Extracting data from semi-structured pdf?

(0) Share

Report

Posted on by pbiadmin123

Hi all,

Wanted to see if anyone has run into this problem before. I have set of pdf's which are semi-structured (attached) and i'm hoping to extract data from the 'pesticide production information' section. The issue is that for each pdf, they can have varying number of pages, and varying number of sections per page (a new section starts when 42. is the first cell in the table). Is there any way to build an AI model that will extract the pesticide production information no matter the pdf format? Or will i have to train the model with a bunch of examples of varying types. Is it possible to utilize power automate to do this? Thanks in advance!

Categories:

AI Builder

Express Chem 3540...

001Form_3540_16_Y...

I have the same question (0)

All responses (1)

Answers (1)

Verified answer

takolota1 4,944 Moderator on at

Like (0)

Report

Extracting data from semi-structured pdf?

I recommend using OCR & GPT in this case, like the set-up in this template:

https://community.powerplatform.com/galleries/gallery-posts/?postid=31e67eea-3f73-47b4-95b7-fe4a7b646389

You can request GPT to output an array with a JSON object for each of the dynamic number of sections found in your PDF, just like the Product Lines in the invoice example.

Was this reply helpful? Yes No