Script for benchmarking mb_detect_encoding

bench_detect_encoding.php

<?php

$path = explode('/', readlink('/proc/self/exe'));
$GLOBALS['commit'] = str_replace('php-', '', array_pop($path));
$GLOBALS['iterations_per_test'] = 5000;

function bench($func, $arg, $iterations) {
  $t1 = microtime(true);
  while ($iterations-- > 0) {
    $func($arg);
  }
  return microtime(true) - $t1;
}

function benchAll($func, $args, $iterations) {
  $elapsed = 0;
  foreach ($args as $arg) {
    $elapsed += bench($func, $arg, $iterations);
  }
  return $elapsed / count($args);
}

/* Test strings */

$ascii_short  = array("a", "This", "cat", "", "APART", "``", "_x_", "@");
$ascii_medium = array(str_repeat("abcde", 20),
                      join(array_map(function($n) { return hash('md5', $n); }, range(1, 5))),
                      str_repeat(join(array_map('chr', range(33, 126))), 2));
$ascii_long   = array(str_repeat('This is a long string! ', 500),
                      join(array_map(function($n) { return hash('md5', $n); }, range(10, 310))),
                      str_repeat(join(array_map('chr', range(33, 126))), 120));

$utf8_short  = array("ひらがな", "Καλημέρα", "АБВГ", "ABC", "", "⡌⠁⠧⠑ ⠼", "∑ f(i)");
$utf8_medium = array("Οὐχὶ ταὐτὰ παρίσταταί μοι γιγνώσκειν, ὦ ἄνδρες ᾿Αθηναῖοι, ὅταν τ᾿ εἰς τὰ πράγματα ἀποβλέψω καὶ ὅταν πρὸς τοὺς",
                     "Зарегистрируйтесь сейчас на Десятую Международную Конференцию по
Unicode, которая состоится 10-12 марта 1997 года в Майнце в Германии.",
                     "ABCDEFGHIJKLMNOPQRSTUVWXYZ /0123456789 abcdefghijklmnopqrstuvwxyz £©µÀÆÖÞßéöÿАБВГДабвгд∀∂∈ℝ∧∪≡∞ ↑↗↨↻⇣",
                     "⡍⠔⠙⠖ ⡊ ⠙⠕⠝⠰⠞ ⠍⠑⠁⠝ ⠞⠕ ⠎⠁⠹ ⠹⠁⠞ ⡊ ⠅⠝⠪⠂ ⠕⠋ ⠍⠹⠪⠝ ⠅⠝⠪⠇⠫⠛⠑⠂ ⠱⠁⠞ ⠹⠻⠑ ⠊⠎ ⠏⠜⠞⠊⠊⠥⠇⠜⠇⠹ ⠙⠑⠁⠙ ⠁⠃⠳⠞",
                     "∮ E⋅da = Q,  n → ∞, ∑ f(i) = ∏ g(i), ∀x∈ℝ: ⌈x⌉ = −⌊−x⌋, α ∧ ¬β = ¬(¬α ∨ β), ℕ ⊆ ℕ₀ ⊂ ℤ ⊂ ℚ ⊂ ℝ ⊂ ℂ, ⊥ < a ≠ b ≡ c ≤ d ≪ ⊤ ⇒ (A ⇔ B)");
$utf8_long   = array(str_repeat("Σὲ γνωρίζω ἀπὸ τὴν κόψη τοῦ σπαθιοῦ τὴν τρομερή, σὲ γνωρίζω ἀπὸ τὴν ὄψη ποὺ μὲ βία μετράει τὴ γῆ.", 100),
                     join(array_map(function($n) { return mb_chr($n, 'UTF-8'); }, range(1, 10000))),
                     str_repeat("გთხოვთ ახლავე გაიაროთ რეგისტრაცია Unicode-ის მეათე საერთაშორისო კონფერენციაზე დასასწრებად, რომელიც გაიმართება 10-12 მარტს,", 80));

$utf16_short  = array_map(function($str) { return mb_convert_encoding($str, 'UTF-16LE', 'UTF-8'); }, $utf8_short);
$utf16_medium = array_map(function($str) { return mb_convert_encoding($str, 'UTF-16LE', 'UTF-8'); }, $utf8_medium);
$utf16_long   = array_map(function($str) { return mb_convert_encoding($str, 'UTF-16LE', 'UTF-8'); }, $utf8_long);

$jp_short  = array("こんにちは", "テストのデータ", "漢字も", "１２３", ""); /* UTF-8 */
$jp_medium = array("むかし、むかし、あるところにおじいさんとおばあさんがいました。おじいさんが山へ木をきりにいけば、おばあさんは川へせんたくにでかけます。「おじいさん、はようもどってきなされ。」「おばあさんもきをつけてな。」まい日やさしくいいあってでかけます。",
                   "静岡県警によると、現場は静岡市葵区柚木の線路内で、人が立ち入り、走行中の新幹線がはねたとみられるという。県警はJR東海からの通報を受けて現場検証を実施、詳しい状況を調べている。",
                   "８千人はっせんにん……、毎年まいとし８千人はっせんにんが全国ぜんこくに８１はちじゅういちある大学だいがく医い学がく部ぶを卒そつ業ぎょうしてゆく。君達きみたちはその８千人はっせんにんのトップの８０はちじゅう人にん");
$jp_long   = array_map(function($str) { return str_repeat($str, 100); }, $jp_medium);

$jis_short  = array_map(function($str) { return mb_convert_encoding($str, 'JIS', 'UTF-8'); }, $jp_short);
$jis_medium = array_map(function($str) { return mb_convert_encoding($str, 'JIS', 'UTF-8'); }, $jp_medium);
$jis_long   = array_map(function($str) { return mb_convert_encoding($str, 'JIS', 'UTF-8'); }, $jp_long);

$sjis_short  = array_map(function($str) { return mb_convert_encoding($str, 'SJIS', 'UTF-8'); }, $jp_short);
$sjis_medium = array_map(function($str) { return mb_convert_encoding($str, 'SJIS', 'UTF-8'); }, $jp_medium);
$sjis_long   = array_map(function($str) { return mb_convert_encoding($str, 'SJIS', 'UTF-8'); }, $jp_long);

function benchMBDetectEncoding($strings, $encodings, $desc) {
  $elapsed = benchAll(function($str) use(&$encodings) {
    mb_detect_encoding($str, $encodings, true);
  }, $strings, $GLOBALS['iterations_per_test']);
  echo "{$GLOBALS['commit']}|mb_detect_encoding|$desc - strict|$elapsed|{$GLOBALS['iterations_per_test']}\n";

/*
  $elapsed = benchAll(function($str) use(&$encodings) {
    mb_detect_encoding($str, $encodings, false);
  }, $strings, $GLOBALS['iterations_per_test']);
  echo "{$GLOBALS['commit']}|mb_detect_encoding|$desc - non-strict|$elapsed|{$GLOBALS['iterations_per_test']}\n";
*/
}

/*
benchMBDetectEncoding($utf8_medium, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'UTF-8, medium');
benchMBDetectEncoding($utf8_long, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'UTF-8, long');
benchMBDetectEncoding($utf16_medium, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'UTF-16LE, medium');
benchMBDetectEncoding($utf16_long, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'UTF-16LE, long');
*/

benchMBDetectEncoding($jis_long, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'JIS, long');

/*
benchMBDetectEncoding($sjis_long, array('UTF-8', 'UTF-16LE', 'UTF-16BE', 'JIS', 'SJIS'), 'SJIS, long');
*/