Get the last "word" of a PATH with different URL formats

Question

I am creating a function that gets the last "word" of a requested url in php, without considering parameters and considering the root as index.

Examples:

URL http://www.teste.com.br/ EXPECTATION index

URL www.teste.com.br/ EXPECTATIVA index

URL test.with EXPECTATION index

URL test.com/ EXPECTATION index

URL www.teste.com.br/test EXPECTATION test

URL http://www.teste.com.br/test TEST EXPECTATION

URL http://test.com/test TEST EXPECTATION

URL https://www.teste.com/test TEST EXPECTATION

URL https://test.com/test TEST EXPECTATION

URL test.com/test/two EXPECTATION 2

URL test.com/test/two/ EXPECTATION TWO

URL test.com/test/two/? variable=test EXPECTATION two

URL test.com/test/two? variable=test EXPECTATION TWO

URL test.com/test/two/? variable=test EXPECTATION two

URL test.com/test? var1=t&var 2=t EXPECTATION test

URL test.com/test/tres#ola EXPECTATION

URL test.com/test? var1=t&var 2=t#ola EXPECTATION test

Using the function basename and working with substr and preg_match I get a certain success rate:

$arr = array(
  array("name"=>"http://www.teste.com.br/","possibleValues"=>array("index")),
  array("name"=>"www.teste.com.br/","possibleValues"=>array("index")),
  array("name"=>"teste.com","possibleValues"=>array("index")),
  array("name"=>"teste.com/","possibleValues"=>array("index")),
  array("name"=>"www.teste.com.br/teste","possibleValues"=>array("teste")),
  array("name"=>"http://www.teste.com.br/teste","possibleValues"=>array("teste")),
  array("name"=>"http://teste.com/teste","possibleValues"=>array("teste")),
  array("name"=>"https://www.teste.com/teste","possibleValues"=>array("teste")),
  array("name"=>"https://teste.com/teste","possibleValues"=>array("teste")),
  array("name"=>"teste.com/teste/dois","possibleValues"=>array("dois")),
  array("name"=>"teste.com/teste/dois/","possibleValues"=>array("dois")),
  array("name"=>"teste.com/teste/dois/?variavel=teste","possibleValues"=>array("dois")),
  array("name"=>"teste.com/teste/dois?variavel=teste","possibleValues"=>array("dois")),
  array("name"=>"teste.com/teste/dois/?variavel=teste","possibleValues"=>array("dois")),
  array("name"=>"teste.com/teste?var1=t&var2=t","possibleValues"=>array("teste")),
  array("name"=>"teste.com/teste/tres#ola","possibleValues"=>array("tres")),
  array("name"=>"teste.com/teste?var1=t&var2=t#ola","possibleValues"=>array("teste"))
);

foreach($arr as $value){
  echo "URL ".$value["name"]."\n";
  echo ( array_search( basename( returnLastWord( $value["name"] ) ), $value["possibleValues"] ) === false ? "FALHOU" : "PASSOU" )." -> expected: ".json_encode( $value["possibleValues"] )." get '".basename( returnLastWord( $value["name"] ) )."'\n\n";
}

function returnLastWord($var){
  preg_match('/[?#]/', $var, $matches, PREG_OFFSET_CAPTURE);
  $after = ( empty( $matches[0][1] ) ? NULL : $matches[0][1] );
  if($after){
    return substr($var, 0, $after);
  }else{
    // echo "aqui\n";
    return $var;
  }
}

https://repl.it/JJYI/7

---

URL http://www.teste.com.br/
FALHOU -> expected: ["index"] get 'www.teste.com.br'

URL www.teste.com.br/
FALHOU -> expected: ["index"] get 'www.teste.com.br'

URL teste.com
FALHOU -> expected: ["index"] get 'teste.com'

URL teste.com/
FALHOU -> expected: ["index"] get 'teste.com'

URL www.teste.com.br/teste
PASSOU -> expected: ["teste"] get 'teste'

URL http://www.teste.com.br/teste
PASSOU -> expected: ["teste"] get 'teste'

URL http://teste.com/teste
PASSOU -> expected: ["teste"] get 'teste'

URL https://www.teste.com/teste
PASSOU -> expected: ["teste"] get 'teste'

URL https://teste.com/teste
PASSOU -> expected: ["teste"] get 'teste'

URL teste.com/teste/dois
PASSOU -> expected: ["dois"] get 'dois'

URL teste.com/teste/dois/
PASSOU -> expected: ["dois"] get 'dois'

URL teste.com/teste/dois/?variavel=teste
PASSOU -> expected: ["dois"] get 'dois'

URL teste.com/teste/dois?variavel=teste
PASSOU -> expected: ["dois"] get 'dois'

URL teste.com/teste/dois/?variavel=teste
PASSOU -> expected: ["dois"] get 'dois'

URL teste.com/teste?var1=t&var2=t
PASSOU -> expected: ["teste"] get 'teste'

URL teste.com/teste/tres#ola
PASSOU -> expected: ["tres"] get 'tres'

URL teste.com/teste?var1=t&var2=t#ola
PASSOU -> expected: ["teste"] get 'teste'

---

I’m having trouble mainly in the first 4 examples, where theoretically it would be the root of the project, ie should get the index

Answer 1

I couldn’t find a simple way to do it, especially with variations of input URL types, what I got was this:

function returnLastWord($var){

    //Remove o protocolo
    $var = preg_replace('~^[^:]+[:][/]{2,}~', '', $var);

    /*
    Pega qualquer coisa que seja um PATH em URLs
    pega o que esta entre o parenteses neste exemplo:
    `site.com/(foo/bar/baz)?querystring=ignorada#hashignorada`
    */
    if (preg_match('~/([^#?]{1,})~', $var, $matches)) {

        //Remove o / do final em urls como `foo/bar/`, para evitar pegar em branco
        $result = rtrim($matches[1], '/');

        //Pega qualquer coisa que estiver no final
        if (preg_match('~[^/]+$~', $result, $matches)) {
          return $matches[0];
        }
    }

    //Se qualquer coisa anterior falhou é porque provavelmente é "index"
    return 'index';
}

Example in Ideone

It is likely that I will review this to make it more performative or simple.

Answer 2

PHP already has a native function to work with URL, but as not all are in the format defined in RFC-3986, the function ends up analyzing incorrectly those that are not standardized. Nothing critical. What happens is that the function considers what should be host as part of path, then a check if there is the character . in the path is therefore necessary, if any, the element in question is the host, not the path, returning, like this, index.

function returnLastWord ($url) {

    // Analisa a URL:
    $url = parse_url($url);

    // Divide o path nas ocorrências de /:
    $parts = explode('/', trim($url["path"], '/'));

    // Busca o último elemento:
    $last = end($parts);

    // Se não estiver vazio e não possuir o caractere ., retorna o valor, senão retorna index:
    return $last && false === strpos($last, '.') ? $last : "index";

}

See working on Ideone.

Answer 3

Using the function parse_url() can facilitate the work.

<?php

$arr = array(
    'http://localhost/',
    'https://localhost/',
    'http://localhost',
    'https://localhost',
    'http://sub.localhost/',
    'http://sub.localhost',
    'http://localhost/foo',
    'http://localhost/foo/bar',
    'http://localhost/?p=1&b=1',
    'localhost/foo',
    'localhost'
);

echo '<table border=1>
<tr><td>URL</td><td>Word</td></tr>';
foreach ($arr as $v) {
    $url_original = null;
    // Normalizing the given URL
    if (preg_match('#^https?://#i', $v) !== 1) {
        $url_original = $v;
        $v = 'http://'.$v;
    }

    $url = parse_url($v);
    echo PHP_EOL.'<tr><td>'.$v.(!empty($url_original)? '<br>('.$url_original.')': '').'</td><td>';
    if (
        isset($url['path'])
        && !empty($url['path'])
        && $url['path'] != '/'
    ) {
        // Path found
        $p = strrpos($url['path'], '/');
        if ($p !== false) {
            echo substr($url['path'], $p+1);
        }
    } else {
        // Empty, no path
        echo 'index';
    }
    echo '</td></tr>';
}
echo '</table>';

Normalization
In order to work even with Urls without schema (http or https), a normalization is made where "http://" is prefixed to string before moving to function parse_url().

Multibyte characters
The section that invokes the function substr() may fail when the URL has multibyte characters. If you want to provide multibyte character support, refer to functions mbstring.