modules/transliteration/transliteration.module

<?php
// $Id: transliteration.module,v 1.7 2009/10/09 12:19:44 smk Exp $

/**
 * @file
 * Provides transliteration for UTF-8 text input and sanitzes file names.
 *
 * Uses data from the Text::Unidecode Perl library.
 * @see http://search.cpan.org/~sburke/Text-Unidecode-0.04/lib/Text/Unidecode.pm
 */

/**
 * Sanitize a file path.
 *
 * Additionally removes invalid characters from a file name (which may include
 * sub-directories) after transliteration.
 *
 * @param $filename
 *   A file name.
 * @param $source_langcode
 *   Optional ISO 639 language code that denotes the language of the input.
 *   Used to apply language-specific variations and defaults to the current
 *   display language. If transliteration takes place during output (instead
 *   of creation) and the source language is not known at that time, it is
 *   recommended to set this argument to 'en' to produce consistent results
 *   for all enabled languages.
 * @return
 *   Cleaned file name.
 */
function transliteration_clean_filename($filename, $source_langcode = NULL) {
  // Transliterate.
  $filename = transliteration_get($filename, '', $source_langcode);
  // Replace whitespace.
  $filename = str_replace(' ', '_', $filename);
  // Remove remaining unsafe characters.
  $filename = preg_replace('![^0-9A-Za-z_./-]!', '', $filename);
  // Force lowercase to prevent issues on case-insensitive file systems.
  $filename = strtolower($filename);

  return $filename;
}

/**
 * Transliterate UTF-8 text to ASCII.
 *
 * Takes an input string in any language and character set, and tries to
 * represent it in ASCII characters by conveying, in Roman letters, the
 * pronunciation expressed by the text in some other writing system.
 *
 * @param $input
 *   UTF-8 text input.
 * @param $unknown
 *   Replacement string for characters that do not have a suitable ASCII
 *   equivalent.
 * @param $source_langcode
 *   Optional ISO 639 language code that denotes the language of the input.
 *   Used to apply language-specific variations and defaults to the current
 *   display language. If transliteration takes place during output (instead
 *   of creation) and the source language is not known at that time, it is
 *   recommended to set this argument to 'en' to produce consistent results
 *   for all enabled languages.
 * @return
 *   Transliterated text.
 */
function transliteration_get($input, $unknown = '?', $source_langcode = NULL) {
  module_load_include('inc', 'transliteration');
  return transliteration_process($input, $unknown, $source_langcode);
}

/**
 * Implementation of hook_init().
 *
 * Transliterate and clean the names of files currently being uploaded.
 */
function transliteration_init() {
  if (!empty($_FILES['files'])) {
    module_load_include('inc', 'transliteration');

    // Figure out language, which is available for node form submits.
    $langcode = NULL;
    if (!empty($_POST['language'])) {
      $languages = language_list();
      if (isset($languages[$_POST['language']])) {
        $langcode = $_POST['language'];
      }
    }
    foreach ($_FILES['files']['name'] as $field => $filename) {
      // Keep a copy of the unaltered file name.
      $_FILES['files']['orig_name'][$field] = $filename;
      $_FILES['files']['name'][$field] = transliteration_clean_filename($filename, $langcode);
    }
  }
}

1	<?php
2	// $Id: transliteration.module,v 1.7 2009/10/09 12:19:44 smk Exp $
3
4	/**
5	* @file
6	* Provides transliteration for UTF-8 text input and sanitzes file names.
7	*
8	* Uses data from the Text::Unidecode Perl library.
9	* @see http://search.cpan.org/~sburke/Text-Unidecode-0.04/lib/Text/Unidecode.pm
10	*/
11
12	/**
13	* Sanitize a file path.
14	*
15	* Additionally removes invalid characters from a file name (which may include
16	* sub-directories) after transliteration.
17	*
18	* @param $filename
19	* A file name.
20	* @param $source_langcode
21	* Optional ISO 639 language code that denotes the language of the input.
22	* Used to apply language-specific variations and defaults to the current
23	* display language. If transliteration takes place during output (instead
24	* of creation) and the source language is not known at that time, it is
25	* recommended to set this argument to 'en' to produce consistent results
26	* for all enabled languages.
27	* @return
28	* Cleaned file name.
29	*/
30	function transliteration_clean_filename($filename, $source_langcode = NULL) {
31	// Transliterate.
32	$filename = transliteration_get($filename, '', $source_langcode);
33	// Replace whitespace.
34	$filename = str_replace(' ', '_', $filename);
35	// Remove remaining unsafe characters.
36	$filename = preg_replace('![^0-9A-Za-z_./-]!', '', $filename);
37	// Force lowercase to prevent issues on case-insensitive file systems.
38	$filename = strtolower($filename);
39
40	return $filename;
41	}
42
43	/**
44	* Transliterate UTF-8 text to ASCII.
45	*
46	* Takes an input string in any language and character set, and tries to
47	* represent it in ASCII characters by conveying, in Roman letters, the
48	* pronunciation expressed by the text in some other writing system.
49	*
50	* @param $input
51	* UTF-8 text input.
52	* @param $unknown
53	* Replacement string for characters that do not have a suitable ASCII
54	* equivalent.
55	* @param $source_langcode
56	* Optional ISO 639 language code that denotes the language of the input.
57	* Used to apply language-specific variations and defaults to the current
58	* display language. If transliteration takes place during output (instead
59	* of creation) and the source language is not known at that time, it is
60	* recommended to set this argument to 'en' to produce consistent results
61	* for all enabled languages.
62	* @return
63	* Transliterated text.
64	*/
65	function transliteration_get($input, $unknown = '?', $source_langcode = NULL) {
66	module_load_include('inc', 'transliteration');
67	return transliteration_process($input, $unknown, $source_langcode);
68	}
69
70	/**
71	* Implementation of hook_init().
72	*
73	* Transliterate and clean the names of files currently being uploaded.
74	*/
75	function transliteration_init() {
76	if (!empty($_FILES['files'])) {
77	module_load_include('inc', 'transliteration');
78
79	// Figure out language, which is available for node form submits.
80	$langcode = NULL;
81	if (!empty($_POST['language'])) {
82	$languages = language_list();
83	if (isset($languages[$_POST['language']])) {
84	$langcode = $_POST['language'];
85	}
86	}
87	foreach ($_FILES['files']['name'] as $field => $filename) {
88	// Keep a copy of the unaltered file name.
89	$_FILES['files']['orig_name'][$field] = $filename;
90	$_FILES['files']['name'][$field] = transliteration_clean_filename($filename, $langcode);
91	}
92	}
93	}
94