feat: enable processing of large video files over 2GB

Update audio extraction to use file-based streaming
instead of in-memory buffers to support processing
video files larger than 2GB.

**Changes:**
- Modified `IAudioExtractor.ExtractMp3FromMp4Async()`
  to accept an output audio stream parameter instead
  of creating a MemoryStream internally
- Updated `AudioExtractor` to write directly to the
  provided audio stream, eliminating the need to load
  entire video into memory
- Enhanced `FFMpegService` to detect FileStream inputs
  and use direct file-to-file processing when possible,
  falling back to pipe-based streaming for other stream types
- Refactored `DefaultCommand` to create output directory
  and audio file stream before extraction, enabling direct
  file-based audio extraction
- Updated all unit and integration tests to accommodate the
  new audio stream parameter

This change prevents `OutOfMemoryException` errors when processing
large video files by streaming data directly to disk rather than
buffering in memory. Additionally, it removes the possibility of
overflowing the 2GB size limit that `MemoryStream` has, which
would result in `IOException` errors when trying to write or
copy to it.
This commit is contained in:
Stevan Freeborn
2025-10-10 16:23:02 -05:00
parent 6dcffa6b75
commit e611522c7b
6 changed files with 57 additions and 38 deletions
@@ -69,13 +69,28 @@ internal sealed class DefaultCommand(
public override async Task<int> ExecuteAsync(CommandContext context, Settings settings) public override async Task<int> ExecuteAsync(CommandContext context, Settings settings)
{ {
_console.MarkupLine($"[blue]Processing stream:[/] {settings.Stream.EscapeMarkup()}"); _console.MarkupLine($"[blue]Processing stream:[/] {settings.Stream.EscapeMarkup()}");
var videoStream = _fileSystem.File.OpenRead(settings.Stream); var videoStream = (FileStream)_fileSystem.File.OpenRead(settings.Stream);
Stream? audioStream = null; var now = _timeProvider.GetUtcNow();
var inputFileName = _fileSystem.Path.GetFileNameWithoutExtension(settings.Stream);
var baseDirectory = ValidateAndGetBaseOutputDirectory();
var artifactsOutputDirectory = $"{now:yyyy_MM_dd_HH_mm_ss}_{inputFileName}";
var outputDirectoryPath = _fileSystem.Path.Combine(baseDirectory, artifactsOutputDirectory);
_fileSystem.Directory.CreateDirectory(outputDirectoryPath);
using var audioStream = new FileStream(
_fileSystem.Path.Combine(outputDirectoryPath, "audio.mp3"),
FileMode.Create,
FileAccess.ReadWrite,
FileShare.ReadWrite,
4096,
FileOptions.Asynchronous
);
await _console.Status() await _console.Status()
.Spinner(Spinner.Known.Dots) .Spinner(Spinner.Known.Dots)
.StartAsync("Extracting audio...", async _ => audioStream = await _audioExtractor.ExtractMp3FromMp4Async(videoStream)); .StartAsync("Extracting audio...", async _ => await _audioExtractor.ExtractMp3FromMp4Async(videoStream, audioStream));
if (audioStream is null) if (audioStream is null)
{ {
@@ -101,14 +116,6 @@ internal sealed class DefaultCommand(
_console.MarkupLine($"[blue]Transcription completed[/] [green]successfully![/]"); _console.MarkupLine($"[blue]Transcription completed[/] [green]successfully![/]");
var now = _timeProvider.GetUtcNow();
var inputFileName = _fileSystem.Path.GetFileNameWithoutExtension(settings.Stream);
var baseDirectory = ValidateAndGetBaseOutputDirectory();
var artifactsOutputDirectory = $"{now:yyyy_MM_dd_HH_mm_ss}_{inputFileName}";
var outputDirectoryPath = _fileSystem.Path.Combine(baseDirectory, artifactsOutputDirectory);
_fileSystem.Directory.CreateDirectory(outputDirectoryPath);
await _fileSystem.File.WriteAllTextAsync( await _fileSystem.File.WriteAllTextAsync(
_fileSystem.Path.Combine(outputDirectoryPath, "transcription.txt"), _fileSystem.Path.Combine(outputDirectoryPath, "transcription.txt"),
string.Join(Environment.NewLine, transcriptionSegments) string.Join(Environment.NewLine, transcriptionSegments)
@@ -25,7 +25,7 @@ public sealed class AudioExtractor : IAudioExtractor
_videoService = videoService ?? throw new ArgumentNullException(nameof(videoService), $"{nameof(videoService)} cannot be null"); _videoService = videoService ?? throw new ArgumentNullException(nameof(videoService), $"{nameof(videoService)} cannot be null");
} }
public async Task<Stream> ExtractMp3FromMp4Async(Stream video) public async Task<Stream> ExtractMp3FromMp4Async(Stream video, Stream audio)
{ {
if (video is null) if (video is null)
{ {
@@ -42,25 +42,29 @@ public sealed class AudioExtractor : IAudioExtractor
throw new ArgumentException("Video stream must be seekable", nameof(video)); throw new ArgumentException("Video stream must be seekable", nameof(video));
} }
if (audio is null)
{
throw new ArgumentNullException(nameof(audio), "Audio stream cannot be null");
}
if (audio.CanWrite is false)
{
throw new ArgumentException("Audio stream must be writable", nameof(audio));
}
var originalPosition = video.Position; var originalPosition = video.Position;
try try
{ {
var mp3Stream = new MemoryStream(); var wasExtracted = await _videoService.ExtractAudioFromVideoAsync(video, audio).ConfigureAwait(false);
using var mp4Stream = new MemoryStream();
await video.CopyToAsync(mp4Stream).ConfigureAwait(false);
mp4Stream.Position = 0;
var wasExtracted = await _videoService.ExtractAudioFromVideoAsync(mp4Stream, mp3Stream).ConfigureAwait(false);
if (wasExtracted is false) if (wasExtracted is false)
{ {
throw new FailedAudioExtractionException("Failed to extract audio from the video stream."); throw new FailedAudioExtractionException("Failed to extract audio from the video stream.");
} }
mp3Stream.Position = 0; audio.Position = 0;
return mp3Stream; return audio;
} }
catch (Exception e) when (e is not FailedAudioExtractionException) catch (Exception e) when (e is not FailedAudioExtractionException)
{ {
@@ -9,11 +9,14 @@ public interface IAudioExtractor
/// Extracts MP3 audio from an MP4 video stream. /// Extracts MP3 audio from an MP4 video stream.
/// </summary> /// </summary>
/// <param name="video">The input video stream.</param> /// <param name="video">The input video stream.</param>
/// <param name="audio">The output audio stream where the extracted MP3 will be written.</param>
/// <returns>A stream containing the extracted MP3 audio.</returns> /// <returns>A stream containing the extracted MP3 audio.</returns>
/// <exception cref="ArgumentNullException">Thrown when the video stream is null.</exception> /// <exception cref="ArgumentNullException">Thrown when the video stream is null.</exception>
/// <exception cref="ArgumentException">Thrown when the video stream is not readable.</exception> /// <exception cref="ArgumentException">Thrown when the video stream is not readable.</exception>
/// <exception cref="ArgumentException">Thrown when the video stream is not seekable.</exception> /// <exception cref="ArgumentException">Thrown when the video stream is not seekable.</exception>
/// <exception cref="ArgumentNullException">Thrown when the audio stream is null.</exception>
/// <exception cref="ArgumentException">Thrown when the audio stream is not writable.</exception>
/// <exception cref="FailedAudioExtractionException">Thrown when the audio extraction fails.</exception> /// <exception cref="FailedAudioExtractionException">Thrown when the audio extraction fails.</exception>
/// <remarks>The method will preserve the passed video stream's data and position.</remarks> /// <remarks>The method will preserve the passed video stream's data and position.</remarks>
Task<Stream> ExtractMp3FromMp4Async(Stream video); Task<Stream> ExtractMp3FromMp4Async(Stream video, Stream audio);
} }
@@ -12,14 +12,17 @@ internal sealed class FFMpegService : IVideoService
{ {
public async Task<bool> ExtractAudioFromVideoAsync(Stream video, Stream audio) public async Task<bool> ExtractAudioFromVideoAsync(Stream video, Stream audio)
{ {
return await FFMpegArguments var inputArguments = video is FileStream videoFileStream
.FromPipeInput(new StreamPipeSource(video)) ? FFMpegArguments.FromFileInput(videoFileStream.Name)
.OutputToPipe( : FFMpegArguments.FromPipeInput(new StreamPipeSource(video));
new StreamPipeSink(audio),
static o => o.DisableChannel(Channel.Video).ForceFormat("mp3") Action<FFMpegArgumentOptions> arguments = static o => o.DisableChannel(Channel.Video).ForceFormat("mp3");
)
.ProcessAsynchronously() var processor = audio is FileStream audioFileStream
.ConfigureAwait(false); ? inputArguments.OutputToFile(audioFileStream.Name, addArguments: arguments)
: inputArguments.OutputToPipe(new StreamPipeSink(audio), addArguments: arguments);
return await processor.ProcessAsynchronously().ConfigureAwait(false);
} }
public async Task CreateClipFromVideoAsync( public async Task CreateClipFromVideoAsync(
@@ -9,13 +9,15 @@ public class AudioExtractorTests
{ {
using var testVideo = TestData.GetTestVideo(); using var testVideo = TestData.GetTestVideo();
using var extractedAudio = TestData.GetExtractedAudio(); using var extractedAudio = TestData.GetExtractedAudio();
using var audioStream = new MemoryStream();
var result = await _sut.ExtractMp3FromMp4Async(testVideo); var result = await _sut.ExtractMp3FromMp4Async(testVideo, audioStream);
var audioBytes = await ConvertStreamToBytesAsync(extractedAudio); var audioBytes = await ConvertStreamToBytesAsync(extractedAudio);
var resultBytes = await ConvertStreamToBytesAsync(result); var resultBytes = await ConvertStreamToBytesAsync(result);
resultBytes.Should().Equal(audioBytes); resultBytes.Should().Equal(audioBytes);
result.Should().BeSameAs(audioStream);
} }
private static async Task<byte[]> ConvertStreamToBytesAsync(Stream stream) private static async Task<byte[]> ConvertStreamToBytesAsync(Stream stream)
@@ -25,7 +25,7 @@ public class AudioExtractorTests
[Fact] [Fact]
public async Task ExtractMp3FromMp4Async_WhenVideoIsNull_ItShouldThrow() public async Task ExtractMp3FromMp4Async_WhenVideoIsNull_ItShouldThrow()
{ {
var action = async () => await _sut.ExtractMp3FromMp4Async(null!); var action = async () => await _sut.ExtractMp3FromMp4Async(null!, new MemoryStream());
await action.Should().ThrowAsync<ArgumentNullException>(); await action.Should().ThrowAsync<ArgumentNullException>();
} }
@@ -36,7 +36,7 @@ public class AudioExtractorTests
var mockStream = new Mock<Stream>(); var mockStream = new Mock<Stream>();
mockStream.Setup(s => s.CanRead).Returns(false); mockStream.Setup(s => s.CanRead).Returns(false);
var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object); var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object, new MemoryStream());
await action.Should().ThrowAsync<ArgumentException>(); await action.Should().ThrowAsync<ArgumentException>();
} }
@@ -48,7 +48,7 @@ public class AudioExtractorTests
mockStream.Setup(s => s.CanRead).Returns(true); mockStream.Setup(s => s.CanRead).Returns(true);
mockStream.Setup(s => s.CanSeek).Returns(false); mockStream.Setup(s => s.CanSeek).Returns(false);
var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object); var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object, new MemoryStream());
await action.Should().ThrowAsync<ArgumentException>(); await action.Should().ThrowAsync<ArgumentException>();
} }
@@ -69,7 +69,7 @@ public class AudioExtractorTests
) )
.ThrowsAsync(new Exception()); .ThrowsAsync(new Exception());
var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object); var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object, new MemoryStream());
await action.Should().ThrowAsync<FailedAudioExtractionException>(); await action.Should().ThrowAsync<FailedAudioExtractionException>();
} }
@@ -90,7 +90,7 @@ public class AudioExtractorTests
) )
.ReturnsAsync(false); .ReturnsAsync(false);
var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object); var action = async () => await _sut.ExtractMp3FromMp4Async(mockStream.Object, new MemoryStream());
await action.Should().ThrowAsync<FailedAudioExtractionException>(); await action.Should().ThrowAsync<FailedAudioExtractionException>();
} }
@@ -111,7 +111,7 @@ public class AudioExtractorTests
) )
.ReturnsAsync(true); .ReturnsAsync(true);
var result = await _sut.ExtractMp3FromMp4Async(mockStream.Object); var result = await _sut.ExtractMp3FromMp4Async(mockStream.Object, new MemoryStream());
result.Should().BeAssignableTo<Stream>(); result.Should().BeAssignableTo<Stream>();
result.Should().BeOfType<MemoryStream>(); result.Should().BeOfType<MemoryStream>();
@@ -137,7 +137,7 @@ public class AudioExtractorTests
) )
.ReturnsAsync(true); .ReturnsAsync(true);
await _sut.ExtractMp3FromMp4Async(stream); await _sut.ExtractMp3FromMp4Async(stream, new MemoryStream());
stream.Position.Should().Be(positionToRead); stream.Position.Should().Be(positionToRead);
} }