meeting-assistant/MeetingAssistant.Tests/AzureSpeechRecognitionPipelineTests.cs

using MeetingAssistant.Recording;
using MeetingAssistant.Transcription;

namespace MeetingAssistant.Tests;

public sealed class AzureSpeechRecognitionPipelineTests
{
    [Fact]
    public async Task FinishedTranscriptUsesLiveDiarizedAzureSegments()
    {
        await using var pipeline = new AzureSpeechRecognitionPipeline(
            new StaticTranscriptionProvider(
            [
                new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(1), "Guest-1", "hello"),
                new TranscriptionSegment(TimeSpan.FromSeconds(1), TimeSpan.FromSeconds(2), "Guest-2", "there")
            ]));

        await pipeline.InitializeAsync(CancellationToken.None);
        await pipeline.WriteAsync(new AudioChunk([1, 0], 16000, 1), CancellationToken.None);
        await pipeline.CompleteAsync(CancellationToken.None);
        var liveSegments = new List<TranscriptionSegment>();
        await foreach (var segment in pipeline.ReadLiveTranscriptAsync(CancellationToken.None))
        {
            liveSegments.Add(segment);
        }

        var finishedSegments = await pipeline.ReadFinishedTranscriptAsync(
            "ignored.wav",
            new SpeechRecognitionPipelineOptions(5),
            CancellationToken.None);

        Assert.Equal(liveSegments, finishedSegments);
        Assert.Collection(
            finishedSegments,
            first => Assert.Equal("Guest-1", first.Speaker),
            second => Assert.Equal("Guest-2", second.Speaker));
    }

    private sealed class StaticTranscriptionProvider : IStreamingTranscriptionProvider
    {
        private readonly IReadOnlyList<TranscriptionSegment> segments;

        public StaticTranscriptionProvider(IReadOnlyList<TranscriptionSegment> segments)
        {
            this.segments = segments;
        }

        public async IAsyncEnumerable<TranscriptionSegment> TranscribeAsync(
            IAsyncEnumerable<AudioChunk> audio,
            SpeechRecognitionPipelineOptions options,
            [System.Runtime.CompilerServices.EnumeratorCancellation] CancellationToken cancellationToken)
        {
            await foreach (var _ in audio.WithCancellation(cancellationToken))
            {
            }

            foreach (var segment in segments)
            {
                yield return segment;
            }
        }
    }
}