Overview
I work for the UC Davis Shared Services Center. A lot of my work is incredibly boring, involving manual reading for invoices and inputtng the data into the the university financial database system, Kuali (kfs.ucdavis.edu). This project automates the process of inputting transportation invoice data forms. It takes a PDF of the invoice as input, extracts the appropriate text using Google Vision, and inputs the data into the form using Selenium.
Ongoing
This is ongoing. Currently, I've been able to extract basic text like invoice/account#s, and check amount. The Selenium scripts puts the data into the form. Ongoing work involves attaching files on the form, and adding functionality for new forms.
Technologies Used
- Google Vision API
- Selenium
- Kuali financial database system
How to use
To run this project, you will need two things: UC Davis Kerberos credentials and Google Vision Authentication:
To get Google Auth, please visit: https://cloud.google.com/docs/authentication/production
Kerberos credentials are only given to those affiliated with the university. I intend to expand the scope of the project later to involve forms generally accessible to the public.
To run this project:
python interact_kfs.py
The following is an image of a raw document ready to be processed, with the important extractable information highlighted in red:
The following is an snippet of the finalized document with the inputted information. Note not all fields in the form are required.
