Handle segmented text same as audio/video.

Now the same pipeline for handling the audio/videos streams will handle
the segmented text streams too.  This doesn't apply to the text output,
only to the MP4 variants.  This also fixes a bug where we added the
X-TIMESTAMP-MAP tag even when there wasn't TS streams; this doesn't
otherwise change the behavior around that tag.

Change-Id: I03f7cea56efa42e96311c00841330629a14aa053

packager/app/test/testdata/hls-only-dash-only-captions/output.m3u8

@@ -3,9 +3,9 @@
 
 #EXT-X-INDEPENDENT-SEGMENTS
 
-#EXT-X-MEDIA:TYPE=AUDIO,URI="stream_2.m3u8",GROUP-ID="default-audio-group",NAME="stream_2",AUTOSELECT=YES,CHANNELS="2"
+#EXT-X-MEDIA:TYPE=AUDIO,URI="stream_1.m3u8",GROUP-ID="default-audio-group",NAME="stream_1",AUTOSELECT=YES,CHANNELS="2"
 
-#EXT-X-MEDIA:TYPE=SUBTITLES,URI="stream_1.m3u8",GROUP-ID="default-text-group",NAME="stream_1",AUTOSELECT=YES
+#EXT-X-MEDIA:TYPE=SUBTITLES,URI="stream_0.m3u8",GROUP-ID="default-text-group",NAME="stream_0",AUTOSELECT=YES
 
 #EXT-X-STREAM-INF:BANDWIDTH=1108115,AVERAGE-BANDWIDTH=1006069,CODECS="avc1.64001e,mp4a.40.2",RESOLUTION=640x360,FRAME-RATE=29.970,AUDIO="default-audio-group",SUBTITLES="default-text-group"
-stream_3.m3u8
+stream_2.m3u8

packager/app/test/testdata/hls-only-dash-only-captions/stream_0.m3u8

@@ -0,0 +1,16 @@
+#EXTM3U
+#EXT-X-VERSION:6
+## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
+#EXT-X-TARGETDURATION:2
+#EXT-X-PLAYLIST-TYPE:VOD
+#EXTINF:1.000,
+bear-english-text-1.vtt
+#EXTINF:1.000,
+bear-english-text-2.vtt
+#EXTINF:1.000,
+bear-english-text-3.vtt
+#EXTINF:1.000,
+bear-english-text-4.vtt
+#EXTINF:1.000,
+bear-english-text-5.vtt
+#EXT-X-ENDLIST

packager/app/test/testdata/hls-only-dash-only-captions/stream_1.m3u8

@@ -3,14 +3,11 @@
 ## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
 #EXT-X-TARGETDURATION:2
 #EXT-X-PLAYLIST-TYPE:VOD
-#EXTINF:1.000,
+#EXT-X-MAP:URI="bear-640x360-audio-init.mp4"
-bear-english-text-1.vtt
+#EXTINF:1.022,
-#EXTINF:1.000,
+bear-640x360-audio-1.m4s
-bear-english-text-2.vtt
+#EXTINF:0.998,
-#EXTINF:1.000,
+bear-640x360-audio-2.m4s
-bear-english-text-3.vtt
+#EXTINF:0.720,
-#EXTINF:1.000,
+bear-640x360-audio-3.m4s
-bear-english-text-4.vtt
-#EXTINF:1.000,
-bear-english-text-5.vtt
 #EXT-X-ENDLIST

packager/app/test/testdata/hls-only-dash-only-captions/stream_2.m3u8

@@ -3,11 +3,11 @@
 ## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
 #EXT-X-TARGETDURATION:2
 #EXT-X-PLAYLIST-TYPE:VOD
-#EXT-X-MAP:URI="bear-640x360-audio-init.mp4"
+#EXT-X-MAP:URI="bear-640x360-video-init.mp4"
-#EXTINF:1.022,
+#EXTINF:1.001,
-bear-640x360-audio-1.m4s
+bear-640x360-video-1.m4s
-#EXTINF:0.998,
+#EXTINF:1.001,
-bear-640x360-audio-2.m4s
+bear-640x360-video-2.m4s
-#EXTINF:0.720,
+#EXTINF:0.734,
-bear-640x360-audio-3.m4s
+bear-640x360-video-3.m4s
 #EXT-X-ENDLIST

packager/app/test/testdata/hls-only-dash-only-captions/stream_3.m3u8

@@ -1,13 +0,0 @@
-#EXTM3U
-#EXT-X-VERSION:6
-## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
-#EXT-X-TARGETDURATION:2
-#EXT-X-PLAYLIST-TYPE:VOD
-#EXT-X-MAP:URI="bear-640x360-video-init.mp4"
-#EXTINF:1.001,
-bear-640x360-video-1.m4s
-#EXTINF:1.001,
-bear-640x360-video-2.m4s
-#EXTINF:0.734,
-bear-640x360-video-3.m4s
-#EXT-X-ENDLIST

packager/app/test/testdata/segmented-webvtt-with-language-override/bear-english-text-1.vtt

@@ -1,5 +1,4 @@
 WEBVTT
-X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:9000
 
 STYLE
 ::cue { color:lime }

packager/app/test/testdata/segmented-webvtt-with-language-override/bear-english-text-2.vtt

@@ -1,5 +1,4 @@
 WEBVTT
-X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:9000
 
 STYLE
 ::cue { color:lime }

packager/app/test/testdata/segmented-webvtt-with-language-override/bear-english-text-3.vtt

@@ -1,5 +1,4 @@
 WEBVTT
-X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:9000
 
 STYLE
 ::cue { color:lime }

packager/app/test/testdata/segmented-webvtt-with-language-override/bear-english-text-4.vtt

@@ -1,5 +1,4 @@
 WEBVTT
-X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:9000
 
 STYLE
 ::cue { color:lime }

packager/app/test/testdata/segmented-webvtt-with-language-override/bear-english-text-5.vtt

@@ -1,5 +1,4 @@
 WEBVTT
-X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:9000
 
 STYLE
 ::cue { color:lime }

packager/app/test/testdata/segmented-webvtt-with-language-override/output.m3u8

@@ -3,4 +3,4 @@
 
 #EXT-X-INDEPENDENT-SEGMENTS
 
-#EXT-X-MEDIA:TYPE=SUBTITLES,URI="stream_1.m3u8",GROUP-ID="default-text-group",LANGUAGE="pt",NAME="stream_1",AUTOSELECT=YES
+#EXT-X-MEDIA:TYPE=SUBTITLES,URI="stream_0.m3u8",GROUP-ID="default-text-group",LANGUAGE="pt",NAME="stream_0",AUTOSELECT=YES

packager/app/test/testdata/vtt-text-to-mp4-with-ad-cues/output.m3u8

@@ -3,9 +3,9 @@
 
 #EXT-X-INDEPENDENT-SEGMENTS
 
-#EXT-X-MEDIA:TYPE=AUDIO,URI="bear-640x360-audio.m3u8",GROUP-ID="default-audio-group",NAME="stream_1",AUTOSELECT=YES,CHANNELS="2"
+#EXT-X-MEDIA:TYPE=AUDIO,URI="bear-640x360-audio.m3u8",GROUP-ID="default-audio-group",NAME="stream_0",AUTOSELECT=YES,CHANNELS="2"
 
-#EXT-X-MEDIA:TYPE=SUBTITLES,URI="bear-english-text.m3u8",GROUP-ID="default-text-group",NAME="stream_0",AUTOSELECT=YES
+#EXT-X-MEDIA:TYPE=SUBTITLES,URI="bear-english-text.m3u8",GROUP-ID="default-text-group",NAME="stream_2",AUTOSELECT=YES
 
 #EXT-X-STREAM-INF:BANDWIDTH=1108115,AVERAGE-BANDWIDTH=1006069,CODECS="avc1.64001e,mp4a.40.2",RESOLUTION=640x360,FRAME-RATE=29.970,AUDIO="default-audio-group",SUBTITLES="default-text-group"
 bear-640x360-video.m3u8

packager/media/base/media_handler.cc

@@ -61,10 +61,10 @@ Status MediaHandler::Initialize() {
 }
 
 Status MediaHandler::Chain(
-    std::initializer_list<std::shared_ptr<MediaHandler>> list) {
+    const std::vector<std::shared_ptr<MediaHandler>>& list) {
   std::shared_ptr<MediaHandler> previous;
 
-  for (auto& next : list) {
+  for (const auto& next : list) {
     // Skip null entries.
     if (!next) {
       continue;

packager/media/base/media_handler.h

@@ -172,8 +172,7 @@ class MediaHandler {
   /// Validate if the handler is connected to its upstream handler.
   bool IsConnected() { return num_input_streams_ > 0; }
 
-  static Status Chain(
+  static Status Chain(const std::vector<std::shared_ptr<MediaHandler>>& list);
-      std::initializer_list<std::shared_ptr<MediaHandler>> list);
 
  protected:
   /// Internal implementation of initialize. Note that it should only initialize

packager/packager.cc

@@ -478,11 +478,11 @@ std::shared_ptr<MediaHandler> CreateEncryptionHandler(
   return std::make_shared<EncryptionHandler>(encryption_params, key_source);
 }
 
-std::unique_ptr<TextChunker> CreateTextChunker(
+std::unique_ptr<MediaHandler> CreateTextChunker(
     const ChunkingParams& chunking_params) {
   const float segment_length_in_seconds =
       chunking_params.segment_duration_in_seconds;
-  return std::unique_ptr<TextChunker>(
+  return std::unique_ptr<MediaHandler>(
       new TextChunker(segment_length_in_seconds));
 }
 
@@ -527,39 +527,6 @@ Status CreateHlsTextJob(const StreamDescriptor& stream,
                               std::move(chunker), std::move(output)});
 }
 
-Status CreateWebVttToMp4TextJob(const StreamDescriptor& stream,
-                                const PackagingParams& packaging_params,
-                                std::unique_ptr<MuxerListener> muxer_listener,
-                                SyncPointQueue* sync_points,
-                                MuxerFactory* muxer_factory,
-                                std::shared_ptr<OriginHandler>* root) {
-  std::shared_ptr<Demuxer> demuxer;
-  RETURN_IF_ERROR(CreateDemuxer(stream, packaging_params, &demuxer));
-  if (!stream.language.empty())
-    demuxer->SetLanguageOverride(stream.stream_selector, stream.language);
-  auto padder = std::make_shared<TextPadder>(kDefaultTextZeroBiasMs);
-  RETURN_IF_ERROR(demuxer->SetHandler(stream.stream_selector, padder));
-
-  auto text_to_mp4 = std::make_shared<WebVttToMp4Handler>();
-  auto muxer = muxer_factory->CreateMuxer(GetOutputFormat(stream), stream);
-  muxer->SetMuxerListener(std::move(muxer_listener));
-
-  // Optional Cue Alignment Handler
-  std::shared_ptr<MediaHandler> cue_aligner;
-  if (sync_points) {
-    cue_aligner = std::make_shared<CueAlignmentHandler>(sync_points);
-  }
-
-  std::shared_ptr<MediaHandler> chunker =
-      CreateTextChunker(packaging_params.chunking_params);
-
-  *root = demuxer;
-
-  return MediaHandler::Chain({std::move(padder), std::move(cue_aligner),
-                              std::move(chunker), std::move(text_to_mp4),
-                              std::move(muxer)});
-}
-
 Status CreateTextJobs(
     const std::vector<std::reference_wrapper<const StreamDescriptor>>& streams,
     const PackagingParams& packaging_params,
@@ -578,7 +545,6 @@ Status CreateTextJobs(
     //    MP4 WEBVTT  --> MP4 WEBVTT  [ unsupported ]
     //    MP4 WEBVTT  --> TEXT WEBVTT [ unsupported ]
     const auto input_container = DetermineContainerFromFileName(stream.input);
-    const auto output_container = GetOutputFormat(stream);
 
     if (input_container != CONTAINER_WEBVTT &&
         input_container != CONTAINER_TTML) {
@@ -586,27 +552,8 @@ Status CreateTextJobs(
                     "Text output format is not support for " + stream.input);
     }
 
-    if (output_container == CONTAINER_MOV) {
-      if (input_container == CONTAINER_TTML) {
-        return Status(error::INVALID_ARGUMENT,
-                      "TTML in MP4 is not supported yet. Please follow "
-                      "https://github.com/google/shaka-packager/issues/87 for "
-                      "the updates.");
-      }
-
-      std::unique_ptr<MuxerListener> muxer_listener =
-          muxer_listener_factory->CreateListener(ToMuxerListenerData(stream));
-
-      std::shared_ptr<OriginHandler> root;
-      RETURN_IF_ERROR(CreateWebVttToMp4TextJob(
-          stream, packaging_params, std::move(muxer_listener), sync_points,
-          muxer_factory, &root));
-
-      job_manager->Add("MP4 text job", std::move(root));
-    } else {
     std::unique_ptr<MuxerListener> hls_listener =
-          muxer_listener_factory->CreateHlsListener(
+        muxer_listener_factory->CreateHlsListener(ToMuxerListenerData(stream));
-              ToMuxerListenerData(stream));
 
     // Check input to ensure that output is possible.
     if (hls_listener && !stream.dash_only) {
@@ -621,8 +568,7 @@ Status CreateTextJobs(
       }
     }
 
-      if (mpd_notifier && !stream.segment_template.empty() &&
+    if (mpd_notifier && !stream.segment_template.empty() && !stream.hls_only) {
-          !stream.hls_only) {
       return Status(error::INVALID_ARGUMENT,
                     "Cannot create text output for MPD with segment output.");
     }
@@ -668,7 +614,6 @@ Status CreateTextJobs(
       }
     }
   }
-  }
 
   return Status::OK;
 }
@@ -722,6 +667,8 @@ Status CreateAudioVideoJobs(
     const bool new_input_file = stream.input != previous_input;
     const bool new_stream =
         new_input_file || previous_selector != stream.stream_selector;
+    // TODO(modmaker): Use a better detector of text streams.
+    const bool is_text = stream.stream_selector == "text";
     previous_input = stream.input;
     previous_selector = stream.stream_selector;
 
@@ -739,27 +686,35 @@ Status CreateAudioVideoJobs(
         demuxer->SetLanguageOverride(stream.stream_selector, stream.language);
       }
 
-      replicator = std::make_shared<Replicator>();
+      std::vector<std::shared_ptr<MediaHandler>> handlers;
-      auto chunker =
+      if (is_text) {
-          std::make_shared<ChunkingHandler>(packaging_params.chunking_params);
+        handlers.emplace_back(
-      auto encryptor = CreateEncryptionHandler(packaging_params, stream,
+            std::make_shared<TextPadder>(kDefaultTextZeroBiasMs));
-                                               encryption_key_source);
-
-      // TODO(vaage) : Create a nicer way to connect handlers to demuxers.
-      if (sync_points) {
-        RETURN_IF_ERROR(
-            MediaHandler::Chain({cue_aligner, chunker, encryptor, replicator}));
-        RETURN_IF_ERROR(
-            demuxer->SetHandler(stream.stream_selector, cue_aligner));
-      } else {
-        RETURN_IF_ERROR(MediaHandler::Chain({chunker, encryptor, replicator}));
-        RETURN_IF_ERROR(demuxer->SetHandler(stream.stream_selector, chunker));
       }
+      if (sync_points) {
+        handlers.emplace_back(cue_aligner);
+      }
+      if (is_text) {
+        handlers.emplace_back(
+            CreateTextChunker(packaging_params.chunking_params));
+      } else {
+        handlers.emplace_back(std::make_shared<ChunkingHandler>(
+            packaging_params.chunking_params));
+      }
+      handlers.emplace_back(CreateEncryptionHandler(packaging_params, stream,
+                                                    encryption_key_source));
+
+      replicator = std::make_shared<Replicator>();
+      handlers.emplace_back(replicator);
+
+      RETURN_IF_ERROR(MediaHandler::Chain(handlers));
+      RETURN_IF_ERROR(demuxer->SetHandler(stream.stream_selector, handlers[0]));
     }
 
     // Create the muxer (output) for this track.
+    const auto output_format = GetOutputFormat(stream);
     std::shared_ptr<Muxer> muxer =
-        muxer_factory->CreateMuxer(GetOutputFormat(stream), stream);
+        muxer_factory->CreateMuxer(output_format, stream);
     if (!muxer) {
       return Status(error::INVALID_ARGUMENT, "Failed to create muxer for " +
                                                  stream.input + ":" +
@@ -776,7 +731,15 @@ Status CreateAudioVideoJobs(
             ? std::make_shared<TrickPlayHandler>(stream.trick_play_factor)
             : nullptr;
 
-    RETURN_IF_ERROR(MediaHandler::Chain({replicator, trick_play, muxer}));
+    // TODO(modmaker): Move to MOV muxer?
+    const auto input_container = DetermineContainerFromFileName(stream.input);
+    auto text_to_mp4 =
+        input_container == CONTAINER_WEBVTT && output_format == CONTAINER_MOV
+            ? std::make_shared<WebVttToMp4Handler>()
+            : nullptr;
+
+    RETURN_IF_ERROR(
+        MediaHandler::Chain({replicator, trick_play, text_to_mp4, muxer}));
   }
 
   return Status::OK;
@@ -806,12 +769,13 @@ Status CreateAllJobs(const std::vector<StreamDescriptor>& stream_descriptors,
     // TODO: Find a better way to determine what stream type a stream
     // descriptor is as |stream_selector| may use an index. This would
     // also allow us to use a simpler audio pipeline.
-    if (stream.stream_selector == "text") {
+    const auto output_format = GetOutputFormat(stream);
+    if (stream.stream_selector == "text" && output_format != CONTAINER_MOV) {
       text_streams.push_back(stream);
     } else {
       audio_video_streams.push_back(stream);
 
-      switch (GetOutputFormat(stream)) {
+      switch (output_format) {
         case CONTAINER_MPEG2TS:
         case CONTAINER_AAC:
         case CONTAINER_MP3: