2022-08-26 15:44:59 +00:00
|
|
|
// Copyright 2017 Google LLC. All rights reserved.
|
2017-09-18 15:43:21 +00:00
|
|
|
//
|
|
|
|
// Use of this source code is governed by a BSD-style
|
|
|
|
// license that can be found in the LICENSE file or at
|
|
|
|
// https://developers.google.com/open-source/licenses/bsd
|
|
|
|
|
2023-10-10 23:51:11 +00:00
|
|
|
#include <packager/media/chunking/text_chunker.h>
|
2017-09-18 15:43:21 +00:00
|
|
|
|
2023-10-13 19:42:47 +00:00
|
|
|
#include <absl/log/check.h>
|
|
|
|
|
2023-10-14 16:36:01 +00:00
|
|
|
#include <packager/macros/status.h>
|
2018-03-23 01:33:01 +00:00
|
|
|
|
2017-09-18 15:43:21 +00:00
|
|
|
namespace shaka {
|
|
|
|
namespace media {
|
|
|
|
namespace {
|
|
|
|
const size_t kStreamIndex = 0;
|
2018-02-05 19:16:00 +00:00
|
|
|
} // namespace
|
2017-09-18 15:43:21 +00:00
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
TextChunker::TextChunker(double segment_duration_in_seconds)
|
|
|
|
: segment_duration_in_seconds_(segment_duration_in_seconds){};
|
2017-09-18 15:43:21 +00:00
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
Status TextChunker::Process(std::unique_ptr<StreamData> data) {
|
|
|
|
switch (data->stream_data_type) {
|
2017-09-18 15:43:21 +00:00
|
|
|
case StreamDataType::kStreamInfo:
|
2018-03-23 01:33:01 +00:00
|
|
|
return OnStreamInfo(std::move(data->stream_info));
|
2017-09-18 15:43:21 +00:00
|
|
|
case StreamDataType::kTextSample:
|
2018-03-23 01:33:01 +00:00
|
|
|
return OnTextSample(data->text_sample);
|
|
|
|
case StreamDataType::kCueEvent:
|
|
|
|
return OnCueEvent(data->cue_event);
|
2017-09-18 15:43:21 +00:00
|
|
|
default:
|
|
|
|
return Status(error::INTERNAL_ERROR,
|
|
|
|
"Invalid stream data type for this handler");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-07-05 22:18:36 +00:00
|
|
|
Status TextChunker::OnFlushRequest(size_t /*input_stream_index*/) {
|
2018-06-06 18:41:21 +00:00
|
|
|
// Keep outputting segments until all the samples leave the system. Calling
|
|
|
|
// |DispatchSegment| will remove samples over time.
|
|
|
|
while (samples_in_current_segment_.size()) {
|
|
|
|
RETURN_IF_ERROR(DispatchSegment(segment_duration_));
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
2018-02-05 17:55:58 +00:00
|
|
|
|
|
|
|
return FlushAllDownstreams();
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
Status TextChunker::OnStreamInfo(std::shared_ptr<const StreamInfo> info) {
|
2018-06-06 18:41:21 +00:00
|
|
|
time_scale_ = info->time_scale();
|
|
|
|
segment_duration_ = ScaleTime(segment_duration_in_seconds_);
|
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
return DispatchStreamInfo(kStreamIndex, std::move(info));
|
|
|
|
}
|
2017-09-18 15:43:21 +00:00
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
Status TextChunker::OnCueEvent(std::shared_ptr<const CueEvent> event) {
|
2018-06-06 18:41:21 +00:00
|
|
|
// We are going to end the current segment prematurely using the cue event's
|
|
|
|
// time as the new segment end.
|
|
|
|
|
|
|
|
// Because the cue should have been inserted into the stream such that no
|
|
|
|
// later sample could start before it does, we know that there should
|
|
|
|
// be no later samples starting before the cue event.
|
|
|
|
|
|
|
|
// Convert the event's time to be scaled to the time of each sample.
|
|
|
|
const int64_t event_time = ScaleTime(event->time_in_seconds);
|
|
|
|
|
|
|
|
// Output all full segments before the segment that the cue event interupts.
|
|
|
|
while (segment_start_ + segment_duration_ < event_time) {
|
|
|
|
RETURN_IF_ERROR(DispatchSegment(segment_duration_));
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
const int64_t shorten_duration = event_time - segment_start_;
|
2018-02-05 19:16:00 +00:00
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
RETURN_IF_ERROR(DispatchSegment(shorten_duration));
|
|
|
|
return DispatchCueEvent(kStreamIndex, std::move(event));
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
Status TextChunker::OnTextSample(std::shared_ptr<const TextSample> sample) {
|
|
|
|
// Output all segments that come before our new sample.
|
2018-06-06 18:41:21 +00:00
|
|
|
const int64_t sample_start = sample->start_time();
|
2018-06-21 18:31:09 +00:00
|
|
|
|
|
|
|
// If we have not seen a sample yet, base all segments off the first sample's
|
|
|
|
// start time.
|
|
|
|
if (segment_start_ < 0) {
|
|
|
|
// Force the first segment to start at the segment that would have started
|
|
|
|
// before the sample. This should allow segments from different streams to
|
|
|
|
// align.
|
|
|
|
segment_start_ = (sample_start / segment_duration_) * segment_duration_;
|
|
|
|
}
|
|
|
|
|
|
|
|
// We need to write all the segments that would have ended before the new
|
|
|
|
// sample started.
|
2018-06-06 18:41:21 +00:00
|
|
|
while (sample_start >= segment_start_ + segment_duration_) {
|
2018-06-21 18:31:09 +00:00
|
|
|
// |DispatchSegment| will advance |segment_start_|.
|
2018-06-06 18:41:21 +00:00
|
|
|
RETURN_IF_ERROR(DispatchSegment(segment_duration_));
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
samples_in_current_segment_.push_back(std::move(sample));
|
2018-03-23 01:33:01 +00:00
|
|
|
|
|
|
|
return Status::OK;
|
|
|
|
}
|
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
Status TextChunker::DispatchSegment(int64_t duration) {
|
|
|
|
DCHECK_GT(duration, 0) << "Segment duration should always be positive";
|
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
// Output all the samples that are part of the segment.
|
2018-06-06 18:41:21 +00:00
|
|
|
for (const auto& sample : samples_in_current_segment_) {
|
2018-03-23 01:33:01 +00:00
|
|
|
RETURN_IF_ERROR(DispatchTextSample(kStreamIndex, sample));
|
2018-01-02 21:58:47 +00:00
|
|
|
}
|
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
// Output the segment info.
|
|
|
|
std::shared_ptr<SegmentInfo> info = std::make_shared<SegmentInfo>();
|
|
|
|
info->start_timestamp = segment_start_;
|
|
|
|
info->duration = duration;
|
|
|
|
RETURN_IF_ERROR(DispatchSegmentInfo(kStreamIndex, std::move(info)));
|
|
|
|
|
|
|
|
// Move onto the next segment.
|
|
|
|
const int64_t new_segment_start = segment_start_ + duration;
|
|
|
|
segment_start_ = new_segment_start;
|
|
|
|
|
|
|
|
// Remove all samples that end before the (new) current segment started.
|
|
|
|
samples_in_current_segment_.remove_if(
|
|
|
|
[new_segment_start](const std::shared_ptr<const TextSample>& sample) {
|
|
|
|
// For the sample to even be in this list, it should have started
|
|
|
|
// before the (new) current segment.
|
|
|
|
DCHECK_LT(sample->start_time(), new_segment_start);
|
|
|
|
return sample->EndTime() <= new_segment_start;
|
|
|
|
});
|
2018-02-05 19:16:00 +00:00
|
|
|
|
2018-03-23 01:33:01 +00:00
|
|
|
return Status::OK;
|
2017-09-18 15:43:21 +00:00
|
|
|
}
|
2018-03-23 01:33:01 +00:00
|
|
|
|
2018-06-06 18:41:21 +00:00
|
|
|
int64_t TextChunker::ScaleTime(double seconds) const {
|
|
|
|
DCHECK_GT(time_scale_, 0) << "Need positive time scale to scale time.";
|
|
|
|
return static_cast<int64_t>(seconds * time_scale_);
|
2018-03-23 01:33:01 +00:00
|
|
|
}
|
2017-09-18 15:43:21 +00:00
|
|
|
} // namespace media
|
|
|
|
} // namespace shaka
|