programming – AntPace.com

September 20, 2023November 6, 2023

Magic Squares in JavaScript

In mathematics, a “magic square” is a matrix of numbers where each row, column, and diagonal add-up to the same number. That number is called the “magic constant“. The integers used are only positive, and do not repeat. The constant sum is determined by the size of the square and is described by a formula:

M = n * ((n^2 + 1) / 2)

Facts about the properties and classifications of these numeric formations have been discussed by scholars for millennia. Knowledge of this topic goes back thousands of years, and can be found referenced throughout the world.

History & Culture

Magic squares have a fascinating historical and cultural significance, often with mystical undertones. They are mentioned in the I Ching, Brhat Samhita, and other works concerned with the transcendental and occult. They can be seen used in art, divination, perfumery, recreational gaming, computer science, and more.

a computer programmer using the Brhat Samhita to generate a talismans — AI images generated using Midjourney

In the Brhat Samhita, the magic square is used as a symbolic representation of the planets. It makes use of magic squares in the creation of talismans for astrological purposes.

For the purposes of this blog post, we’ll view them through the lens of software engineering.

3×3 Magic Squares

There’s so much to cover on this topic. I’ll narrow it down to 3×3 lattices (magic squares can actually be any size), specifically in the context of the JavaScript programming language. A quad of numbers, like the one pictured above, can be described as a two-dimensional array:

const magicSquare = [[4, 9, 2], [3, 5, 7], [8, 1, 6]];

Squares of this size always have a magic constant of 15. And, the number 5 will be in the middle. There’s additional logic that explains which numerals can appear where and why. Those ideas are explored in the comments section of a HackerRank coding challenge titled “Forming a Magic Square”.

HackerRank Coding Challenge

This coding problem found on HackerRank asks programmers to figure out what it would take to convert a 3×3 matrix of integers into a magic square. The input array is almost valid, but requires a few replacements. For each change, we must track the difference between the numbers and return the total variance – referred to as the “cost”. The correct answer will be the lowest cost required to convert the input data into a magic square.

The difficulty of this assignment is considered “medium”. The first step is realizing that there are a finite number of valid magic square configurations. As it turns out, there are exactly eight 3×3 permutations:

const magicSquares = [[[4, 9, 2], [3, 5, 7], [8, 1, 6]], 
                [[6, 1, 8], [7, 5, 3], [2, 9, 4]], 
                [[8, 1, 6], [3, 5, 7], [4, 9, 2]], 
                [[2, 9, 4], [7, 5, 3], [6, 1, 8]], 
                [[8, 3, 4], [1, 5, 9], [6, 7, 2]], 
                [[4, 3, 8], [9, 5, 1], [2, 7, 6]], 
                [[6, 7, 2], [1, 5, 9], [8, 3, 4]], 
                [[2, 7, 6], [9, 5, 1], [4, 3, 8]]];

Starting with any one of these, we can generate the other seven programmatically. The subsequent arrangements can be derived through rotation and reflection. Using JavaScript, I rotate an initial seed square three times to have the first four series. Then, I flip each one of those to get the final records.

function generateMagicSquares(magicSquare1){
	const magicSquares = [];
	magicSquares.push(magicSquare1);

	// we need to rotate it 3 times to get all rotations
	for(let i = 0; i < 3; i++){
		var rotation = magicSquares[i].map((val, index) => magicSquares[i].map(row => row[index]).reverse());
		// console.log(rotation)
		magicSquares.push(rotation);
	}

	// and then flip each one
	for(let i = 0; i < 4; i++){
		var flipped = magicSquares[i].map((_, colIndex) => magicSquares[i].map(row => row[colIndex]));
		magicSquares.push(flipped);
	}
	
	return magicSquares;
}

const magicSquare1 = [[4, 9, 2], [3, 5, 7], [8, 1, 6]];
const magicSquares = generateMagicSquares(magicSquare1);
console.log(magicSquares);

To solve this exercise, we’ll take the input array and compare it to each of the valid magic squares. We keep track of the cost on each iteration, and finally return the minimum.

function formingMagicSquare(s){
	const magicSquares = [[[4, 9, 2], [3, 5, 7], [8, 1, 6]], [[6, 1, 8], [7, 5, 3], [2, 9, 4]], [[8, 1, 6], [3, 5, 7], [4, 9, 2]], [[2, 9, 4], [7, 5, 3], [6, 1, 8]], [[8, 3, 4], [1, 5, 9], [6, 7, 2]], [[4, 3, 8], [9, 5, 1], [2, 7, 6]], [[6, 7, 2], [1, 5, 9], [8, 3, 4]], [[2, 7, 6], [9, 5, 1], [4, 3, 8]]];
	
	// let minCost = 100000;
	let minCost = Number.MAX_SAFE_INTEGER;
	let cost = 0;
	for(let i = 0; i < magicSquares.length; i++){
		cost = determineCost(s, magicSquares[i]);
		if(cost < minCost){
			minCost = cost;
		}
	}
	return minCost;
}

You’ll notice that the initial minimum cost is set to a very high number. As I loop through each of the valid magic squares, I check if the cost is lower than the current minimum and then replace the value.

With each comparison, subtraction is used to determine the cost to complete the transformation. That code loops through each digit of each row on both 2D arrays. The absolute value of the difference between each coordinate is tallied and returned.

function determineCost(inputArray, validMagicSquare){
	let cost = 0;
	for(let i = 0; i < 3; i++){ // each row
		
		for(let j = 0; j < 3; j++){ // each digit

			cost += Math.abs(inputArray[i][j] - validMagicSquare[i][j]);
		}
	}
	return cost;
}

The working code all stitched together looks like this:

<script>

function generateMagicSquares(magicSquare1){
	const magicSquares = [];
	magicSquares.push(magicSquare1);

	// we need to rotate it 3 times to get all rotations
	for(let i = 0; i < 3; i++){
		var rotation = magicSquares[i].map((val, index) => magicSquares[i].map(row => row[index]).reverse());
		// console.log(rotation)
		magicSquares.push(rotation);
	}

	// and then flip each one
	for(let i = 0; i < 4; i++){
		var flipped = magicSquares[i].map((_, colIndex) => magicSquares[i].map(row => row[colIndex]));
		magicSquares.push(flipped);
	}
	
	return magicSquares;
}

function determineCost(inputArray, validMagicSquare){
	let cost = 0;
	
	for(let i = 0; i < 3; i++){ // each row
		
		for(let j = 0; j < 3; j++){ // each digit

			cost += Math.abs(inputArray[i][j] - validMagicSquare[i][j]);
		}
	}

	return cost;

}

function formingMagicSquare(s){
	// const magicSquares = [[[4, 9, 2], [3, 5, 7], [8, 1, 6]], [[6, 1, 8], [7, 5, 3], [2, 9, 4]], [[8, 1, 6], [3, 5, 7], [4, 9, 2]], [[2, 9, 4], [7, 5, 3], [6, 1, 8]], [[8, 3, 4], [1, 5, 9], [6, 7, 2]], [[4, 3, 8], [9, 5, 1], [2, 7, 6]], [[6, 7, 2], [1, 5, 9], [8, 3, 4]], [[2, 7, 6], [9, 5, 1], [4, 3, 8]]];
	const magicSquare1 = [[4, 9, 2], [3, 5, 7], [8, 1, 6]];
	const magicSquares = generateMagicSquares(magicSquare1);
	
	// let minCost = 100000;
	let minCost = Number.MAX_SAFE_INTEGER;
	let cost = 0;
	for(let i = 0; i < magicSquares.length; i++){
		cost = determineCost(s, magicSquares[i]);
		if(cost < minCost){
			minCost = cost;
		}
	}
	return minCost;
}

const finalCost = formingMagicSquare([[4, 9, 2], [3, 5, 7], [8, 1, 5]]);
console.log(finalCost);
</script>

This solution was not intuitive to me and took some research and experimentation. It was interesting to learn about the concept of magic squares (and other shapes) along the way.

March 9, 2023January 22, 2024

Binary Search in JavaScript

“Binary search” is a computer science term used in programming and software engineering. It is a way of determining if a value exists in an ordered data set. It’s considered a textbook algorithm. It is useful because it reduces the search space by half on each iteration.

Imagine we are asked to search for an integer in an array that is sorted in ascending order. The challenge is to return the target number’s index.

We could use a built in JavaScript function indexOf():

var search = function(nums, target) {
    var answer = nums.indexOf(target);
    console.log(answer);
    return answer;
};
console.log(search([-1,0,3,4,5,9,12], 5)); // returns '4'

But that is an abstraction, which can be slow and expensive. Instead, we should implement our own binary search code.

Binary Search Implementation

Binary search starts in the middle of a collection, and checks if that is the requested term. The initial middle position is determined by taking the length of the array and dividing it by two. If the array has an even number of elements, then the ‘halfway position’ is considered close enough. I use a JavaScript Math function to round down, making sure I am not left with a decimal floating number:

cursor =  Math.floor( ((left + right) / 2) ); // about the middle

If the search term is not found, the position (the cursor) is moved. Since the collection is ordered from smallest to largest we know which direction to move the search cursor. If the target is less than the current value, we shift one to the left. If the target is greater than the current value, we shift two to the right.

This continues until the target query is found. That sequence is ran inside of a loop. The loop repeats while the left boundary is less than (or equal to) the right. The search cursor’s index is changed by adjusting the left or right bounds. Once adjusted, the cursor is recalculated on the subsequent iteration.

var search = function(nums, target) {
    var left = 0; // first element, initial left boundary
    var right = nums.length;
    right = right - 1; // last element, initial right boundary
    var cursor = 0;
    while(left <= right){
        cursor =  Math.floor( ((left + right) / 2) ); // about the middle to start
        console.log(cursor + ": " + nums[cursor]);
        if(nums[cursor] === target){
            return cursor; // query found!
        }
        if(target < nums[cursor]){
            right = cursor - 1; // move cursor to the left by one
            console.log("cursor moved to the left  by one")
        }else{
            left = cursor + 1; // move cursor to the right by two
            console.log("cursor moved to the right by two")
        }
    }
    return "false"; // not found
};
console.log(search([-1,0,1,2,3,4,5,6,7,8,9,10,11,12], 10));

The above code starts to search in the central location, nums[6] == 5. Look at the output below to see how the cursor moves until it finds the target value of 10 (at a zero-based index of 11):

This algorithm has an O(log n) runtime complexity. That Big-O notation describes how it would scale as the input increases. Imagine an array with thousands of elements. O(log n) means that the time to execute this code will increase linearly only as the number of input items increase exponentially. That means our code is performant and efficient.

Square root algorithm using binary search

Another implementation example of using binary search in a JavaScript function is to calculate the square root of a number. Now, of course, Javascript has a built in Math.sqrt() function that is highly performant compared to any custom code we will write. But, it is a good exercise to think about how this works.

The square root of a number is a value that, when multiplied by itself, gives the original number. To solve for it when given an input, we are going to take a guess at what the answer (square root) is, starting with itself divided by two (giving us the midpoint between it and zero). We multiply that midpoint by itself to check our guess. It will only be correct on the first iteration if the input is the number four (four divided by two is two; two times two is four). Otherwise, we will move the cursor (halving the search range) to either the left or the right, depending on if the square of our guess is larger or smaller than the original input.

The concept of “moving the cursor” helps to visualize how the binary search algorithm zeroes in on the square root by iteratively narrowing down the search range based on the comparison of the square of the midpoint to the target number.

I accounted for two special cases in my code. For negative numbers we return false because negative numbers have imaginary square roots. For numbers smaller than one we set the range’s end to one because the square root of a fraction is larger than the fraction itself. Additionally, I set a precision threshold (0.0000000001), because some numbers have square roots that are irrational.

function binarySearchSquareRoot(n, precision = 1e-10) {
    if (n < 0) {
        throw new Error('Input must be a non-negative number');
    }

    let start = 0;
    let end = n;

    // If the number is less than 1, set end to 1 to handle fractions
    if (n < 1) {
        end = 1;
    }

    while (end - start > precision) {
        const mid = (start + end) / 2;
        const midSquared = mid * mid;

        if (midSquared === n) {
            return mid;  // Found exact square root
        } else if (midSquared < n) {
            start = mid;
        } else {
            end = mid;
        }
    }

    return (start + end) / 2;  // Return approximate square root
}

// Usage:
const number = 25;
const squareRoot = binarySearchSquareRoot(number);
console.log(`The square root of ${number} is approximately ${squareRoot}`);

Alternatively, the Babylonian algorithm can be implemented to solve for square roots.

March 7, 2021October 11, 2023

How to create a Shopify app with PHP

Developing a marketplace app for your SAAS will grow organic traffic and lets users find you. Potential customers can discover your digital product where they are already looking. Software services that occupy the eCommerce space have a chance to help Shopify store owners grow their businesses. Creating a public Shopify app benefits both developers and the merchant user-base in the Shopify App Store ecosystem.

Recently, Shopify announced that it is decreasing the share of profit that it takes from developers. Each year, developers keep all of their revenue up to the first one-million dollars.

Why Build a Marketplace App?

The short answer: Discoverability.

A few years ago, I built a fitness tracking app for a niche sport. It was a hobby project to better track my BJJ training. Since then, I continue to average ~10 registrations weekly without any marketing efforts.

Consistent BJJ Tracker sign-ups are driven from Google Play. Even though it is only a web app (PWA), I was able to bundle it into an APK file using Trusted Web Activities and Digital Asset Links. Having an app listed in a marketplace leads to new users finding it naturally.

My next side project was a SAAS for split testing & conversion optimization. It helps websites A/B test to figure out what front-end changes lead to more sales, sign-ups, etc. The Shopify App Store was as perfect fit to attract shop owners to use SplitWit. I decided to build a public, embedded, Shopify app to reach new prospects.

I’ll explain how I did it, along with examples of building another one, all using PHP. This guide will make launching a Shopify App Store app easy, fast, repeatable.

Creating a Public Shopify App with PHP

Embedded Shopify apps display in a merchant’s admin dashboard. They are meant to “add functionality to Shopify stores“. They are hosted on the developer’s infrastructure and are loaded via iFrame within Shopify. You can create a new app in your Shopify Partners dashboard to get started.

Since the SplitWit SAAS already existed as a subscription web app built on the LAMP stack, I only had to handle Shopify specific authorization and payments. I could essentially load the existing app in the dashboard’s iFrame after authentication. The new code I wrote contains methods for checking installation status, building the oAuth URL, subscribing users to recurring application charges, and more.

When I created my next Shopify app, Click to Call, I leveraged that code and refactored it to be more reusable. Configurable parameters let me set the database and API credentials dynamically.

Step 1: Check Installation Status

When the app’s url loads in the merchant admin dashboard, the first step is to check installation status for that shop.

public function checkInstallationStatus(){
	$conn = $this->conn;
	$shop = $_GET['shop'];

	//check if app is already installed or not
	$statement = $conn->prepare("SELECT * FROM `shopify_installation_complete` WHERE shop = :shop");
	$statement->execute(['shop' => $shop]);
	$count = $statement->rowCount();
	if($count == 0){
		//app is not yet installed
		return false;
	}else{
		//app is already installed
		$row = $statement->fetch();
		return $row;
	}

}

You’ll notice we don’t hit any Shopify API for this. Instead, our own database is queried. We manually track if the app has already been installed in a MySql table.

Two database tables are used to manage the Shopify App installation:

CREATE TABLE IF NOT EXISTS `shopify_authorization_redirect` (
    `shopify_authorization_redirect_id` int NOT NULL AUTO_INCREMENT,
    `shop` varchar(200),
    `nonce` varchar(500),
    `scopes` varchar(500),
    created_date datetime DEFAULT CURRENT_TIMESTAMP,
    updated_date datetime ON UPDATE CURRENT_TIMESTAMP,
    PRIMARY KEY (`shopify_authorization_redirect_id`)
);

 
CREATE TABLE IF NOT EXISTS `shopify_installation_complete` (
    `shopify_installation_complete_id` int NOT NULL AUTO_INCREMENT,
    `splitwit_account_id` int,
    `splitwit_project_id` int,
    `shop` varchar(200),
    `access_token` varchar(200),
    `scope` varchar(200),
    `expires_in` int,
    `associated_user_scope` varchar(200),
    `associated_user_id` BIGINT,
    `associated_user_first_name` varchar(200),
    `associated_user_last_name` varchar(200),
    `associated_user_email` varchar(200),
    `associated_user_email_verified` varchar(10),
    `associated_user_account_owner` varchar(10),
    `associated_user_account_locale` varchar(10),
    `associated_user_account_collaborator` varchar(10),
    created_date datetime DEFAULT CURRENT_TIMESTAMP,
    updated_date datetime ON UPDATE CURRENT_TIMESTAMP,
    PRIMARY KEY (`shopify_installation_complete_id`)
);

The root file (index.php) looks like this:

<?php

require '/var/www/html/service-layer/shopify-app-service.php';
include 'shopify-creds.php'; // $api_key, $secret, $app_db, $app_slug
use SplitWit\ShopifyService\ShopifyService;
$shopify_service = new ShopifyService($api_key, $secret, $app_db);

$already_installed = $shopify_service->checkInstallationStatus();

if(!$already_installed){
    $install_redirect_url = $shopify_service->buildAuthorizationUrl(false, $app_slug);
}else{
    $install_redirect_url = $shopify_service->buildAuthorizationUrl(true, $app_slug);
}
header('Location: ' . $install_redirect_url );

?>

After determining the merchant’s installation status, the next step is to authenticate them.

oAuth Authentication

The Shopify developer resources explain “how to ask for permission” with oAuth. The merchant user needs to be redirected to a Shopify URL:

https://{shop}.myshopify.com/admin/oauth/authorize?client_id={api_key}&scope={scopes}&redirect_uri={redirect_uri}&state={nonce}&grant_options[]={access_mode}

I wrote a PHP class “ShopifyService” (shopify-app-service.php) to handle all of the Shopify specific logic. The method buildAuthorizationUrl() builds the Shopify authorization URL. It accepts a boolean parameter set according to the merchant’s installation status. That value toggles the authorization URL’s redirect URI, directing the code flow through either first-time installation or re-authentication.

The built URL includes query params: an API key, a nonce, the scope of permission being requested, and a redirect URI. The shop name is the subdomain, and can be pulled as GET data delivered to your app.

The API key can be found in the developer’s partner dashboard in the app’s overview page.

A nonce (“number used once”) is used as an oAuth state parameter. It serves to “link requests and callbacks to prevent cross-site request forgery attacks.” We save that random value in our database to check against during the oAuth callback.

The redirect URI (the oAuth callback) is dynamic based on the users installation status.

public function buildAuthorizationUrl($reauth = false, $slug= "shopify-app"){
	$conn = $this->conn;
	$requestData = $this->requestData;
	$scopes = "write_script_tags"; //write_orders,read_customers, read_content
	$nonce = bin2hex(random_bytes(10));
	$shop = $requestData['shop'];

	//first check if there is already a record for this shop. If there is, delete it first.
	$statement = $conn->prepare("SELECT * FROM `shopify_authorization_redirect` WHERE shop = :shop");
	$statement->execute(['shop' => $shop]);
	$count = $statement->rowCount();
	
	if($count > 0){
		$statement = $conn->prepare("DELETE FROM `shopify_authorization_redirect` WHERE shop = :shop");
		$statement->execute(['shop' => $shop]);
	}

	$statement = $conn->prepare("INSERT INTO `shopify_authorization_redirect` (shop, nonce, scopes) VALUES (:shop, :nonce, :scopes)");
	$statement->bindParam(':shop', $shop);
	$statement->bindParam(':nonce', $nonce);
	$statement->bindParam(':scopes', $scopes);
	$statement->execute();

	
	$redirect_uri = "https://www.splitwit.com/".$slug."/authorize-application";
	
	if($reauth){ //change the redirect URI
		$redirect_uri = "https://www.splitwit.com/".$slug."/reauthorize-application";
	}

	$redirect_url = "https://".$shop."/admin/oauth/authorize?client_id=". $this->api_key ."&scope=".$scopes."&redirect_uri=". $redirect_uri ."&state=".$nonce . "&grant_options[]=per-user";

	return $redirect_url;

}

Both possible redirect URLs needed to be white-listed in the “App setup” page.

A location header routes the user. If the app hasn’t been installed yet, Shopify prompts the merchant to confirm authorization.

If the app was already installed then the oAuth grant screen is skipped entirely and the merchant is immediately routed to the /reauthorize-application resource instead (and ultimately lands on the app home screen).

Install App & Register the Merchant User

What actually happens when “Install app” is clicked? The user is redirected from Shopify permissions screen back to our app ( /authorize-application ), calling the authorizeApplication() function. That method receives four values as GET parameters: ‘shop’, ‘state’, ‘hmac’, ‘code’

The ‘shop’ name is used to look up the nonce value we saved when the Shopify authorization URL was first built. We compare it to the ‘state’ parameter. This security step ensures that the callback request is valid and is not fraudulent. We also check that the shop name provided matches a valid Shopify hostname. Here is the relevant code, pulled from authorizeApplication():

$conn = $this->conn; 
$requestData = $this->requestData;
$requiredKeys = ['code', 'hmac', 'state', 'shop'];
foreach ($requiredKeys as $required) {
    if (!in_array($required, array_keys($requestData))) {
        throw new Exception("The provided request data is missing one of the following keys: " . implode(', ', $requiredKeys));
    }
}

//lookup and validate nonce
$shop = $requestData['shop'];

$statement = $conn->prepare("SELECT * FROM `shopify_authorization_redirect` WHERE shop = :shop");
$statement->execute(['shop' => $shop]);
$count = $statement->rowCount();
if($count == 0){
    throw new Exception("Nonce not found for this shop.");
}
$row = $statement->fetch();
$nonce = $row['nonce'];
//

//make sure the 'state' parameter provided matches the stored nonce
$state = $requestData['state'];
if($state !== $nonce){
    throw new Exception("Nonce does not match provided state.");
}
//

//validate the shop name
$pattern = "/[a-zA-Z0-9][a-zA-Z0-9\-]*\.myshopify\.com[\/]?/";
if(!preg_match($pattern, $shop)) {
    throw new Exception("The shop name is an invalid Shopify hostname.");
}

“Every request or redirect from Shopify” includes a HMAC value that can be used to verify its authenticity. Here is how I do it in PHP:

public function verifyHmac($requestData){
	// verify HMAC signature. 
	// https://help.shopify.com/api/getting-started/authentication/oauth#verification
	if( !isset($requestData['hmac'])){
		return false;
	}

	$hmacSource = [];

	foreach ($requestData as $key => $value) {
	    
	    if ($key === 'hmac') { continue; }

	    // Replace the characters as specified by Shopify in the keys and values
	    $valuePatterns = [
	        '&' => '%26',
	        '%' => '%25',
	    ];
	    $keyPatterns = array_merge($valuePatterns, ['=' => '%3D']);
	    $key = str_replace(array_keys($keyPatterns), array_values($keyPatterns), $key);
	    $value = str_replace(array_keys($valuePatterns), array_values($valuePatterns), $value);

	    $hmacSource[] = $key . '=' . $value;
	}

	sort($hmacSource);
	$hmacBase = implode('&', $hmacSource);
	$hmacString = hash_hmac('sha256', $hmacBase, $this->secret);
	// Verify that the signatures match
    if ($hmacString !== $requestData['hmac']) {
        return false;
    }else{
    	return true;
    }
}

That method is called in the class construct function, to be sure it happens every time.

The ‘code’ parameter is the access code. It is exchanged for an access token by sending a request to the shop’s access_token endpoint. We record that token to the ‘shopify_installation_complete’ table along with relevant data.

To fully complete the installation, app specific project records are saved. For SplitWit, this means a user-account is created along with an initial project. Any JavaScript tags are injected onto the merchant’s site using the Shopify admin API ScriptTag resource. Linking a privately hosted JS file, unique to each merchant, allows our app to dynamically update the shop website. You can learn about how that snippet tag works in another post that explains how the visual editor is built.

Lastly, a webhook is created to listen for when this app in uninstalled ( ‘topic’ => ‘app/uninstalled’ ) in order to call our “uninstallApplication” method. Webhooks allow you to listen for certain events in a shop, and run code based on data about what happened.

Once installation is complete, our server returns a header that reloads the app.

Below is the original authorizeApplication() method. Eventually, I moved app specific logic into its own files after refactoring this class to support another SAAS, Click to Call.

public function authorizeApplication(){
	$conn = $this->conn; 
	$requestData = $this->requestData;
	$requiredKeys = ['code', 'hmac', 'state', 'shop'];
        foreach ($requiredKeys as $required) {
           if (!in_array($required, array_keys($requestData))) {
             throw new Exception("The provided request data is missing one of the following keys: " . implode(', ', $requiredKeys));
           }
        }

	//lookup and validate nonce
	$shop = $requestData['shop'];
	
	$statement = $conn->prepare("SELECT * FROM `shopify_authorization_redirect` WHERE shop = :shop");
	$statement->execute(['shop' => $shop]);
	$count = $statement->rowCount();
	if($count == 0){
        throw new Exception("Nonce not found for this shop.");
	}
	$row = $statement->fetch();
	$nonce = $row['nonce'];
	//
	
	//make sure the 'state' parameter provided matches the stored nonce
	$state = $requestData['state'];
	if($state !== $nonce){
        throw new Exception("Nonce does not match provided state.");
	}
	//
	
	//validate the shop name
	$pattern = "/[a-zA-Z0-9][a-zA-Z0-9\-]*\.myshopify\.com[\/]?/";
	if(!preg_match($pattern, $shop)) {
        throw new Exception("The shop name is an invalid Shopify hostname.");
	}
	//

	$already_installed = $this->checkInstallationStatus();
	//if it is already installed, then lets update the access token 
        if(!$already_installed){
    	  //install the app
    	
	  //exchange the access code for an access token by sending a request to the shop’s access_token endpoint
	  $code = $requestData['code'];
	  $post_url = "https://" . $shop . "/admin/oauth/access_token";
		
	  $params = [
            'client_id'    => $this->api_key,
            'client_secret'    => $this->secret,
            'code'    => $code
          ];

          $curl_response_json = $this->curlApiUrl($post_url, $params);
	  $access_token = $curl_response_json['access_token'];
		
          $statement = $conn->prepare("INSERT INTO `shopify_installation_complete` (shop, access_token, scope, expires_in, associated_user_scope, associated_user_id, associated_user_first_name, associated_user_last_name, associated_user_email, associated_user_email_verified, associated_user_account_owner, associated_user_account_locale, associated_user_account_collaborator) VALUES (:shop, :access_token, :scope, :expires_in, :associated_user_scope, :associated_user_id, :associated_user_first_name, :associated_user_last_name, :associated_user_email, :associated_user_email_verified, :associated_user_account_owner, :associated_user_account_locale, :associated_user_account_collaborator)");
		
	  $statement->bindParam(':shop', $shop);
		
	  $statement->bindParam(':access_token', $access_token);
	  $statement->bindParam(':scope', $curl_response_json['scope']);
	  $statement->bindParam(':expires_in', $curl_response_json['expires_in']);
	  $statement->bindParam(':associated_user_scope', $curl_response_json['associated_user_scope']);
	  $statement->bindParam(':associated_user_id', $curl_response_json['associated_user']['id']);
	  $statement->bindParam(':associated_user_first_name', $curl_response_json['associated_user']['first_name']);
	  $statement->bindParam(':associated_user_last_name', $curl_response_json['associated_user']['last_name']);
	  $statement->bindParam(':associated_user_email', $curl_response_json['associated_user']['email']);
	  $statement->bindParam(':associated_user_email_verified', $curl_response_json['associated_user']['email_verified']);
	  $statement->bindParam(':associated_user_account_owner', $curl_response_json['associated_user']['account_owner']);
	  $statement->bindParam(':associated_user_account_locale', $curl_response_json['associated_user']['locale']);
	  $statement->bindParam(':associated_user_account_collaborator', $curl_response_json['associated_user']['collaborator']);

	  $statement->execute();
	  $installation_complete_id = $conn->lastInsertId();
		 
	  if(isset($curl_response_json['associated_user']['email']) && strlen($curl_response_json['associated_user']['email']) > 0){

		$store_name = explode(".", $shop);
		$store_name = ucfirst($store_name[0]);

		//create account
		$method = "thirdPartyAuth";
		$user_service_url = "https://www.splitwit.com/service-layer/user-service.php?third_party_source=shopify&method=" . $method . "&email=".$curl_response_json['associated_user']['email']."&companyname=" .$store_name . "&first=" . $curl_response_json['associated_user']['first_name'] . "&last=" . $curl_response_json['associated_user']['last_name'] ;
			
		$params = [];

		$curl_user_response_json = $this->curlApiUrl($user_service_url, $params);
		
		$account_id = $curl_user_response_json['userid']; 
			
		$method = "createProject";
			
		$project_service_url = "https://www.splitwit.com/service-layer/project-service.php?method=" . $method . "&accountid=" . $account_id;

		$params = [
	            'projectname'    => $store_name . " Shopify",
	            'projectdomain'    => "https://".$shop,
	            'projectdescription'    => ""
	        ];

		$curl_project_response_json = $this->curlApiUrl($project_service_url, $params);
		$project_id = $curl_project_response_json['projectid'];
		$snippet = $curl_project_response_json['snippet'];
			
			

		//inject JS snippet into site
		// https://shopify.dev/docs/admin-api/rest/reference/online-store/scripttag#create-2020-04
		$create_script_tag_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/script_tags.json";
		$params = [
                    'script_tag' => [
                     'event' => 'onload',
                     'src' => 'https://www.splitwit.com/snippet/' . $snippet
                    ]
        	];

        	$headers = array(
		  'X-Shopify-Access-Token:' . $access_token,
		  'content-type: application/json'
		);

		$json_string_params = json_encode($params);

		$create_script_curl_response_json = $this->curlApiUrl($create_script_tag_url, $json_string_params, $headers);
		
		//shopify app should only ever have access to this one project.
		//write accountID and ProjectID to this shopify_installation_complete record.

		$statement = $conn->prepare("UPDATE `shopify_installation_complete` SET splitwit_account_id = ?, splitwit_project_id = ? WHERE shopify_installation_complete_id = ?");

		$statement->execute(array($account_id, $project_id, $installation_complete_id));
			
    	}
		
    	//create webhook to listen for when app in uninstalled.
	//https://{username}:{password}@{shop}.myshopify.com/admin/api/{api-version}/{resource}.json
	// https://shopify.dev/docs/admin-api/rest/reference/events/webhook#create-2020-04
	$create_webhook_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/webhooks.json";
	$params = [
                'webhook' => [
                    'topic' => 'app/uninstalled',
                    'address' => 'https://www.splitwit.com/service-layer/shopify-app-service?method=uninstallApplication',
                    'format' => 'json'
                ]
        ];

	$headers = array(
		'X-Shopify-Access-Token:' . $access_token,
		'content-type: application/json'
	);
		
	$json_string_params = json_encode($params);

    	$create_webhook_curl_response_json = $this->curlApiUrl($create_webhook_url, $json_string_params, $headers);
		
	//installation complete.
   }

   header('Location: ' . "https://" . $shop . "/admin/apps/splitwit");
	
}

You’ll notice that I call a custom method that abstracts the PHP cURL (client URL library) methods. This helps me avoid repeating the same code in multiple places.

public function curlApiUrl($url, $params, $headers = false, $use_post = true, $use_delete = false, $use_put = false){
		
	$curl_connection = curl_init();
	// curl_setopt($curl_connection, CURLOPT_FOLLOWLOCATION, true);
	if($headers){
	    curl_setopt($curl_connection, CURLOPT_HTTPHEADER, $headers);
	    curl_setopt($curl_connection, CURLOPT_HEADER, false);
	}
	curl_setopt($curl_connection, CURLOPT_URL, $url);
    
       // TODO: refactor these three conditions into one, that accepts the RESTful request type!!
       if($use_post){
	    curl_setopt($curl_connection, CURLOPT_POST, true);
	    curl_setopt($curl_connection, CURLOPT_POSTFIELDS, $params);
	}
        if($use_delete){
	    curl_setopt($curl_connection, CURLOPT_CUSTOMREQUEST, "DELETE");
	}
        if($use_put){
	    curl_setopt($curl_connection, CURLOPT_CUSTOMREQUEST, "PUT");
	    curl_setopt($curl_connection, CURLOPT_POSTFIELDS, $params);
	}
	//end TODO

	curl_setopt($curl_connection, CURLOPT_RETURNTRANSFER, true);
	$curl_response = curl_exec($curl_connection);
	$curl_response_json = json_decode($curl_response,true);
	curl_close($curl_connection);
	return $curl_response_json;
}

The merchant will be able to find the app in the ‘Apps’ section of their dashboard. Shopify remembers that permission was granted by the merchant.

Returning User Log in

The oAuth grant screen will not show again when the app is selected in the future. As the installation status returns true, our code will flow into the reAuthenticate() method. The same validation checks are performed and a new access token is received.

   
public function reAuthenticate(){
    $conn = $this->conn; 
    $requestData = $this->requestData;
    $requiredKeys = ['code', 'hmac', 'state', 'shop'];
    foreach ($requiredKeys as $required) {
        if (!in_array($required, array_keys($requestData))) {
            throw new Exception("The provided request data is missing one of the following keys: " . implode(', ', $requiredKeys));
            // return;
        }
    }

    //lookup and validate nonce
    $shop = $requestData['shop'];
    
    $statement = $conn->prepare("SELECT * FROM `shopify_authorization_redirect` WHERE shop = :shop");
    $statement->execute(['shop' => $shop]);
    $count = $statement->rowCount();
    if($count == 0){
        throw new Exception("Nonce not found for this shop.");
    }
    $row = $statement->fetch();
    $nonce = $row['nonce'];
    //
    
    //make sure the 'state' parameter provided matches the stored nonce
    $state = $requestData['state'];
    if($state !== $nonce){
        throw new Exception("Nonce does not match provided state.");
    }
    //
    
    //validate the shop name
    $pattern = "/[a-zA-Z0-9][a-zA-Z0-9\-]*\.myshopify\.com[\/]?/";
    if(!preg_match($pattern, $shop)) {
        throw new Exception("The shop name is an invalid Shopify hostname.");
    }

    //exchange the access code for an access token by sending a request to the shop’s access_token endpoint
    $code = $requestData['code'];
    $post_url = "https://" . $shop . "/admin/oauth/access_token";
    
    $params = [
        'client_id'    => $this->api_key,
        'client_secret'    => $this->secret,
        'code'    => $code
    ];

    $curl_response_json = $this->curlApiUrl($post_url, $params);
    $access_token = $curl_response_json['access_token'];
    
    $statement = $conn->prepare("UPDATE `shopify_installation_complete` SET access_token = ? WHERE shop = ?");
    $statement->execute(array($access_token, $shop));

    header('Location: ' . "/home?shop=".$shop);
}

The merchant is routed to the app’s /home location. A few session variables are set and the user interface is loaded.

<?php
require '/var/www/html/service-layer/shopify-app-service.php';
use SplitWit\ShopifyService\ShopifyService;
$shopify_service = new ShopifyService();
include '/var/www/html/head.php'; 
?>
<style>
	.back-to-projects{
		display: none;
	}   
</style>

<body class="dashboard-body">
    <?php 
    // log the user out...
    $sess_service = new UserService();
    $sess_service -> logout();
    //logout destroys the session. make sure to start a new one.
    if (session_status() == PHP_SESSION_NONE) {
        session_start();
    }
    // ...then log them in
    $already_installed = $shopify_service->checkInstallationStatus();
    $shopify_service->makeSureRecordsExist($already_installed);
    $projectid = $shopify_service->splitwit_project_id; 
    $accountid = $shopify_service->splitwit_account_id; 
    $_SESSION['accountid'] = $accountid;
    $_SESSION['userid'] =  $accountid;
    $_SESSION['email'] = $already_installed['associated_user_email'];
    
    $sess_service -> login();
    $_SESSION['active'] = true;
    include '/var/www/html/includes/experiments-ui.php'; 
    
    ?>
</body>
</html>

The method makeSureRecordsExist() checks that the SplitWit user account and project records exist, as a failsafe. The .back-to-projects CTA is hidden because Shopify users only have access to one project for their shop. The app is installed for free, while premium functionality requires a subscription after a one-week trial.

When building my second Shopify app, I started with an empty home screen UI:

<?php
require '/var/www/html/service-layer/shopify-app-service.php';
include 'shopify-creds.php';
use SplitWit\ShopifyService\ShopifyService;
$shopify_service = new ShopifyService($api_key, $secret, $app_db);
?>
<html>
<head>
</head>
<body>
   
</body>
<script>


    if(self!==top){
        // if loaded outside of an iframe, redirect away to a marketing page
        window.location = "https://www.splitwit.com";
    }


</script>
</html>

Subscription Payment

SplitWit’s codebase is originally used as a non-Shopify, stand-alone, web app SAAS. It uses Stripe as a payment gateway. Shopify requires apps to use their Billing API instead. To remedy this, I’m able to write Shopify specific front-end code with a simple JavaScript check. I leverage the browser’s window.self property to check if my app’s code is running in the top most window (opposed to being nested in an iFrame).

if(self!==top){
	// shopify app
	$(".activate-cta").remove();
	if(window.pastDueStatus || window.customerId.length === 0){
		$(".activate-loading").show();

		$.ajax({
			url:"/service-layer/shopify-app-service?method=createRecurringApplicationCharge",
			complete: function(response){
				console.log(response)
				//user is redirected to shopify confirmation screen
				$(".activate-loading").hide();
				$(".activate-cta-top").attr("href", response.responseText).show();
			}
		});
	}
	 
	$(".back-cta").click(function(){
		window.history.back();
	})
	$(".reset-pw-cta").hide()

}else{

	$(".activate-cta").click(function(){
		$(".stripe-payment-modal").show();
	});
	
	$(".back-cta").hide();
}

If it’s not the top most window, I assume the code is running in Shopify. I’ll change the click-event listener on the .activate-cta element to create a recurring subscription charge. An AJAX call is made to our PHP end-point that hits Shopify’s RecurringApplicationCharge API.

public function createRecurringApplicationCharge(){
	
	$conn = $this->conn;
	$statement = $conn->prepare("SELECT * FROM `shopify_installation_complete` WHERE splitwit_account_id = :splitwit_account_id");
	$statement->execute(['splitwit_account_id' => $_SESSION['accountid']]);
	$row = $statement->fetch();
	$shop = $row['shop'];
	$access_token = $row['access_token'];
	
	$create_recurring_charge_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/recurring_application_charges.json";
	$params = [
        'recurring_application_charge' => [
            'name' => 'Basic Plan',
            'price' => 25.0,
            // 'return_url' => "https://" . $shop . "/admin/apps/splitwit",
            // 'test' => true,
            'return_url' => "https://www.splitwit.com/service-layer/shopify-app-service?method=confirmSubscription"
        ]
	];
	$headers = array(
		'X-Shopify-Access-Token: ' . $access_token,
	 	'content-type: application/json'
	);
	$json_string_params = json_encode($params);

	$create_recurring_charge_curl_response_json = $this->curlApiUrl($create_recurring_charge_url, $json_string_params, $headers);
	echo $create_recurring_charge_curl_response_json['recurring_application_charge']['confirmation_url'];
}

The charge ID (delivered by the Shopify request to our ‘return_url’), payment processor, and subscription expiry date are saved to our database on call-back before returning a location header to reload the app.

public function confirmSubscription(){
	
	$conn = $this->conn;
	$statement = $conn->prepare("SELECT * FROM `shopify_installation_complete` WHERE splitwit_account_id = :splitwit_account_id");
	$statement->execute(['splitwit_account_id' => $_SESSION['accountid']]);
	$row = $statement->fetch();
	$shop = $row['shop'];
 
 	$charge_id = $_REQUEST['charge_id'];
	//write shopify billing ID to db
	$sql = "UPDATE `account` SET payment_processor = ?, billing_customer_id = ?, current_period_end = ?, past_due = 0 WHERE accountid = ?"; 
	$result = $conn->prepare($sql); 
	$current_period_end = new \DateTime();  //we need the slash here (before DateTime class name), since we're in a different namespace (declared at the top of this file)
	$current_period_end->modify( '+32 day' );
	$current_period_end = $current_period_end->format('Y-m-d H:i:s'); 
	$payment_processor = "shopify";
	$result->execute(array($payment_processor, $charge_id, $current_period_end, $_SESSION['accountid']));
	
	//redirect to app
	header('Location: ' . "https://" . $shop . "/admin/apps/splitwit");

}

That charge ID (saved to our database in a column titled “billing_customer_id”) can later be passed back to Shopify to delete the recurring charge.

Cancel a Subscription

Once a subscription is active, I can check the payment processor saved the the account’s DB record to toggle the “cancel account” functionality from Stripe to Shopify.

<?php if ($account_row['payment_processor'] == "shopify"){ ?>
	//hit shopify service

	$(".cancel-cta").click(function(){
		//
		$.ajax({
			url:"/service-layer/shopify-app-service?method=cancelSubscription",
			complete: function(response){
				window.location.reload();
			}
		});
	});

<?php }else{ ?>
	//hit the stripe service
	
	$(".cancel-cta").click(function(){
		$(".cancel-subscription-modal").show();
	});

<?php }?>

The cancelSubscription method hits the same Shopify recurring_application_charges API, but uses a DELETE request. It also deletes the Shopify billing ID from our records.

public function cancelSubscription(){
	
	$conn = $this->conn;
	$statement = $conn->prepare("SELECT * FROM `shopify_installation_complete` WHERE splitwit_account_id = :splitwit_account_id");
	$statement->execute(['splitwit_account_id' => $_SESSION['accountid']]);
	$row = $statement->fetch();
	$shop = $row['shop'];
	$access_token = $row['access_token'];

	$statement = $conn->prepare("SELECT * FROM `account` WHERE accountid = :accountid");
	$statement->execute(['accountid' => $_SESSION['accountid']]);
	$account_row = $statement->fetch();
	$charge_id = $account_row['billing_customer_id'];


	$delete_recurring_charge_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/recurring_application_charges/#" . $charge_id . ".json";

	$params = [];
	$headers = array(
		'X-Shopify-Access-Token: ' . $access_token,
	 	'content-type: application/json'
	);
	$json_string_params = json_encode($params);
	$delete = true;

	$delete_recurring_charge_curl_response_json = $this->curlApiUrl($delete_recurring_charge_url, $json_string_params, $headers, $delete);

	//delete shopify billing ID from db
	$empty_string = "";
	$sql = "UPDATE `account` SET payment_processor = ?, billing_customer_id = ? WHERE accountid = ?"; 
	$result = $conn->prepare($sql); 
	$result->execute(array($empty_string, $empty_string, $_SESSION['accountid']));
	
	echo $delete_recurring_charge_curl_response_json;


}

I can use these same recurring application API end-point functions with minimal adjustments for other Shopify apps that I build. After refactoring, I am able to specify an app database as a GET parameter in the AJAX calls to my Shopify PHP service.

Uninstall the App

Merchants can choose to delete apps from their shop. This will remove it from their list of installed apps. If they try installing it again, they will be re-promoted for permissions. When an app is deleted, a webhook is notified so that code can handle server-side uninstall logic:

The payment processor and billing ID associated with the merchant’s account is set to an empty string. The ‘shopify_installation_complete’ shop record is deleted.

public function uninstallApplication(){
	$conn = $this->conn; 
	
	$res = '';
	$hmac_header = $_SERVER['HTTP_X_SHOPIFY_HMAC_SHA256'];
	$topic_header = $_SERVER['HTTP_X_SHOPIFY_TOPIC'];
	$shop_header = $_SERVER['HTTP_X_SHOPIFY_SHOP_DOMAIN'];
	$data = file_get_contents('php://input'); //similar to $_POST
	$decoded_data = json_decode($data, true);
	$verified = $this->verifyWebhook($data, $hmac_header);

	if( $verified == true ) {
	  if( $topic_header == 'app/uninstalled' || $topic_header == 'shop/update') {
	    if( $topic_header == 'app/uninstalled' ) {
			$domain = $decoded_data['domain'];

			$statement1 = $conn->prepare("SELECT * FROM `shopify_installation_complete` WHERE shop = ?");
			$statement1->execute(array($domain));
			$row = $statement1->fetch();
			$accountid = $row['splitwit_account_id'];

			//delete shopify billing ID from db
			$empty_string = "";
			$result = $conn->prepare("UPDATE `account` SET payment_processor = ?, billing_customer_id = ? WHERE accountid = ?"); 
			$result->execute(array($empty_string, $empty_string, $accountid));

			$statement = $conn->prepare("DELETE FROM `shopify_installation_complete` WHERE shop = ?");
			$statement->execute(array($domain));

	    } else {
	      $res = $data;
	    }
	  }
	} else {
	  $res = 'The request is not from Shopify';
	}

}

Any webhook requests have the HMAC delivered as a header (instead of a query param, as in the case of oAuth requests) and is processed differently. “The HMAC verification procedure for OAuth is different from the procedure for verifying webhooks“. The method verifyWebhook() takes care of it:

public function verifyWebhook($data, $hmac_header){
  $calculated_hmac = base64_encode(hash_hmac('sha256', $data, $this->secret, true));
  return hash_equals($hmac_header, $calculated_hmac);
}

Cache Busting

When project changes are recorded in the app, the merchant’s snippet file is updated. We need to be sure that their website recognizes the latest version. In a separate class (that handles project & snippet logic) I make a HTTP request to my method that re-writes the script tag.

public function updateSnippetScriptTag(){
	$projectid = $_GET['projectid'];
	$conn = $this->conn;
	$sql = "SELECT * FROM `shopify_installation_complete` WHERE splitwit_project_id = ?"; 
	$result = $conn->prepare($sql); 
	$result->execute(array($projectid));
	$row = $result->fetch(\PDO::FETCH_ASSOC);
	$number_of_rows = $result->rowCount();
	if($number_of_rows == 1){
		$access_token = $row['access_token'];
		$shop = $row['shop'];
		$sql = "SELECT * FROM `project` WHERE projectid = ?"; 
		$project_result = $conn->prepare($sql); 
		$project_result->execute(array($projectid));
		$project_row = $project_result->fetch(\PDO::FETCH_ASSOC);
		$snippet = $project_row['snippet'];			

		$script_tag_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/script_tags.json";
		$headers = array(
		  'X-Shopify-Access-Token:' . $access_token,
		  'content-type: application/json'
		);
		$params = [];
		$json_string_params = json_encode($params);
		$use_post = false;
		//get existing script tag
		$get_script_curl_response_json = $this->curlApiUrl($script_tag_url, $json_string_params, $headers, $use_post);
		$tags = $get_script_curl_response_json['script_tags'];
	
		foreach ($tags as $tag) {
			$id = $tag['id'];
			$delete_script_tag_url = "https://" . $this->api_key . ":" . $this->secret . "@" . $shop . "/admin/api/2020-04/script_tags/" . $id . ".json";
			$use_delete = true;
			$delete_script_curl_response_json = $this->curlApiUrl($delete_script_tag_url, $json_string_params, $headers, $use_post, $use_delete);
		}
		 
		//add snippet
		$snippet = "https://www.splitwit.com/snippet/" . $snippet . "?t=" . time();
		$params = [
			'script_tag' => [
				'event' => 'onload',
				'src' => $snippet 
			]
		];
		$json_string_params = json_encode($params);
		$create_script_curl_response_json = $this->curlApiUrl($script_tag_url, $json_string_params, $headers);	 

	}
}

Once our Shopify app is built and tested we can begin to prepare for submission to the Shopify App Market.

Preparing for production

Shopify allows you to test your app on a development store.

After debugging your code locally, make sure it works end-to-end in Shopify’s environment.

Even though the app is “unlisted”, and has not yet been accepted into the Shopify App Market, you’ll still be able to work through the entire UX flow.

GDPR mandatory webhooks

“Each app developer is responsible for making sure that the apps they build for the Shopify platform are GDPR compliant.” Every app is required to provide three webhook end-points to help manage the data it collects. These end-points make requests to to view stored customer data, delete customer data, and delete shop data. After handling the request, an HTTP status of 200/OK should be returned. PHP lets us do that with its header() function:

header("HTTP/1.1 200 OK");

These GDPR webhook subscriptions can be managed on the “App setup” page.

App Listing

Before submitting your app to the Shopify App Market, you’ll need to complete “Listing Information”. This section includes the app’s name, icon, description, pricing details, and more. It is encouraged to include screenshots and a demonstration video. Detailed app review instructions, along with screenshots and any on-boarding information, will help move the approval process along more quickly.

Approval Process

Complete the setup and listing sections, and submit your app.

You’ll receive an email letting you know that testing will begin shortly.

You may be required to make updates based on feedback from Shopify’s review process. After making any required changes, your application will be listed on the Shopify App Store. Below is an example of feedback that I had received:

To remedy the first required change I added additional onboarding copy to the app’s listing and included a demonstration YouTube video.

The second point was fixed by stopping any links from opening in new tabs. (Although, the reviewer’s note about ad blocking software stopping new tabs from opening is bogus).

The third issue was resolved by making sure the graphic assets detailed in my app listing were consistent.

Soon after making these changes, my app was finally approved and listed.

Keep Building

While writing this article I extended and refactored my PHP code to support multiple apps. I added configuration files to keep database settings modular. The Shopify PHP class can serve as back-end to several implementations. If you have any questions about how to build a Shopify app, or need my help, send me a message.

Update:

I wrote a subsequent post about building another Shopify app. It’s called SplitWit Click to Call. It explains the creative details that go into shipping a fulling working SAAS. I dive into new features that are only available to Shopify themes running the latest OS2.0 experience.

January 24, 2021December 1, 2022

Visual Website Editor in PHP

Tools for non-programmers to manage websites are growing in demand. No-code solutions increase bandwidth. Many SaaS products promise marketers and project managers a way to get things done without a developer. A visual website editor baked into an app is an important technology for digital product builders. The ability for a software service to interact with a client’s website can be a critical selling point.

SplitWit A/B Testing is an example of a web app that had that requirement. It lets user specify an existing website and displays it in an editor window. The user can then click on elements, make changes, and add new content. Those changes are actually applied to the live website by a JavaScript snippet. This post explains how it was built using PHP and jQuery.

Showing the website via iFrame

Displaying an already existing website on your app is easy with an iFrame. But, since users should be able to view and edit any page they choose, challenges arise. If a non-SSL URL is used (http:// instead of https://), the editor app will throw a “mixed-content warning”. If the page has X-FRAME-OPTIONS set to DENY, then the request will fail. To get around these issues I load an internal page into the iFrame that contains all of the DOM code from the one specified.

var pageIframe = $('<iframe id="page-iframe" style="" src="/page-to-edit.php?baseUrl='+baseUrl+'&url=<?php echo $url; ?>"></iframe>').appendTo(".page-iframe-wrapper");

The page to be edited is entered into an input field. That 3rd party webpage’s URL is passed along to my internal “page-to-edit.php” as a query parameter. From there, I get its full code using PHP Simple HTML DOM Parser. You’ll notice that I also pass along a “base url” – which is the root domain of the page. I’m able to grab it using JavaScript.

<?php
$url = "";
if(isset($_GET['url']) && strlen($_GET['url']) > 0){
  $url = $_GET['url'];
}
?>
function removeSubdirectoryFromUrlString(url, ssl){
  
  var ssl = ssl || false;
  if(url.indexOf("https://")){
    ssl = true;
  }

  url = url.replace("http://", "");
  url = url.replace("https://", "");
  var pathArray = url.split("/")
  url = pathArray[0];
  if(ssl){
    url = "https://" + url;
  }else{
    url = "http://" + url;
  }

  return url;
}

var url = "<?php echo $url; ?>";
var ssl = false;
var pageIframe;
 
if(url.length > 0){

  if(url.indexOf("https:") !== -1){
    ssl = true
  }
   
  var baseUrl = removeSubdirectoryFromUrlString(url, ssl);
  if(baseUrl.slice(-1) !== "/"){
    baseUrl = baseUrl + "/";
  }
}

I need to append that to any assets (images, scripts, etc.) that use relative references – or else they won’t load.

<?php 
// Report all errors
error_reporting(E_ALL);
ini_set("display_errors", 1);
require 'simple_html_dom.php';


class HtmlDomParser {
	
	/**
	 * @return \simplehtmldom_1_5\simple_html_dom
	 */
	static public function file_get_html() {
		return call_user_func_array ( '\simplehtmldom_1_5\file_get_html' , func_get_args() );
	}

	/**
	 * get html dom from string
	 * @return \simplehtmldom_1_5\simple_html_dom
	 */
	static public function str_get_html() {
		return call_user_func_array ( '\simplehtmldom_1_5\str_get_html' , func_get_args() );
	}
}

$base_url = $_GET['baseUrl'];
$url = $_GET['url'];
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, TRUE);
$html = curl_exec($ch);
$redirectedUrl = curl_getinfo($ch, CURLINFO_EFFECTIVE_URL);
$url = $redirectedUrl;

$parse = parse_url($url);

$base_url = $parse["scheme"] . "://" . $parse["host"];

$html = @HtmlDomParser::file_get_html($url);

if(substr($base_url, 0, 1) == "/"){
    $base_url = substr_replace($base_url ,"",-1);
}

if($html === FALSE) { 
    echo "Sorry, we do not have permission to analyze that website. ";
    return;
}
foreach($html->find('script') as $element){
    $src = $element->src;
    // echo "<script>console.log('starting src: ".$src."')</script>";
    
    if (strlen($src) > 0 && strpos($src, '//') === false){
        if(substr($src, 0, 1) !== "/"){
            $src = "/" . $src;
        }
        $element->src = $base_url . $src;
    }
    if(strlen($element->integrity) > 0){
        $element->integrity = "";
    }
    // echo "<script>console.log('final src: ".$base_url . $src."')</script>";

    // echo $element->src . "\n";
} 
foreach($html->find('link') as $element){
    $src = $element->href;
    
    if (strlen($src) > 0 && strpos($src, '//') === false){
        if(substr($src, 0, 1) !== "/"){
            $src = "/" . $src;
        }
        $element->href = $base_url . $src;
    }
    if(strlen($element->integrity) > 0){
        $element->integrity = "";
    }
   
}
foreach($html->find('a') as $element){
    $src = $element->href;
    
    if (strlen($src) > 0 && strpos($src, '//') === false){
        if(substr($src, 0, 1) !== "/"){
            $src = "/" . $src;
        }
        $element->href = $base_url . $src;
    } 
  
}
foreach($html->find('img') as $element){
    $src = $element->src;
    if (strlen($src) > 0 && strpos($src, '//') === false){
        if(substr($src, 0, 1) !== "/"){
            $src = "/" . $src;
        }

        $element->src = $base_url . $src;
    } 
   
}
foreach($html->find('source') as $element){
    $src = $element->srcset;
    $sources = explode(",",$src);
    $src = trim($sources[0]);

    if (strlen($src) > 0 && strpos($src, '//') === false){
        if(substr($src, 0, 1) !== "/"){
            $src = "/" . $src;
        }

        $element->srcset = $base_url . $src;
    } 
}

echo $html;

?>

I check against five different element types for assets that could need to be updated: <script>, <link>, <a>, <img>, and <source>. In version 1.0, I had missed the <link> and <source> elements.

Now that I was able to load any website into my editor UI, I had to think about layout. I split the screen in half, having the right side display the website, and the left side showing editor options.

Update: Eventually, I noticed that some pages were not loading properly. I found it to be two issues: a security error & a permission-denied response. The security error was a PHP problem. The permission-denied (401) response was due to server settings. Some sites don’t want to be accessed by bots and try to ensure that an actual user is making the request. I was able to fix it by passing context settings to the ‘file_get_html’ method.

$arrContextOptions=array(
    "ssl"=>array(
        "verify_peer"=>false,
        "verify_peer_name"=>false,
    ),
    "http" => array(
        "header" => "User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36"
    )
); 
$html = file_get_html($url, false, stream_context_create($arrContextOptions));

Element selection

Page-clicks in the editor window need to be intercepted to stop links from being followed and any other actions from happening. Instead, any thing clicked is selected by the editor and its details shown in the options panel.

As elements are moused over, a semi-transparent red highlight is applied as a visual cue. It is achieved by injecting CSS into the iFrame and adding mouseenter and mouseout event listeners.

pageIframe.on('load', function(){
    var style = "<style>.highlighted{background-color:rgba(255, 0, 0, 0.5);} </style>";         
    pageIframe.contents().find("body").prepend(style);
     
    pageIframe.contents().find("body *").mouseenter(function(){
	$(this).addClass('highlighted'); 
    }).mouseout(function(){
	$(this).removeClass('highlighted');   
    });

});

When something is clicked, normal functionality is stopped by using the web browser’s preventDefault() method. Once the element is grabbed (using the this keyword), we want to be sure that we’re as deep into the DOM tree as possible. That way, our “Text / HTML” content won’t contain unnecessary code tags.

To do so, we use a while loop to iterate an arbitrary number of times. Eight iterations seems to be the number that gets us to the heart of any selected content, without taking too much time.

First, we check if the element has any child nodes – if it does, we grab the first one. If that node happens to be a <style> or <script> tag, we move onto the second instead. As long as that new node is not a formatting element (<strong>, <em>, etc.), we set it as the new element before continuing our loop.

testSelectorEl = $(this);
var i = 8;
while(i > 0){
	if ( $(testSelectorEl).children().length > 0 ) {
	  nextEl = $(testSelectorEl).find(">:first-child");
	  if(nextEl.is( "style" ) || nextEl.is( "script" ) || nextEl.is( "noscript" )){
	     nextEl = $(testSelectorEl).find(">:nth-child(2)");
	  }
	  if ( !nextEl.is( "u" ) && !nextEl.is( "i" ) && !nextEl.is( "strong" ) && !nextEl.is( "em" )) {
	    testSelectorEl = nextEl;
	  }
	}
	i--;
}

The above code is added to the click event listener on all elements within the iFrame. Next, below that code, we determine a unique CSS selector by which we can later reference the element. This is important for writing any changes to the JavaScript snippet to effect changes on the 3rd-party website.

var node = testSelectorEl;
var path = "";
while (node.length) {
  var realNode = node[0], name = realNode.localName;
  if (!name) break;
  name = name.toLowerCase();

  var parent = node.parent();

  var siblings = parent.children(name);
  if (siblings.length > 1) { 
      name += ':eq(' + siblings.index(realNode) + ')';
  }

  path = name + (path ? '>' + path : '');
  node = parent;
}
var value = path;
$(".selector-input").val(value);

The final iFrame on-load function looks like this:

var url = "<?php echo $url; ?>";
var ssl = false;
var pageIframe;
 
if(url.length > 0){

  if(url.indexOf("https:") !== -1){
    ssl = true
  }
   
  var baseUrl = removeSubdirectoryFromUrlString(url, ssl);
  if(baseUrl.slice(-1) !== "/"){
    baseUrl = baseUrl + "/";
  }

  var style = "<style>.highlighted{background-color:rgba(255, 0, 0, 0.5);} </style>";


  var pageIframe = $('<iframe id="page-iframe" style="" src="/new-page-to-edit.php?baseUrl='+baseUrl+'&url=<?php echo $url; ?>"></iframe>').appendTo(".page-iframe-wrapper");
     
  pageIframe.on('load', function(){

    pageIframe.contents().find("body").prepend(style);
       
    pageIframe.contents().find("body *").mouseenter(function(){

      $(this).addClass('highlighted'); 

      testSelectorEl = $(this);
        
    }).mouseout(function(){

      $(this).removeClass('highlighted');   

    }).click(function(e){

      e.preventDefault();
      e.stopPropagation();
      
      //dig deeper down the dom
      var i = 8;
      while(i > 0){
        if ( $(testSelectorEl).children().length > 0 ) {
          nextEl = $(testSelectorEl).find(">:first-child");
          if(nextEl.is( "style" ) || nextEl.is( "script" ) || nextEl.is( "noscript" )){
             nextEl = $(testSelectorEl).find(">:nth-child(2)");
          }
          if ( !nextEl.is( "u" ) && !nextEl.is( "i" ) && !nextEl.is( "strong" ) && !nextEl.is( "em" )) {
            testSelectorEl = nextEl;
          }
        }
        i--;
      }
      
      var node = testSelectorEl;
      var path = "";
      while (node.length) {
          var realNode = node[0], name = realNode.localName;
          if (!name) break;
          name = name.toLowerCase();

          var parent = node.parent();

          var siblings = parent.children(name);
          if (siblings.length > 1) { 
              name += ':eq(' + siblings.index(realNode) + ')';
          }

          path = name + (path ? '>' + path : '');
          node = parent;
      }
      var value = path;
                
      $(".selector-input").val(value); //for html insert section (redundant for change element section)
      if(! $(".insert-html-wrap").is(':visible')){
        selectNewElement(value); // prepare editor options        
        $(".page-editor-info").offset().top; //scroll user to selector input
      }
      return false;
    });

   
   //make sure images load   
   pageIframe.contents().find("img").each(function(){
      var src = $(this).attr("src");
      if(src && src.length > 0 && src.indexOf("//") == -1){  //if not absolute reference
        if(src.charAt(0) !== "/"){
          src = "/" + src;
        }
        $(this).attr("src", baseUrl + src);
      }
   });
   
   //make sure links aren't followed  
   pageIframe.contents().find("a").each(function(){
      var href = $(this).attr("href");
      $(this).attr("href", "");
      $(this).attr("data-href", href);
   });

   pageIframe.contents().find("body").attr("style", "cursor: pointer !important");
    
   $(".loading-iframe").hide();

  }); //page-iframe load
   
}else{
  //no URL found
   $(".loading-iframe").hide();
}

Lastly, we prepare the editor options panel. This involves setting the original content for the selected element and removing any newly added features (modals, sticky-bars) that were not saved.

Editor Options

After an element is clicked its content and style properties are loaded into the editor panel. The original values are stored in memory. As users edit values a change indicator icon is revealed and the right-panel editor view is updated in real time.

var testSelectorEl;
var testSelectorElPath = "";
var testSelectorElHtml = "";
var testSelectorElImage = "";
var testSelectorElLink = "";
var originalVisibilityState = "";
var originalValues = [];
originalValues['height'] = "";
originalValues['width'] = "";
originalValues['margin'] = "";
originalValues['padding'] = "";
originalValues['border'] = "";
originalValues['font-family'] = "";
originalValues['font-weight'] = "";
originalValues['font-style']= "";
originalValues['text-decoration'] = "";
originalValues['background'] = "";
originalValues['css'] = "";
originalValues['classes'] = "";

$(".html-input").keyup(function(){
    var value = $(this).val();  
    if (value !== testSelectorElHtml){
      $(this).parent().find(".change-indicator").show();
    }else{
      $(this).parent().find(".change-indicator").hide();
    }

    if($(".change-indicator").is(":visible")){
      $(".element-change-save-btn").removeAttr("disabled");
    }else{
      $(".element-change-save-btn").attr("disabled", "disabled");
    }

    var selector = $(".selector-input").val();
    var iFrameDOM = $("iframe#page-iframe").contents()
    iFrameDOM.find(selector).html(value);
    
  });

Besides changing existing elements, new ones can also be added. The “insert content” section, also based on selecting an element, lets users add new text, html, or images before or after whatever has been clicked. In both sections, the CSS selector can also be manually typed (instead of clicking). Adding or editing images is handled by a custom built image upload gallery that leverages AWS S3 and PHP.

Out-of-the-box elements, such as sticky bars and modals, can also be added with a few clicks and configurations. The HTML and CSS for those are pre-built, with variables for any options that may be set. Any changes made are saved to the database in relation to the user account, project, experiment, and variation.

function addSticky(){
	$conn = $this->connection;
	$accountid = $this->accountid;
	$variationid = $_GET['variationid'];
	$experimentid = $_GET['experimentid'];
	$text = $_POST['text'];
	$color = $_POST['color'];
	$background = $_POST['background'];
	$position = $_POST['position'];
	$linkurl = $_POST['linkurl'];

	$insertStatement = "INSERT INTO `variationchange` (accountid, variationid, experimentid, selector, changetype, changecode, content) VALUES (:accountid, :variationid, :experimentid, :selector, :changetype, :changecode, :content)";
	$changetype = "stickybar";
	$selector = "body";

	$link_opening = "";
	$link_closing = "";

	if(strlen($linkurl) > 0){
		$link_opening = "<a style='color:".$color."' href='".$linkurl."'>";
		$link_closing = "</a>";
	}

	$sticky_html = "<div style='font-weight:bold;".$position.":0;position:fixed;z-index:100000;left:0px;text-align:center;padding:8px 20px;width:100%;background:".$background.";color:".$color."' id='splitwit-sticky'><p style='margin:0px'>".$link_opening.$text.$link_closing."</p></div>";

	$changecode = '$("body").append("'.$sticky_html.'")';

	$stmt = $conn->prepare($insertStatement);	
	$stmt->bindParam(':accountid', $accountid);
	$stmt->bindParam(':variationid', $variationid);
	$stmt->bindParam(':experimentid', $experimentid);
	$stmt->bindParam(':selector', $selector);
	$stmt->bindParam(':changetype', $changetype);
	$stmt->bindParam(':changecode', $changecode);
	$stmt->bindParam(':content', $text);
	$stmt->execute();	

	$this->writeSnippetFile($variationid);
}

Writing changes to the snippet

Every project has a unique JavaScript file that users must add to their webpage. The file is hosted by SplitWit, so website owners only need to copy/paste a snippet. The WordPress and Shopify plugins automatically add the snippet, making it even more friendly to non-developers.

Each project may contain multiple experiments with their own changes, metrics, variations, and conditions. After the data is produced by a SQL join statement, it is massaged into a nested object and parsed in JSON. That JavaScript output is concatenated to the necessary libraries and helper functions.

function writeSnippetFile($variationid=false, $experimentid=false, $projectid=false){
	
	$conn = $this->connection;
	$variationid = $variationid || false;
	$experimentid = $experimentid || false;
	
	if(isset($_GET['variationid'])){
		$variationid = $_GET['variationid'];
	}
	if(isset($_GET['experimentid'])){
		$experimentid = $_GET['experimentid'];
	}


	if($variationid){
		$variationid = $_GET['variationid'];
		$sql = "SELECT experiment.projectid FROM `variation` right join `experiment` on variation.experimentid = experiment.experimentid WHERE variationid = ?"; 
		$result = $conn->prepare($sql); 
		$result->execute(array($variationid));
		$experiment_row = $result->fetch(PDO::FETCH_ASSOC);
	}elseif($experimentid){
		$experimentid = $_GET['experimentid'];
		$sql = "SELECT projectid FROM `experiment` WHERE experimentid = ?"; 
		$result = $conn->prepare($sql); 
		$result->execute(array($experimentid));
		$experiment_row = $result->fetch(PDO::FETCH_ASSOC);
	}
	
	if(!$projectid){
		$projectid = $experiment_row['projectid'];
	}

	$sql = "SELECT experiment.experimentid, experiment.status, experimentcondition.experimentconditionid, variation.variationid, variationchange.variationchangeid, variationchange.changecode, variation.css, variation.javascript, experimentcondition.url, experimentcondition.matchtype, experimentcondition.conditiontype, metric.metricid, metric.type, metric.input, metric.urlmatch FROM `experiment` left join `metric` on metric.experimentid = experiment.experimentid join `experimentcondition` on experimentcondition.experimentid = experiment.experimentid join `variation` on variation.experimentid = experiment.experimentid left join `variationchange` on variationchange.variationid = variation.variationid WHERE experiment.projectid = ?"; 
	
	// echo "<br />"."variationid: ".$variationid . "<br />";
	// echo "<br />"."experimentid: ".$experimentid . "<br />";
	// echo "<br />"."projectid: ".$projectid . "<br />";

	// $this->status = "SQL: " . $sql;
	$result = $conn->prepare($sql); 
	$result->execute(array($projectid));
	 
	$experiment_row = $result->fetchAll(PDO::FETCH_ASSOC);
	 
	//turn flat array, into a nested one
	$endResult = array();
	foreach($experiment_row as $row){
	    if (!isset($endResult[$row['experimentid']])){
	        $endResult[$row['experimentid']] = array(
	            'experimentid' => $row['experimentid'],
	            'status' => $row['status'],
	            'conditions' => array(),
	            'variations' => array(),
	            'metrics' => array()
	        );
	    }

	    if (!isset($endResult[$row['experimentid']]['conditions'][$row['experimentconditionid']])){
		    $endResult[$row['experimentid']]['conditions'][$row['experimentconditionid']] = array(
		        'experimentconditionid' => $row['experimentconditionid'],
		        'url' => $row['url'],
		        'matchtype' => $row['matchtype'],
		        'conditiontype' => $row['conditiontype']
		    );
		}

	    if (!isset($endResult[$row['experimentid']]['variations'][$row['variationid']])){
		    $endResult[$row['experimentid']]['variations'][$row['variationid']] = array(
		        'variationid' => $row['variationid'],
		        'javascript' => $row['javascript'],
		        'css' => $row['css'],
		        'changes' => array()
		    );
		}

	    if (!isset($endResult[$row['experimentid']]['variations'][$row['variationid']]['changes'][$row['variationchangeid']])){
		   	$endResult[$row['experimentid']]['variations'][$row['variationid']]['changes'][$row['variationchangeid']] = array(
		        'variationchangeid' => $row['variationchangeid'],
		        'changecode' => $row['changecode']
		    );
		}

	    if (!isset($endResult[$row['experimentid']]['metrics'][$row['metricid']])){
		    $endResult[$row['experimentid']]['metrics'][$row['metricid']] = array(
		        'metricid' => $row['metricid'],
		        'type' => $row['type'],
		        'input' => $row['input'],
		        'urlmatch' => $row['urlmatch']
		    );
		}
		
	}

	$json_output = json_encode($endResult);
	// echo $json_output;
	// get snippet file name.
	$sql = "SELECT snippet from `project` where projectid=?";
	$result = $conn->prepare($sql); 
	$result->execute(array($projectid));
	$projectrow = $result->fetch(PDO::FETCH_ASSOC);
	$filename = $projectrow['snippet'];
	// echo $filename;
	$snippet_template = file_get_contents("/var/www/html/snippet/snippet-template.min.js");
	// concat json to snippet template. write file			
	$myfile = fopen("/var/www/html/snippet/".$filename, "w") or die("Unable to open file!");
	//if there are any animation changes, include the necessary library.
	$txt = "";
	if(strpos($json_output, "addClass('animated") !== false){
		$txt .= 'var head=document.getElementsByTagName("head")[0],link=document.createElement("link");link.rel="stylesheet";link.type="text/css";link.href="https://www.splitwit.com/css/animate.min.css";link.media="all";head.appendChild(link); ';
	}
	$txt .= "window.splitWitExperiments = ".$json_output . "\n" . $snippet_template;
	
	fwrite($myfile, $txt) or die("Unable to save file!");

	fclose($myfile);

	//if this project is for a shopify app, we need to update the snippet cache
	$update_snippet_url = "https://www.splitwit.com/service-layer/shopify-app-service.php?method=updateSnippetScriptTag&projectid=".$projectid;
	$params = [];
	$curl_update_snippet = $this->curlApiUrl($update_snippet_url, $params);

}

One of the helper functions is a method that checks if the current URL matches the experiment’s conditions to run on:

function testUrl(testurl, conditions){
	
	if(testurl.search(/(?:[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?\.)+[a-z0-9][a-z0-9-]{0,61}[a-z0-9]/) < 0){
		console.log("no domain")
		return window.inputError($(".test-url-input"), "Please test a valid URL.");
	}
	var valid = false;

	for (i = 0; i < conditions.length; i++) { 
		 var url = conditions[i].url;
		 var matchtype = conditions[i].matchtype;
		 var conditiontype = conditions[i].conditiontype;

		 //exact match
		 if(matchtype == "exact" && conditiontype == "target" && url == testurl){
		 	valid = true;
		 }
		 if(matchtype == "exact" && conditiontype == "exclude" && url == testurl){
		 	valid = false;
		 }

		 //basic match
		 if(matchtype == "basic"){
		 	//strip off querystring, hashtags, protocols, and www from url and testurl - then compare
		 	var cleanTestUrl = testurl.toLowerCase();
		 	var cleanUrl = url.toLowerCase();

		 	if (cleanTestUrl.indexOf("?") > 0) {
					cleanTestUrl = cleanTestUrl.substring(0, cleanTestUrl.indexOf("?"));
				}
		 	if (cleanUrl.indexOf("?") > 0) {
					cleanUrl = cleanUrl.substring(0, cleanUrl.indexOf("?"));
				}
		 	if (cleanTestUrl.indexOf("&") > 0) {
					cleanTestUrl = cleanTestUrl.substring(0, cleanTestUrl.indexOf("&"));
				}
		 	if (cleanUrl.indexOf("&") > 0) {
					cleanUrl = cleanUrl.substring(0, cleanUrl.indexOf("&"));
				}
		 	if (cleanTestUrl.indexOf("#") > 0) {
					cleanTestUrl = cleanTestUrl.substring(0, cleanTestUrl.indexOf("#"));
				}
		 	if (cleanUrl.indexOf("#") > 0) {
					cleanUrl = cleanUrl.substring(0, cleanUrl.indexOf("#"));
				}
				cleanTestUrl = cleanTestUrl.replace(/^(?:https?:\/\/)?(?:www\.)?/i, "");
				cleanUrl = cleanUrl.replace(/^(?:https?:\/\/)?(?:www\.)?/i, "");
				cleanTestUrl = cleanTestUrl.replace(/\/$/, "");
				cleanUrl = cleanUrl.replace(/\/$/, ""); //remove trailing slash

				// console.log(cleanTestUrl);
				// console.log(cleanUrl);
				if(conditiontype == "target" && cleanUrl == cleanTestUrl){
					valid = true;
				}
				if(conditiontype == "exclude" && cleanUrl == cleanTestUrl){
					valid = false;
				}

		 }
		 
		 //substring match
		 if(matchtype == "substring"){
		 	if(testurl.includes(url) && conditiontype == "target"){
					valid = true;
		 	}
		 	if(testurl.includes(url) && conditiontype == "exclude"){
					valid = false;
		 	}
		 }
		 
	} //end conditions loop
	
	$(".test-url-msg").hide();		
	if(valid){
		$(".valid-msg").fadeIn();
	}else{
		$(".invalid-msg").fadeIn();
	}

}

While the snippet code is specific to this use-case (experimental A/B UI changes), the visual editor can be used in a variety of other contexts. You can look through more of the code I used in this GitHub repository I created.