Almost all of us have gone through the tedious process of filing expense claims, which involves copying over information from the receipt to the claims form. Not only is this process painstaking, but made even worse when you cannot simply Ctrl-C and Ctrl-V information from the receipt because the file is an image.
I myself have been through this before, as such I decided to make receiptScanner; An Automatic Receipt Parser to address this. receiptScanner is able to extract necessary information from a receipt that you will most likely need to file for expense claims, such as "Receipt No, Total Amount Paid, Date of Expense made" etc. This information will be saved as a .csv file
This repo's source code is heavily based on the invoice2data, so I would like to thank the authors for creating the original package.
Install the necessary required packages through the provided requirements.txt
file using Anaconda prompt.
pip install -r requirements.txt
This repo uses Tesseract-OCR, to install Tesseract-OCR for Windows, visit the
following link and download the
appropriate Tesseract installer .exe depending on your computer specs. If
Tesseract-OCR is successfully installed, you should find a folder labelled
Tesseract-OCR under C:\Program Files\Tesseract-OCR.
Note that some have reported it being found under C:\Program Files (x86) or
C:\Users\USER\AppData\Local\Tesseract-OCR instead. Do take note if you are
unable to locate it under C:\Program Files as originally expected.
- Clone this repo:
git clone https://github.com/leonardlohky/receiptScanner
Similar to the original invoice2data package, receipt scanner extracts texts
from receipt PDF files or images using Tesseract. After that, it searches for
regex in the extracted text using a YAML-based template system to extract the
necessary field data. It then saves the result as a CSV file.
To run the application, navigate to the location of main.py script and type
in the following in an Anaconda prompt.
cd src\receiptScanner
python main.py <INPUT_FILES_OR_FOLDER>
#Example
python main.py .../docs/receipt1.pdf
python main.py .../doc/receipts_folder
Distributed under the MIT License