MindWorkAI/AI-Studio
1
0 Fork 0
mirror of https://github.com/MindWorkAI/AI-Studio.git synced 2026-11-05 07:27:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Added transcription of audio and video files to local data sources (#1081)

Browse Source 
Co-authored-by: Paul Koudelka <106623909+PaulKoudelka@users.noreply.github.com>
This commit is contained in:
Thorsten SommerandPaul Koudelka authored and GitHub committed 2026-10-18 12:06:15 +02:00
1 parent 8642ce2d81
commit 752f7d828a
82 files changed
+3231 -302
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

app/MindWork AI Studio/Assistants/BatchProcessing/AssistantBatchProcessing.razor.Content.cs
+14 -88
View File
@@ -1,7 +1,4 @@
using System.Text;
using AIStudio.Tools.Media;
using AIStudio.Tools.Rust;
namespace AIStudio.Assistants.BatchProcessing;
@@ -69,108 +66,37 @@ public partial class AssistantBatchProcessing
private async Task<string?> LoadMediaTranscriptAsync(BatchProcessingFileResult fileResult, CancellationToken token)
{
var transcriptFilePath = GetTranscriptFilePath(fileResult.FilePath);
if (File.Exists(transcriptFilePath))
{
try
{
var existingTranscript = await File.ReadAllTextAsync(transcriptFilePath, token);
if (!string.IsNullOrWhiteSpace(existingTranscript))
{
this.Logger.LogInformation("Reusing the existing batch transcript '{TranscriptFilePath}' for media file '{MediaFilePath}'.", transcriptFilePath, fileResult.FilePath);
return existingTranscript;
}
this.Logger.LogWarning("The existing batch transcript '{TranscriptFilePath}' for media file '{MediaFilePath}' is empty and will be replaced.", transcriptFilePath, fileResult.FilePath);
}
catch (OperationCanceledException)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.CANCELED, T("The batch run was canceled."));
return null;
}
catch (Exception e)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, string.Format(T("Was not able to read the existing transcript: {0}"), e.Message), e);
return null;
}
}
if (!this.MediaTranscriptionService.HasUsableTranscriptionProvider)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, T("No usable transcription provider is configured."));
return null;
}
var transcription = await this.MediaTranscriptionService.TranscribeAsync(fileResult.FilePath, token);
if (transcription.Status is MediaTranscriptionResultStatus.CANCELLED)
var provider = this.MediaTranscriptionService.GetUsableProvider();
var transcript = await PersistentMediaTranscript.ReadOrCreateAsync(MediaSidecarTranscripts.INSTANCE, fileResult.FilePath, provider, this.MediaTranscriptionService.TranscribeAsync, this.Logger, token);
if (transcript.Status is PersistentMediaTranscriptStatus.CANCELLED)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.CANCELED, T("The batch run was canceled."));
return null;
}
if (transcription.Status is not MediaTranscriptionResultStatus.SUCCEEDED)
if (transcript.Status is not PersistentMediaTranscriptStatus.READY)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, transcription.UserMessage);
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, transcript.UserMessage);
return null;
}
if (string.IsNullOrWhiteSpace(transcription.Text))
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, T("The transcription provider returned an empty transcript."));
return null;
}
return await this.StoreMediaTranscriptAsync(fileResult, transcriptFilePath, transcription.Text);
}
private async Task<string?> StoreMediaTranscriptAsync(BatchProcessingFileResult fileResult, string transcriptFilePath, string transcript)
{
var tempFilePath = transcriptFilePath + ".tmp";
try
{
// Complete the small persistence step even if cancellation arrived
// after transcription, so the expensive provider result can be
// reused when the interrupted batch is continued.
await File.WriteAllTextAsync(tempFilePath, transcript, new UTF8Encoding(false), CancellationToken.None);
File.Move(tempFilePath, transcriptFilePath, true);
this.Logger.LogInformation("Stored the batch transcript '{TranscriptFilePath}' next to media file '{MediaFilePath}'.", transcriptFilePath, fileResult.FilePath);
return transcript;
return await File.ReadAllTextAsync(transcript.TranscriptPath, token);
}
catch (OperationCanceledException)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.CANCELED, T("The batch run was canceled."));
return null;
}
catch (Exception e)
{
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, string.Format(T("Was not able to store the transcript next to the media file: {0}"), e.Message), e);
this.FinishFileResult(fileResult, BatchProcessingFileStatus.FAILED, string.Format(T("Was not able to read the existing transcript: {0}"), e.Message), e);
return null;
}
finally
{
try
{
if (File.Exists(tempFilePath))
File.Delete(tempFilePath);
}
catch (Exception e)
{
this.Logger.LogWarning(e, "Was not able to remove the temporary batch transcript '{TempFilePath}'.", tempFilePath);
}
}
}
private static bool IsTranscribableMedia(string filePath) => FileTypes.IsAllowedPath(filePath, FileTypes.AUDIO, FileTypes.VIDEO);
private static bool IsTranscribableMedia(string filePath) => PersistentMediaTranscript.IsMediaPath(filePath);
private static string GetTranscriptFilePath(string mediaFilePath) => mediaFilePath + TRANSCRIPT_FILE_SUFFIX;
private static bool HasReusableTranscript(string mediaFilePath)
{
var transcriptFilePath = GetTranscriptFilePath(mediaFilePath);
try
{
return File.Exists(transcriptFilePath) && new FileInfo(transcriptFilePath).Length > 0;
}
catch
{
// The concrete read error is reported when the affected file is
// processed. Here we only decide whether a provider is required.
return File.Exists(transcriptFilePath);
}
}
private static bool HasReusableTranscript(string mediaFilePath) => MediaSidecarTranscripts.HasReusableTranscript(mediaFilePath);
}