Merge pull request #82 from Andrius521/master

Improve lrc parsing and some fixes
mantas-done · Aug 14, 2023 · 2610635 · 2610635
2 parents 21f1e8e + 000f2c9
commit 2610635
Show file tree

Hide file tree

Showing 3 changed files with 47 additions and 19 deletions.
diff --git a/src/Code/Converters/LrcConverter.php b/src/Code/Converters/LrcConverter.php
@@ -6,37 +6,44 @@
 
 class LrcConverter implements ConverterContract
 {
-    protected static $regexp = '/\[\s*(\d{2}:\d{2}(?:[:.]\d{1,3})?)\s*](.*)/s';
+    protected static $regexp = '/\[\s*(\d{2}:\d{2}(?:[:.]\d{1,3})?)\s*]/';
     protected static $time_offset_regexp = '/\[offset:\s*\+?(-?\d+)\s*]/s';
 
     public function canParseFileContent($file_content)
     {
-        return preg_match(self::$regexp, $file_content) === 1;
+        return preg_match(self::$regexp . 's', $file_content) === 1;
     }
 
     public function fileContentToInternalFormat($file_content)
     {
         $timestamp_offset = self::timestampsOffset($file_content);
         $lines = explode("\n", $file_content);
         $internal_format = [];
-        $i = 0;
         foreach ($lines as $line) {
-            $found = preg_match(self::$regexp, $line, $match);
+            $found = preg_match_all(self::$regexp, $line, $timestamps);
             if ($found === 0) {
                 continue;
             }
 
-            $internal_format[$i] = [
-                'start' => static::lrcTimeToInternal($match[1], $timestamp_offset),
-                'end' => null,
-                'lines' => [trim($match[2])]
-            ];
+            $text = str_replace($timestamps[0], '', $line);
 
+            foreach ($timestamps[1] as $timestamp) {
+                $internal_format[] = [
+                    'start' => static::lrcTimeToInternal($timestamp, $timestamp_offset),
+                    'end' => null,
+                    'lines' => [trim($text)]
+                ];
+            }
+        }
+
+        usort($internal_format, function ($a, $b) {
+            return $a['start'] <=> $b['start'];
+        });
+
+        for ($i = 0; $i < count($internal_format); $i++) {
             if (isset($internal_format[$i - 1]) && $internal_format[$i - 1]['end'] === null) {
                 $internal_format[$i - 1]['end'] = $internal_format[$i]['start'];
             }
-
-            $i++;
         }
 
         //TODO: Currently last line's end time is start + 1sec, but it might be calculated differently

diff --git a/src/Code/Converters/TxtConverter.php b/src/Code/Converters/TxtConverter.php
@@ -214,9 +214,9 @@ public static function fillStartAndEndTimes(array $internal_format)
         // sort times
         usort($internal_format, function ($a, $b) {
             if ($a['start'] === $b['start']) {
-                return $a['end'] > $b['end'];
+                return $a['end'] <=> $b['end'];
             }
-            return $a['start'] > $b['start'];
+            return $a['start'] <=> $b['start'];
         });
 
         // fill ends
@@ -359,12 +359,14 @@ private static function timestampsFromLine(string $line)
         if (isset($timestamps[0][1])) {
             $result['end'] = $timestamps[0][1];
         }
-        $text_before_timestamp = substr($line, 0, strpos($line, $result['start']));
-        if ($result['start'] && self::hasText($text_before_timestamp)) {
-            $result = [
-                'start' => null,
-                'end' => null,
-            ];
+        if ($result['start']) {
+            $text_before_timestamp = substr($line, 0, strpos($line, $result['start']));
+            if (self::hasText($text_before_timestamp)) {
+                $result = [
+                    'start' => null,
+                    'end' => null,
+                ];
+            }
         }
 
         return $result;

diff --git a/tests/formats/LrcTest.php b/tests/formats/LrcTest.php
@@ -80,4 +80,23 @@ public function testParseLrcWithNegativeTimeOffset()
 
         $this->assertEquals($expected, $actual);
     }
+
+    public function testParseGroupedLines()
+    {
+        $given = <<< TEXT
+[00:01.10] First
+[00:02.20][00:05.00] [grouped]
+[00:03:25] Third
+TEXT;
+        $actual = (new Subtitles())->loadFromString($given)->getInternalFormat();
+
+        $expected = (new Subtitles())
+            ->add(1.1, 2.2, 'First')
+            ->add(2.2, 3.25, '[grouped]')
+            ->add(3.25, 5, 'Third')
+            ->add(5, 6, '[grouped]')
+            ->getInternalFormat();
+
+        $this->assertEquals($expected, $actual);
+    }
 }