Extraer informacion de pdf

Help Forum
,
1

hola a todos, necesito extraer informacion de un pdf, una factura, puedo seleccionar campos del pdf, eso quiere decir que es nativo, pero al momento de usar get text o find element, me toma la hoja completa, no lo que necesito seleccionar, necesito ayudar para poder resolver esto, extraer la informacion y escribirla en un excel, estoy detenida en el proceso de extraccion debido a esto.
agradeceria su ayuda o algun xaml de ejemplo.
muchas gracias!

2

Hi @mively

You can use Document Understanding to extract data from the Invoice. If you are completely New to UiPath and want a simpler approach, you can extract the native text and then use Regex to extract the data you want.

Can you show me a sample of the PDF you wish to extract? share it if possible, I can help with a xaml.

3

Hi @mively

Can you try with Extract Document Data activity

Regards,

4

image
image443×420 12 KB

estaria bien asi la actividad ?

5

me podria mas dar informacion de como hacerlo con document understanding, usando la actividad Extract Document Data
soy nueva en esto y me gustaria aprender y hacerlo de buena forma, lo agradeceria!

1 Like
6

@mively

Please check below images

A user interface for extracting document data, featuring options for the type of project, extractor, prompt, and file input path. (Captioned by AI)
A user interface for extracting document data, featuring options for the type of project, extractor, prompt, and file input path. (Captioned by AI)510×340 10.1 KB

The image displays a software interface for extracting document data by defining prompts for various field names such as PO Due Date, Revise, Description, and Header. (Captioned by AI)
The image displays a software interface for extracting document data by defining prompts for various field names such as PO Due Date, Revise, Description, and Header. (Captioned by AI)826×567 26.4 KB

Regards,

1 Like
7

no me aparece la misma informacion que a ti cuando utilizo la actividad

8

@mively

Can you try to upgrade the UiPath.DocumentUnderstanding.Activities version to 2.12.1

The image shows a list of dependencies for a UiPath project with the UiPath.DocumentUnderstanding.Activities package highlighted. (Captioned by AI)
The image shows a list of dependencies for a UiPath project with the UiPath.DocumentUnderstanding.Activities package highlighted. (Captioned by AI)457×386 60.8 KB

A dropdown menu in a software interface, displaying various document types, highlighting the "Generative Extractor" option. (Captioned by AI)
A dropdown menu in a software interface, displaying various document types, highlighting the "Generative Extractor" option. (Captioned by AI)395×617 53.9 KB

Below are the examples of prompts

The image shows a dictionary configuration screen for extracting document data, with fields for PO Due Date, Revise, Description, and Header each having corresponding generative prompts. (Captioned by AI)
The image shows a dictionary configuration screen for extracting document data, with fields for PO Due Date, Revise, Description, and Header each having corresponding generative prompts. (Captioned by AI)827×571 26.4 KB

Regards,

9

listo, me podrias explicar como funciona la actividad en si, que pongo en promp o como puedo extraer la informacion que yo necesito ? por favor

10

@mively

Based on the fields you need from the invoice, you should include them in the prompt. Please watch the video below for a better understanding

Regards,

2 Likes
11

Hi @mively

Abre UiPath Studio y haz clic en Administrar Paquetes. Ve a la pestaña Oficial, busca UiPath.DocumentUnderstanding.Activities y selecciónalo. Haz clic en Instalar y luego en Guardar para confirmar la instalación. Espera a que se complete el proceso y reinicia UiPath Studio si es necesario. Una vez instalado, podrás ver las actividades de Document Understanding en el panel de Actividades y usarlas en tus automatizaciones.

Si te ha sido útil, márcalo como solución.

1 Like
12

@mively

First what kind of document is it?

Check if any of the predefined models are of same type like invoice,purchase order etc so that you can use the model and in the output you would have the fields to get the data

If you dont have any predefined model then go with generative extractor and provide what all fields you need and how do you identify that field like for example if you need a total cost…say you need total and it would have a prefix total and you. Eed the value for that…like that create prompts if no predefined model works

If you want all data as datatable then output.resultasdatatable will give the data as wel which is one of the methods…similarly you can get individual fields as well

Cheers

13

@mively

Please go throw below thread and attached video

1 Like
14

hola! la información que se extrae de esta forma, como se podría ir escribiendo en un Excel, en las columnas que correspondan, tengo una idea, pero me gustaría saber diferentes camino para poder hacerlo.
Me ayudaría demasiado.
Saludos!

1 Like
15

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.