dsc-uob · abdelaziz-mahdy · Dec 23, 2024 · Dec 23, 2024 · Dec 23, 2024
diff --git a/lib/src/utils/regexes.dart b/lib/src/utils/regexes.dart
@@ -105,7 +105,7 @@ class SrtRegex extends SubtitleRegexObject {
 /// dart code.
 class TtmlRegex extends SubtitleRegexObject {
   static const String _regex =
-      r'<p[\w\d:="\s]+?begin="(\d{1,}:)?(\d{1,}:)?(\d{1,}).(\d{1,})s?"[\w\d:="\s]+end="(\d{1,}:)?(\d{1,}:)?(\d{1,}).(\d{1,})s?"[\w\d:="\s]*>(\D+)<\/p>';
+      r'<p[\w\d:="\s]+?begin="(\d{1,}):?(\d{1,}):?(\d{1,}).(\d{1,})s?"[\w\d:="\s]+end="(\d{1,}):?(\d{1,}):?(\d{1,}).(\d{1,})s?"[\w\d:="\s]*>(\D+)<\/p>';
   const TtmlRegex()
       : super(
           pattern: _regex,

diff --git a/lib/src/utils/subtitle_parser.dart b/lib/src/utils/subtitle_parser.dart
@@ -22,8 +22,10 @@ abstract class ISubtitleParser {
   /// Normalize the text data of subtitle, remove unnecessary characters.
   String normalize(String txt) {
     return txt
-        .replaceAll(RegExp(r'<\/?[\w.]+\/?>|\n| {2,}'), ' ')
+        .replaceAll(RegExp(r'<\/?[\w.]+\/?>| {2,}'), ' ')
         .replaceAll(RegExp(r' {2,}'), ' ')
+        // Remove multiple new lines
+        .replaceAll(RegExp(r'\n{2,}'), '\n')
         .trim();
   }
 }
@@ -56,8 +58,9 @@ class SubtitleParser extends ISubtitleParser {
     final pattern = regexObject.pattern;
 
     var regExp = RegExp(pattern);
-    var matches = regExp
-        .allMatches(object.data.replaceAll('\r', '').replaceAll('\r\n', '\n'));
+    var cleanedData = object.data.replaceAll('\r', '').replaceAll('\r\n', '\n');
+
+    var matches = regExp.allMatches(cleanedData);
 
     return _decodeSubtitleFormat(
       matches,

diff --git a/test/dfxp_subtitle b/test/dfxp_subtitle
@@ -0,0 +1,13 @@
+<details> <summary>Sample DFXP/TTML content</summary>
+
+
+<?xml version="1.0" encoding="UTF-8"?>
+<tt xmlns="http://www.w3.org/ns/ttml">
+  <body>
+    <div>
+      <p begin="00:00:00.500" end="00:00:02.000">Hello from DFXP</p>
+      <p begin="00:00:02.000" end="00:00:04.000">Still the same TTML style</p>
+    </div>
+  </body>
+</tt>
+</details>
diff --git a/test/dfxp_subtitle_test.dart b/test/dfxp_subtitle_test.dart
@@ -0,0 +1,38 @@
+import 'dart:io';
+import 'package:test/test.dart';
+import 'package:subtitle/subtitle.dart';
+
+void main() {
+  group('DFXP Subtitle Parsing', () {
+    test('Test parsing of DFXP format from file', () async {
+      // Read the DFXP subtitle content
+      final subtitleFile = File('test/dfxp_subtitle');
+      final data = await subtitleFile.readAsString();
+
+      // Parse the subtitle string with the correct type
+      var controller = SubtitleController(
+        provider: SubtitleProvider.fromString(
+          data: data,
+          type: SubtitleType.dfxp,
+        ),
+      );
+      await controller.initial();
+      final subtitles = controller.subtitles;
+
+      // Basic checks
+      expect(subtitles.length, 2);
+
+      // First
+      expect(subtitles[0].index, 1);
+      expect(subtitles[0].start.toString(), '0:00:00.500000');
+      expect(subtitles[0].end.toString(), '0:00:02.000000');
+      expect(subtitles[0].data, 'Hello from DFXP');
+
+      // Second
+      expect(subtitles[1].index, 2);
+      expect(subtitles[1].start.toString(), '0:00:02.000000');
+      expect(subtitles[1].end.toString(), '0:00:04.000000');
+      expect(subtitles[1].data, 'Still the same TTML style');
+    });
+  });
+}
diff --git a/test/srt_subtitle b/test/srt_subtitle
@@ -0,0 +1,11 @@
+1
+00:00:00,000 --> 00:00:01,000
+Hello world!
+
+2
+00:00:01,001 --> 00:00:03,500
+This is a test for SRT subtitle format.
+
+3
+00:00:03,600 --> 00:00:05,000
+Subtitle parsing is fun!
diff --git a/test/srt_subtitle_test.dart b/test/srt_subtitle_test.dart
@@ -0,0 +1,44 @@
+import 'dart:io';
+import 'package:test/test.dart';
+import 'package:subtitle/subtitle.dart';
+
+void main() {
+  group('SRT Subtitle Parsing', () {
+    test('Test parsing of SRT format from file', () async {
+      // Read the SRT subtitle content from the file
+      final subtitleFile = File('test/srt_subtitle');
+      final data = await subtitleFile.readAsString();
+
+      // Parse the subtitle string with the correct type
+      var controller = SubtitleController(
+        provider: SubtitleProvider.fromString(
+          data: data,
+          type: SubtitleType.srt,
+        ),
+      );
+      await controller.initial();
+      final subtitles = controller.subtitles;
+
+      // Basic checks
+      expect(subtitles.length, 3);
+
+      // First subtitle line
+      expect(subtitles[0].index, 1);
+      expect(subtitles[0].start.toString(), '0:00:00.000000');
+      expect(subtitles[0].end.toString(), '0:00:01.000000');
+      expect(subtitles[0].data, 'Hello world!');
+
+      // Second subtitle line
+      expect(subtitles[1].index, 2);
+      expect(subtitles[1].start.toString(), '0:00:01.001000');
+      expect(subtitles[1].end.toString(), '0:00:03.500000');
+      expect(subtitles[1].data, 'This is a test for SRT subtitle format.');
+
+      // Third subtitle line
+      expect(subtitles[2].index, 3);
+      expect(subtitles[2].start.toString(), '0:00:03.600000');
+      expect(subtitles[2].end.toString(), '0:00:05.000000');
+      expect(subtitles[2].data, 'Subtitle parsing is fun!');
+    });
+  });
+}
diff --git a/test/string_subtitle b/test/string_subtitle
@@ -1,6 +1,7 @@
 1
 00:00:02,250 --> 00:00:03,509
-Why am I here?
+Why am I
+here?
 
 2
 00:00:03,960 --> 00:00:05,539

diff --git a/test/string_subtitle_test.dart b/test/string_subtitle_test.dart
@@ -24,7 +24,7 @@ void main() {
       expect(subtitles[0].index, 1);
       expect(subtitles[0].start.toString(), '0:00:02.250000');
       expect(subtitles[0].end.toString(), '0:00:03.509000');
-      expect(subtitles[0].data, 'Why am I here?');
+      expect(subtitles[0].data, 'Why am I\nhere?');
 
       expect(subtitles[9].index, 10);
       expect(subtitles[9].start.toString(), '0:00:16.729000');

diff --git a/test/ttml_subtitle b/test/ttml_subtitle
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<tt xmlns="http://www.w3.org/ns/ttml">
+  <body>
+    <div>
+      <p begin="00:00:01.000" end="00:00:03.000">Hello from TTML</p>
+      <p begin="00:00:03.000" end="00:00:04.500">This is a TTML subtitle</p>
+      <p begin="00:00:05.000" end="00:00:07.000">Goodbye from TTML</p>
+    </div>
+  </body>
+</tt>
diff --git a/test/ttml_subtitle_test.dart b/test/ttml_subtitle_test.dart
@@ -0,0 +1,44 @@
+import 'dart:io';
+import 'package:test/test.dart';
+import 'package:subtitle/subtitle.dart';
+
+void main() {
+  group('TTML Subtitle Parsing', () {
+    test('Test parsing of TTML format from file', () async {
+      // Read the TTML subtitle content from the file
+      final subtitleFile = File('test/ttml_subtitle');
+      final data = await subtitleFile.readAsString();
+
+      // Parse the subtitle string with the correct type (TTML)
+      var controller = SubtitleController(
+        provider: SubtitleProvider.fromString(
+          data: data,
+          type: SubtitleType.ttml,
+        ),
+      );
+      await controller.initial();
+      final subtitles = controller.subtitles;
+
+      // Basic checks
+      expect(subtitles.length, 3);
+
+      // First TTML subtitle
+      expect(subtitles[0].index, 1);
+      expect(subtitles[0].start.toString(), '0:00:01.000000');
+      expect(subtitles[0].end.toString(), '0:00:03.000000');
+      expect(subtitles[0].data, 'Hello from TTML');
+
+      // Second TTML subtitle
+      expect(subtitles[1].index, 2);
+      expect(subtitles[1].start.toString(), '0:00:03.000000');
+      expect(subtitles[1].end.toString(), '0:00:04.500000');
+      expect(subtitles[1].data, 'This is a TTML subtitle');
+
+      // Third TTML subtitle
+      expect(subtitles[2].index, 3);
+      expect(subtitles[2].start.toString(), '0:00:05.000000');
+      expect(subtitles[2].end.toString(), '0:00:07.000000');
+      expect(subtitles[2].data, 'Goodbye from TTML');
+    });
+  });
+}
diff --git a/test/vtt_subtitle b/test/vtt_subtitle
@@ -0,0 +1,14 @@
+WEBVTT
+
+1
+00:00:00.000 --> 00:00:02.000
+Hello from VTT
+
+2
+00:00:02.000 --> 00:00:03.500
+This is a test 
+for VTT format
+
+3
+00:00:04.000 --> 00:00:05.000
+Goodbye!
diff --git a/test/vtt_subtitle_test.dart b/test/vtt_subtitle_test.dart
@@ -0,0 +1,44 @@
+import 'dart:io';
+import 'package:test/test.dart';
+import 'package:subtitle/subtitle.dart';
+
+void main() {
+  group('VTT Subtitle Parsing', () {
+    test('Test parsing of VTT format from file', () async {
+      // Read the VTT subtitle content from the file
+      final subtitleFile = File('test/vtt_subtitle');
+      final data = await subtitleFile.readAsString();
+
+      // Parse the subtitle string with the correct type
+      var controller = SubtitleController(
+        provider: SubtitleProvider.fromString(
+          data: data,
+          type: SubtitleType.vtt,
+        ),
+      );
+      await controller.initial();
+      final subtitles = controller.subtitles;
+
+      // Basic checks
+      expect(subtitles.length, 3);
+
+      // First
+      expect(subtitles[0].index, 1);
+      expect(subtitles[0].start.toString(), '0:00:00.000000');
+      expect(subtitles[0].end.toString(), '0:00:02.000000');
+      expect(subtitles[0].data, 'Hello from VTT');
+
+      // Second
+      expect(subtitles[1].index, 2);
+      expect(subtitles[1].start.toString(), '0:00:02.000000');
+      expect(subtitles[1].end.toString(), '0:00:03.500000');
+      expect(subtitles[1].data, 'This is a test \nfor VTT format');
+
+      // Third
+      expect(subtitles[2].index, 3);
+      expect(subtitles[2].start.toString(), '0:00:04.000000');
+      expect(subtitles[2].end.toString(), '0:00:05.000000');
+      expect(subtitles[2].data, 'Goodbye!');
+    });
+  });
+}