从文件中获取类名

I have a php file which contains only one class. how can I know what class is there by knowing the filename? I know I can do something with regexp matching but is there a standard php way? (the file is already included in the page that is trying to figure out the class name).

There are multiple possible solutions to this problem, each with their advantages and disadvantages. Here they are, it's up to know to decide which one you want.

Tokenizer

This method uses the tokenizer and reads parts of the file until it finds a class definition.

Advantages

  • Do not have to parse the file entirely
  • Fast (reads the beginning of the file only)
  • Little to no chance of false positives

Disadvantages

  • Longest solution

Code

$fp = fopen($file, 'r');
$class = $buffer = '';
$i = 0;
while (!$class) {
    if (feof($fp)) break;

    $buffer .= fread($fp, 512);
    $tokens = token_get_all($buffer);

    if (strpos($buffer, '{') === false) continue;

    for (;$i<count($tokens);$i++) {
        if ($tokens[$i][0] === T_CLASS) {
            for ($j=$i+1;$j<count($tokens);$j++) {
                if ($tokens[$j] === '{') {
                    $class = $tokens[$i+2][1];
                }
            }
        }
    }
}

Regular expressions

Use regular expressions to parse the beginning of the file, until a class definition is found.

Advantages

  • Do not have to parse the file entirely
  • Fast (reads the beginning of the file only)

Disadvantages

  • High chances of false positives (e.g.: echo "class Foo {";)

Code

$fp = fopen($file, 'r');
$class = $buffer = '';
$i = 0;
while (!$class) {
    if (feof($fp)) break;

    $buffer .= fread($fp, 512);
    if (preg_match('/class\s+(\w+)(.*)?\{/', $buffer, $matches)) {
        $class = $matches[1];
        break;
    }
}

Note: The regex can probably be improved, but no regex alone can do this perfectly.

Get list of declared classes

This method uses get_declared_classes() and look for the first class defined after an include.

Advantages

  • Shortest solution
  • No chance of false positive

Disadvantages

  • Have to load the entire file
  • Have to load the entire list of classes in memory twice
  • Have to load the class definition in memory

Code

$classes = get_declared_classes();
include 'test2.php';
$diff = array_diff(get_declared_classes(), $classes);
$class = reset($diff);

Note: You cannot simply do end() as others suggested. If the class includes another class, you will get a wrong result.


This is the Tokenizer solution, modified to include a $namespace variable containing the class namespace, if applicable:

$fp = fopen($file, 'r');
$class = $namespace = $buffer = '';
$i = 0;
while (!$class) {
    if (feof($fp)) break;

    $buffer .= fread($fp, 512);
    $tokens = token_get_all($buffer);

    if (strpos($buffer, '{') === false) continue;

    for (;$i<count($tokens);$i++) {
        if ($tokens[$i][0] === T_NAMESPACE) {
            for ($j=$i+1;$j<count($tokens); $j++) {
                if ($tokens[$j][0] === T_STRING) {
                     $namespace .= '\\'.$tokens[$j][1];
                } else if ($tokens[$j] === '{' || $tokens[$j] === ';') {
                     break;
                }
            }
        }

        if ($tokens[$i][0] === T_CLASS) {
            for ($j=$i+1;$j<count($tokens);$j++) {
                if ($tokens[$j] === '{') {
                    $class = $tokens[$i+2][1];
                }
            }
        }
    }
}

Say you have this class:

namespace foo\bar {
    class hello { }
}

...or the alternative syntax:

namespace foo\bar;
class hello { }

You should have the following result:

var_dump($namespace); // \foo\bar
var_dump($class);     // hello

You could also use the above to detect the namespace a file declares, regardless of it containing a class or not.

You can do this in two ways:

  • complex solution: open the file and through regex extract the class-name (like /class ([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*)/)
  • simply solution: name all your php files with the class-name contained (eg: the class TestFoo in the file TestFoo.php or TestFoo.class.php)

You may be able to use the autoload function.

function __autoload($class_name) {
    include "special_directory/" .$class_name . '.php';
}

And you can echo $class_name. But that requires a directory with a single file.

But it is standard practice to have one class in each file in PHP. So Main.class.php will contain Main class. You may able to use that standard if you are the one coding.

You could get all declared classes before you include the file using get_declared_classes. Do the same thing after you have included it and compare the two with something like array_diff and you have your newly added class.

You can make PHP do the work by just including the file and get the last declared class:

$file = 'class.php'; # contains class Foo

include($file);
$classes = get_declared_classes();
$class = end($classes);
echo $class; # Foo

If you need to isolate that, wrap it into a commandline script and execute it via shell_exec:

$file = 'class.php'; # contains class Foo

$class = shell_exec("php -r \"include('$file'); echo end(get_declared_classes());\"");    
echo $class; # Foo

If you dislike commandline scripts, you can do it like in this question, however that code does not reflect namespaces.

$st = get_declared_classes();
include "classes.php"; //one or more classes in file, contains class class1, class2, etc...

$res = array_values(array_diff_key(get_declared_classes(),$st));
print_r($res); # Array ([0] => class1 [1] => class2 [2] ...)

This sample returns all classes. If you're looking for a class which is derived of a specific one, use is_subclass_of

$php_code = file_get_contents ( $file );
    $classes = array ();
    $namespace="";
    $tokens = token_get_all ( $php_code );
    $count = count ( $tokens );

    for($i = 0; $i < $count; $i ++)
    {
        if ($tokens[$i][0]===T_NAMESPACE)
        {
            for ($j=$i+1;$j<$count;++$j)
            {
                if ($tokens[$j][0]===T_STRING)
                    $namespace.="\\".$tokens[$j][1];
                elseif ($tokens[$j]==='{' or $tokens[$j]===';')
                    break;
            }
        }
        if ($tokens[$i][0]===T_CLASS)
        {
            for ($j=$i+1;$j<$count;++$j)
                if ($tokens[$j]==='{')
                {
                    $classes[]=$namespace."\\".$tokens[$i+2][1];
                }
        }
    }
    return $classes;

Thanks to some people from Stackoverflow and Github, I was able to write this amazing fully working solution:

/**
 * get the full name (name 
amespace) of a class from its file path
 * result example: (string) "I\Am\The\Namespace\Of\This\Class"
 *
 * @param $filePathName
 *
 * @return  string
 */
public function getClassFullNameFromFile($filePathName)
{
    return $this->getClassNamespaceFromFile($filePathName) . '\\' . $this->getClassNameFromFile($filePathName);
}


/**
 * build and return an object of a class from its file path
 *
 * @param $filePathName
 *
 * @return  mixed
 */
public function getClassObjectFromFile($filePathName)
{
    $classString = $this->getClassFullNameFromFile($filePathName);

    $object = new $classString;

    return $object;
}





/**
 * get the class namespace form file path using token
 *
 * @param $filePathName
 *
 * @return  null|string
 */
protected function getClassNamespaceFromFile($filePathName)
{
    $src = file_get_contents($filePathName);

    $tokens = token_get_all($src);
    $count = count($tokens);
    $i = 0;
    $namespace = '';
    $namespace_ok = false;
    while ($i < $count) {
        $token = $tokens[$i];
        if (is_array($token) && $token[0] === T_NAMESPACE) {
            // Found namespace declaration
            while (++$i < $count) {
                if ($tokens[$i] === ';') {
                    $namespace_ok = true;
                    $namespace = trim($namespace);
                    break;
                }
                $namespace .= is_array($tokens[$i]) ? $tokens[$i][1] : $tokens[$i];
            }
            break;
        }
        $i++;
    }
    if (!$namespace_ok) {
        return null;
    } else {
        return $namespace;
    }
}

/**
 * get the class name form file path using token
 *
 * @param $filePathName
 *
 * @return  mixed
 */
protected function getClassNameFromFile($filePathName)
{
    $php_code = file_get_contents($filePathName);

    $classes = array();
    $tokens = token_get_all($php_code);
    $count = count($tokens);
    for ($i = 2; $i < $count; $i++) {
        if ($tokens[$i - 2][0] == T_CLASS
            && $tokens[$i - 1][0] == T_WHITESPACE
            && $tokens[$i][0] == T_STRING
        ) {

            $class_name = $tokens[$i][1];
            $classes[] = $class_name;
        }
    }

    return $classes[0];
}

I modified Nette\Reflection\AnnotationsParser that so it returns an array of namespace+classname that are defined in the file

$parser = new PhpParser();
$parser->extractPhpClasses('src/Path/To/File.php');


class PhpParser
{
    public function extractPhpClasses(string $path)
    {
        $code = file_get_contents($path);
        $tokens = @token_get_all($code);
        $namespace = $class = $classLevel = $level = NULL;
        $classes = [];
        while (list(, $token) = each($tokens)) {
            switch (is_array($token) ? $token[0] : $token) {
                case T_NAMESPACE:
                    $namespace = ltrim($this->fetch($tokens, [T_STRING, T_NS_SEPARATOR]) . '\\', '\\');
                    break;
                case T_CLASS:
                case T_INTERFACE:
                    if ($name = $this->fetch($tokens, T_STRING)) {
                        $classes[] = $namespace . $name;
                    }
                    break;
            }
        }
        return $classes;
    }

    private function fetch(&$tokens, $take)
    {
        $res = NULL;
        while ($token = current($tokens)) {
            list($token, $s) = is_array($token) ? $token : [$token, $token];
            if (in_array($token, (array) $take, TRUE)) {
                $res .= $s;
            } elseif (!in_array($token, [T_DOC_COMMENT, T_WHITESPACE, T_COMMENT], TRUE)) {
                break;
            }
            next($tokens);
        }
        return $res;
    }
}

I spent lots of productive time looking for a way around this. From @netcoder's solution its obvious there are lots of cons in all the solutions so far.

So I decided to do this instead. Since most PHP classes has class name same as filename, we could get the class name from the filename. Depending on your project you could also have a naming convention. NB: This assume class does not have namespace

<?php
    $path = '/path/to/a/class/file.php';

    include $path;

    /*get filename without extension which is the classname*/
    $classname = pathinfo(basename($path), PATHINFO_FILENAME);

    /* you can do all this*/
    $classObj = new $classname();

    /*dough the file name is classname you can still*/
    get_class($classObj); //still return classname