Azure AI Document Intelligence
An Azure service that turns documents into usable data. Previously known as Azure Form Recognizer.
1,528 questions
Running document intelligence to extract Layout but the output given is from a diffeent model (general-document)
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(108.80000000000001, 13%, 20%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EJM%3C/text%3E%3C/svg%3E)
Jacopo Morabito
5
Reputation points
Hi, while using the Python SDK to request the analysis of a PDF document from the Document Intellicence service, we are getting different result between docs tested on document intelligence studio end the results coming back from the SDK call.
More details:
when we test the Layout (modelid="prebilt-layout") model to extract tables in the Azure Document Intelligence Studio we get the tables extracted correctly. From the SDK, the tables are extracted with some missing columns, the weird thing is that the result is the same as if the model used is the General Documents (modelid="prebuilt-document"). Is it possible that the Prebuilt-layout model is not yet supported or that there is some parameter I am missing?
Runtime Python3.11
API version used: tested all of them
{count} vote
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(22.400000000000002, 11%, 12%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ERM%3C/text%3E%3C/svg%3E)
Ramr-msft 17,731 Reputation points2024-01-28T05:07:55.73+00:00 Thanks for the question, Can you please share the pdf document to check. With the Document Intelligence 2023-10-31-preview, the general document model (prebuilt-document) is deprecated. Going forward, to extract key-value pairs from documents, use the
prebuilt-layoutmodel with the optional query string parameterfeatures=keyValuePairsenabled.API Version: Ensure that you’re using the same API version in both the SDK and the Studio. Different versions may have different features or behaviors.
- Model ID: Double-check that you’re using the correct model ID (
prebuilt-layout) in your SDK calls. - Parameters: Make sure you’re passing the correct parameters in your SDK calls. For example, if you want to extract key-value pairs, you should enable the
features=keyValuePairsoption.
- Model ID: Double-check that you’re using the correct model ID (
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(67.2, 39%, 16%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EAR%3C/text%3E%3C/svg%3E)
Aditya R Bhat 0 Reputation points2024-05-29T09:12:57.0233333+00:00 Same issue happening with me as well. I am trying to extract data from a PDF. If i run latest DI version deployed in East US in DI studio, it is able to extract everything. But if i extract using Python SDK, it is missing a specific paragraph. I am using the prebuilt-layout model.
We are planning to scale the project to extract content from 2000 PDFs. Please tell me how to fix this.
I am using Doc Int API version 2023-10-31-preview
Sign in to comment