Patch for real UTF-16 input, UTF-8 output and tree structure

diff --git a/core/main.php b/core/main.php
index ce88cb9..6aa032c 100644
--- a/core/main.php
+++ b/core/main.php
@@ -26,7 +26,9 @@ $optionsConfig = [
     'chronicle' => ['AI chronicle. Provide a directory name from the data directory.', 'gf'],
     'language' => ['Resulting language. Provide a file name from the core/generators directory (without .php extension).', 'nasc'],
     'split' => ["\t" . 'Split result by classes.', true],
-    'join' => ["\t" . 'Join split classes into one file. Provide a directory which contains the classes.txt file.', null]
+    'join' => ["\t" . 'Join split classes into one file. Provide a directory which contains the classes.txt file.', null],
+    'tree' => ["\t" . 'Split result in tree structure.', false],
+    'utf8' => ["\t" . 'Encode output in UTF-8 instead of UTF-16LE', false],
 ];

 $longopts = [];
@@ -76,14 +78,47 @@ $parser = new Parser($data);
 $generatorClass = ucfirst($options['language']) . 'Generator';
 $generator = new $generatorClass(); /** @var GeneratorInterface $generator */

+class readutf16le_filter extends php_user_filter {
+    function filter($in, $out, &$consumed, $closing) {
+        while ($bucket = stream_bucket_make_writeable($in)) {
+            # printf("filter: %s\n", to_hex($bucket->data));
+            $bucket->data = iconv('UTF-16LE', 'UTF-8',
+                strlen($bucket->data) && substr($bucket->data, 0, 2) == "\xff\xfe"
+                    ? substr($bucket->data, 2)
+                    : $bucket->data);
+            $consumed += $bucket->datalen;
+            stream_bucket_append($out, $bucket);
+        }
+        return PSFS_PASS_ON;
+    }
+}
+
+stream_filter_register('readutf16le', 'readutf16le_filter');
+
 $file = fopen($options['input'], 'r');
+
+stream_filter_append($file, 'readutf16le');
+
 $line = 0;

+function rmr($dir) {
+    $files = array_diff(scandir($dir), array('.', '..'));
+    foreach ($files as $file) {
+        $path = $dir . '/' . $file;
+        if (is_dir($path)) {
+            rmr($path);
+        } else {
+            unlink($path);
+        }
+    }
+    rmdir($dir);
+}
+
 if (!$options['split']) {
     $outputFile = $options['output'] . '.' . $options['language'];

     // workaround for NASC: write BOM
-    if ($options['language'] === 'nasc') {
+    if ($options['language'] === 'nasc' && !$options['utf8']) {
         file_put_contents($outputFile, pack('S', 0xFEFF));
     } else {
         file_put_contents($outputFile, '');
@@ -92,12 +127,14 @@ if (!$options['split']) {
     mkdir($options['output']);
 } else {
     echo 'Cleaning output directory: ' . $options['output'] . "\n\n";
-    array_map('unlink', glob($options['output'] . '/*'));
+    rmr($options['output']);
+    mkdir($options['output']);
 }

+$tree = [];
+
 while ($file && !feof($file)) {
     $string = trim(fgets($file));
-    $string = preg_replace('/[^\s\x20-\x7E]/', '', $string); // remove non-ASCII characters
     $line++;

     if (!$string) {
@@ -131,17 +168,36 @@ while ($file && !feof($file)) {
             $outputFile = $options['output'] . '.' . $options['language'];

             // workaround for NASC: convert to UTF-16LE BOM
-            if ($options['language'] === 'nasc') {
+            if ($options['language'] === 'nasc' && !$options['utf8']) {
                 file_put_contents($outputFile, iconv('UTF-8', 'UTF-16LE', $code), FILE_APPEND);
             } else {
                 file_put_contents($outputFile, $code, FILE_APPEND);
             }
         } else {
-            $outputFile = $options['output'] . '/' . $name . '.' . $options['language'];
-            file_put_contents($options['output'] . '/classes.txt', $name . '.' . $options['language'] . "\n", FILE_APPEND);
+            if ($options['tree']) {
+                if ($class->getSuper() !== null) {
+                    $tree[$class->getName()] = $class->getSuper();
+                }
+                $filename = '';
+                $current = $class->getName();
+                for (;;) {
+                    $current = $tree[$current];
+                    if (!$current) break;
+                    $filename = $current . '/' . $filename;
+                }
+                $dirname = $options['output'] . '/' . $filename;
+                if (!is_dir($dirname)) {
+                    mkdir($dirname, 0777, true);
+                }
+                $filename = $filename . $class->getName();
+            } else {
+                $filename = $name;
+            }
+            $outputFile = $options['output'] . '/' . $filename . '.' . $options['language'];
+            file_put_contents($options['output'] . '/classes.txt', $filename . '.' . $options['language'] . "\n", FILE_APPEND);

             // workaround for NASC: convert to UTF-16LE BOM
-            if ($options['language'] === 'nasc') {
+            if ($options['language'] === 'nasc' && !$options['utf8']) {
                 file_put_contents($outputFile, pack('S', 0xFEFF) . iconv('UTF-8', 'UTF-16LE', $code));
             } else {
                 file_put_contents($outputFile, $code);