Saath Baara Utara OCR, The 7/12 OCR

By X Æ A-3

Hosted Link:

http://caaca1b72e90.ngrok.io/OCR712/ [ please use the choose file button, hosting was causing issues with drag and drop :) ] (link is down)

This projectis aimed at helping the Global Parli Foundation in their mission to improve rural India through a replicable model of Rural Rejuvenation.

In order to achieve this we aim to

• build a hosted OCR web service
• which converts the 7/12 extract (“Saath Baara Utara”)
• to an editable excel file.

Winning project of the Code For Change Hackathon

Detailed Documentation in ppt

PPT

Basic Prototype

Final Design

Our Features!

• Upload singular or multiple Saath Baara Utara pdfs at once using drag and drop or browse.
• Google ocr converts each of these files to a text document in devnagri script
• Using pandas and python we extract usefull information from these converted text. Basically, we extract variables from the text documents and create columns for a excel file based on them for easy readability and comparision between multiple 7/12 Extracts.
• FINAL EXCEL

Advantages!

• In our opinion the biggest advantage we provide is the multiple file support => EXCEL as this will be helpfull for the NGO to compare thousands of 7/12 extracts at a glance using excel functionalities.
• Our next advantage would be fast and reliable OCR service, with a error rate of only 1.2 %.
• No one in the market currently provide all these functionalities bundled together.

Basic Steps to Run the Code:

clone the repo

pip install -r requirements.txt

Refer requirements_NOTE.txt

python manage.py migrate

python manage.py makemigrations

python manage.py runserver

Run at http://127.0.0.1:8000/OCR712/