Extract OCR from a pdf/ppt has only images(No Text)

[yamuna83]

I am using kernel memory to process set of documents(pdf, docx, images, ppt, pptx). Works fine.
When I process a pdf file with only images then it is not extracting ocr.
Getting this error:
warn: Microsoft.KernelMemory.DocumentStorage.AzureBlobs.AzureBlobsStorage[0] The file user-documents/000d0000-ac13-0242-b807-08de3cc174d8/Digital_Strategy.pdf.extract.txt is empty warn: Microsoft.KernelMemory.Handlers.SaveRecordsHandler[0] Pipeline 'user-documents/000d0000-ac13-0242-b807-08de3cc174d8': step save_records: no records found, cannot save, moving to next pipeline step.

My current OCR settings:
"ImageOcrType": "AzureAIDocIntel",

Any help would be ppreciated.

[dluc]

hi @yamuna83, unfortunately that's a known limitation and the version of KM you are using has been archived. That said, the fix should be simple enough, patching this file

kernel-memory/archived/km-v1/service/Core/Handlers/TextExtractionHandler.cs

Lines 186 to 235 in 94b69d3

	private async Task<(string text, FileContent content, bool skipFile)> ExtractTextAsync(
	DataPipeline.FileDetails uploadedFile,
	BinaryData fileContent,
	CancellationToken cancellationToken)
	{
	// Define default empty content
	var content = new FileContent(MimeTypes.PlainText);
	if (string.IsNullOrEmpty(uploadedFile.MimeType))
	{
	uploadedFile.Log(this, $"File MIME type is empty, ignoring the file {uploadedFile.Name}");
	this._log.LogWarning("Empty MIME type, file '{0}' will be ignored", uploadedFile.Name);
	return (text: string.Empty, content, skipFile: true);
	}
	// Checks if there is a decoder that supports the file MIME type. If multiple decoders support this type, it means that
	// the decoder has been redefined, so it takes the last one.
	var decoder = this._decoders.LastOrDefault(d => d.SupportsMimeType(uploadedFile.MimeType));
	if (decoder is not null)
	{
	this._log.LogDebug("Extracting text from file '{0}' mime type '{1}' using extractor '{2}'",
	uploadedFile.Name, uploadedFile.MimeType, decoder.GetType().FullName);
	content = await decoder.DecodeAsync(fileContent, cancellationToken).ConfigureAwait(false);
	}
	else
	{
	uploadedFile.Log(this, $"File MIME type not supported: {uploadedFile.MimeType}. Ignoring the file {uploadedFile.Name}.");
	this._log.LogWarning("File MIME type not supported: {0} - ignoring the file {1}", uploadedFile.MimeType, uploadedFile.Name);
	return (text: string.Empty, content, skipFile: true);
	}
	var textBuilder = new StringBuilder();
	foreach (var section in content.Sections)
	{
	var sectionContent = section.Content.Trim();
	if (string.IsNullOrEmpty(sectionContent)) { continue; }
	textBuilder.Append(sectionContent);
	// Add a clean page separation
	if (section.SentencesAreComplete)
	{
	textBuilder.AppendLineNix();
	textBuilder.AppendLineNix();
	}
	}
	var text = textBuilder.ToString().Trim();
	return (text, content, skipFile: false);