Going along with Neville K's answer here is an example of how my company handles RESTful api calls.
First we have a php file that handles the calls with a switch statement. Routing the different actions to said functions and classes.
/* Class file that is called on this page */
include_once "$_SERVER[DOCUMENT_ROOT]/classes/class.myclass.php";
/**
* This function makes it simpler to stop it from working for debugging purposes.
* All we have to do is comment out the one line of code apiCall($_REQUEST);
* You could have this outside of the function and it would work just as well.
* @param type $REQUEST
*/
function apiCall($REQUEST) {
$con = new MyClass();
switch ($REQUEST['action']) {
case 'getList':
/* Setting the content type to json means that the developer can
* expect a response in the form of parseable json.
*/
header('Content-Type: application/json');
echo json_encode($con->getList($REQUEST));
case 'setValue':
header('Content-Type: application/json');
echo json_encode($con->setValue($REQUEST));
case 'login':
if ($con->login($REQUEST)) {
header('Location: /index.php');
} else {
header('Content-Type: /login.php?status=Failed+Login');
}
default:
header('Content-Type: application/json');
/* If an invalid action was sent in, then this error message will be sent
* back to the user
*/
echo json_encode(['status' => 'Invalid API Call']);
}
}
/* Using $_REQUEST allows developers to access the api via GET or POST */
apiCall($_REQUEST);
Then we handle all the logic in the different classes we called.
class MyClass {
public function getList($REQUEST) {
$id = $REQUEST['id'];
/* code */
return ['status' => 'ok', 'results' => $array];
}
public function setList($REQUEST) {
/* code */
return ['status' => 'ok'];
}
public function login($REQUEST) {
/* code */
$_SESSION['user_id'] = $user_id;
return $login_successful;
}
}
Using JSON is good for applications that send information via AJAX calls. Using the header('Location:') are good for form submissions without ajax.
You can then use JavaScript ajax calls or for submissions based on how you handle the submission of data.
Example of using jQuery.getJSON
$.getJSON('/switch.php', $.param({id: id, action: 'getList'}), function (json) {
if (json) {
/*code*/
}
});
Then you would pass a hidden input with action in it to the switch page for regular form submissions.
<form action="/switch.php" method="post">
<!--hidden input named action to direct which switch to use-->
<input name="action" value="login" type="hidden"/>
<input name="username"/>
<input name="password" type="password"/>
<input type="submit"/>
</form>
These examples are for html/JavaScript web applications. If you are using JAVA, Python, .NET, or some other language, it would be as simple as using the REST API and parsing out the JSON to figure out how to handle your application logic.
You can even run a php to php api call using file_get_contents or curl.
$data = [
'action' => 'setValue',
'information' => 'More'
];
$json = json_decode(file_get_contents('/switch.php?' . http_build_query($data)),true);
if(!empty($json)){
/*code*/
}
You could create a seperate page for each call and not have to worry about passing in an action to every request. But then your filetree starts to look like this.
/api/loginSubmit.php
/api/login.php
/api/getListFromId.php
/api/getList.php
/api/setValues.php
/api/getValues.php
It gets really tedious to traverse all these files to figure out where the problem is.
