Finding pagenumber of table from DocumentIntelligenceClient

Thibault Verlinde 120

Hiya

I'm currently working a project that analyzes documents into a search index. I'm trying to change from the Form Recognizer nuget to the Document Intelligence nuget packet. When I do, there is a problem selecting the analyzed tables. In the Form Recognizer, you could get the tables on a single page by doing this:

tables.Where(t => t.BoundingRegions[0].PageNumber == pageNumber).ToList();

In the Document Intelligence (version 1.0.0 beta 2), It's impossible to do this, since the BoundingRegions are for some reason 0, even though there are tables on the page analyzed.
Some help/guidance would be appreciated!

Thanks in advance!

VasaviLankipalle-MSFT 15,956 Reputation points

2024-03-18T20:22:05.08+00:00

Hello @Thibault Verlinde , Thanks for using Microsoft Q&A Platform.

Did you get a chance to check this documentation? https://learn.microsoft.com/en-us/dotnet/api/azure.ai.documentintelligence.boundingregion?view=azure-dotnet-preview

Is it possible to share the query and it's results with latest Document Intelligence nuget packet?
VasaviLankipalle-MSFT 15,956 Reputation points

2024-03-18T20:43:55.3966667+00:00
Hello @Thibault Verlinde , as mentioned in this thread, if you are working on Word document, please refer to the following information.

Versions 2024-02-29-preview, 2023-10-31-preview, and later support Microsoft office (DOCX, XLSX, PPTX) and HTML files. The following features are not supported:

There are no angle, width/height and unit with each page object.

For each object detected, there is no bounding polygon or bounding region.

Page range (pages) is not supported as a parameter.

No lines object.
Thibault Verlinde 120 Reputation points

2024-03-19T08:05:20.5833333+00:00
Hiya @VasaviLankipalle-MSFT ,
The documentation doesn't mention anything about the number of pages that the document intelligence client returns. I agree that page range isn't a supported parameter, but output stays one page.
This is a part of the used code (i'm analyzing a .docx file of 10 pages long):

var documentToAnalyze = new AnalyzeDocumentContent() { UrlSource = fileUri, }; Console.WriteLine("Analyzing..."); var operation = await _diClient.AnalyzeDocumentAsync(WaitUntil.Completed, modelId, documentToAnalyze); var results = operation.Value; var pages = results.Pages; Console.WriteLine($"The document has {pages.Count} page(s)."); //This will say: 1 page

As mentioned in the other thread you linked, this is expected behaviour.

Could there be another way to calculate the page number of a character in the analyzed output?

Thanks in advance!

Share via

Finding pagenumber of table from DocumentIntelligenceClient