sql"}

Like in the REPL example towards the end of the Accessing the Database section, the :db.sql/query-fn component comes from kit-sql. Unlike in that example:

  • we pass specific database connection information
  • we only make available the query functions from one specific file.

For more information on how components like :db.sql/query-fn work, see Accessing Components.

Creating a controller for the guestbook

We'll create a new controller that will be responsible for saving new messages in the database. Let's create a new namespace called kit.guestbook.web.controllers.guestbook, this namespace should be placed in a corresponding file called guestbook.clj under the src/clj/kit/guestbook/web/controllers/ folder. We'll add the following content to the namespace:

(ns kit.guestbook.web.controllers.guestbook
  (:require
   [clojure.tools.logging :as log]   
   [ring.util.http-response :as http-response]))

(defn save-message!
  [{:keys [query-fn]} {{:strs [name message]} :form-params :as request}]
  (log/debug "saving message" name message)
  (try
    (if (or (empty? name) (empty? message))
      (cond-> (http-response/found "/")
        (empty? name)
        (assoc-in [:flash :errors :name] "name is required")
        (empty? message)
        (assoc-in [:flash :errors :message] "message is required"))
      (do
        (query-fn :save-message! {:name name :message message})
        (http-response/found "/")))
    (catch Exception e
      (log/error e "failed to save message!")
      (-> (http-response/found "/")
          (assoc :flash {:errors {:unknown (.getMessage e)}})))))

As you can see, the namespace contains a save-message! function that executes the query to add a new message to the guestbook table. The query is accessed from the first argument which is the Integrant system map that's passed to the handler. The query-fn key contains a map of the query functions. The names of these functions are inferred from the -- :name comments in the SQL templates found in the resources/sq/queries.sql file.

Our function will grab the form-params key from the request that contains the form data and attempt to save the message in the database. The controller will redirect back to the home page, and if set any errors as a flash session on the response.

Creating Pages and Handling Form Input

The routes for the HTML pages are defined in the kit.guestbook.web.routes.pages namespace. Let's reference our kit.guestbook.web.controllers.guestbook and kit.guestbook.web.routes.utils namespaces in the namespace declaration.

(ns kit.guestbook.web.routes.pages
  (:require
    ...    
    [kit.guestbook.web.controllers.guestbook :as guestbook]))

We can now add the logic for rendering the messages from the database by updating the home-page handler function to look as follows:

(defn home [{:keys [query-fn]} {:keys [flash] :as request}]
  (layout/render request "home.html" {:messages (query-fn :get-messages {})
                                      :errors (:errors flash)}))

The function now renders the home.html template, and passes into it the messages from the database (using the :messages key), and any errors (using the :errors key).

Finally, we'll add the /save-message route in the page-routes function. This route will pass the request to the guestbook/save-message! function we defined above when the form post happens:

(defn page-routes [opts]
  [["/" {:get (partial home opts)}]
   ["/save-message" {:post (partial guestbook/save-message! opts)}]])

Now that we have our controllers set up, let's open home.html template located in the resources/html directory. Currently, it simply renders a static page. We'll update our content div to iterate over the messages and print each one in a list:

<div class="content container">
  <div class="columns">
    <div class="column">
      <h3>Messages</h3>
      <ul class="messages">
        {% for item in messages %}
        <li>
          <time>{{item.timestamp}}</time>
          <p>{{item.message}}</p>
          <p> - {{item.name}}</p>
        </li>
        {% endfor %}
      </ul>
    </div>
  </div>
</div>

As you can see above, we use a for iterator to walk through the messages. Since each message is a map with the message, name, and timestamp keys, we can access them by name.

Finally, we'll create a form to allow users to submit their messages. We'll populate the name and message values if they're supplied and render any errors associated with them. Note that the forms also uses the csrf-field tag that's required for cross-site request forgery protection.

<div class="columns">
  <div class="column">
      {% if errors.unknown %}
      <div class="notification is-danger">{{errors.unknown}}</div>
      {% endif %}
      <form method="POST" action="/save-message">
          {% csrf-field %}
          <p>
              <label>
              Name:
              <input class="input" type="text" name="name" value="{{name}}" />
              </label>
          </p>
          {% if errors.name %}
          <div class="notification is-danger">{{errors.name}}</div>
          {% endif %}
          <p>
              <label>
              Message:
              <textarea class="textarea" name="message">{{message}}</textarea>
              </label>
          </p>
          {% if errors.message %}
          <div class="notification is-danger">{{errors.message}}</div>
          {% endif %}
          <input type="submit" class="button is-primary is-outlined has-text-dark" value="comment" />
      </form>
  </div>
</div>

Our final content div should look as follows:

<div class="content container">
  <div class="columns">
      <div class="column">
          <h3>Messages</h3>
          <ul class="messages">
              {% for item in messages %}
              <li>
                  <time>{{item.timestamp}}</time>
                  <p>{{item.message}}</p>
                  <p> - {{item.name}}</p>
              </li>
              {% endfor %}
          </ul>
      </div>
  </div>
  <div class="columns">
      <div class="column">
          {% if errors.unknown %}
          <div class="notification is-danger">{{errors.unknown}}</div>
          {% endif %}
          <form method="POST" action="/save-message">
              {% csrf-field %}
              <p>
                  <label>
                  Name:
                  <input class="input" type="text" name="name" value="{{name}}" />
                  </label>
              </p>
              {% if errors.name %}
              <div class="notification is-danger">{{errors.name}}</div>
              {% endif %}
              <p>
                  <label>
                  Message:
                  <textarea class="textarea" name="message">{{message}}</textarea>
                  </label>
              </p>
              {% if errors.message %}
              <div class="notification is-danger">{{errors.message}}</div>
              {% endif %}
              <input type="submit" class="button is-primary is-outlined has-text-dark" value="comment" />
          </form>
      </div>
  </div>
</div>

Our site should now be functional, but it looks a little bland. Let's add a bit of style to it using Bulma CSS framework. We'll add the following reference in the head of our template:

<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/bulma@0.9.3/css/bulma.min.css">

Finally, we'll update the screen.css file located in the resources/public/css folder to format our form nicer:

ul {
	list-style: none;
}

ul.messages li {
	position: relative;
	font-size: 16px;
	padding: 5px;
	border-bottom: 1px dotted #ccc;
}

li:last-child {
	border-bottom: none;
}

li time {
	font-size: 12px;
	padding-bottom: 20px;
}

form, .error {
	padding: 30px;
	margin-bottom: 50px;
	position: relative;
}

When we reload the page in the browser we should be greeted by the guestbook page. We can test that everything is working as expected by adding a comment in our comment form.

To learn more about HTML templating options you can use with Kit, see HTML Templating.

Adding some tests

Tests are found under the test source path.

We can now run clj -M:test in the terminal to see that our database interaction works as expected.

Packaging the application

You can package your application for standalone deployment by running the following command:

clj -Sforce -T:build all

This will create a runnable jar that you can run using the following commands:

export JDBC_URL="jdbc:sqlite:guestbook_dev.db"
java -jar target/guestbook-standalone.jar

Note that we have to supply the JDBC_URL environment variable when running as a jar, as it's not packaged with the application.


Complete source listing for the tutorial is available here.