FIR Classifier

This is the prototype developed for the classification of FIRs written in the language Kannada, it is deployed at : https://fir-classification.herokuapp.com/

Working model

• Page to enter the case of the text

• Page to upload the PDF

• Sample case statement : XYZ received a call that he had got a job offer and was asked to send his personal documents and money after which he was not given a job.The classification is aptly obtained as the type JOB SCAM

Local Setup

Steps to follow to locally setup the project

• Ensure to have Python with version >= 3.6
• Preferably create a virtual environment
• Clone the repository
• Install all the dependencies by pip install -r requirements.txt
• To install POPPLER one of the dependencies, this article can be followed for those on Windows : https://newbedev.com/how-to-install-poppler-on-windows, for those on Ubuntu or Mac OS this canbe followed : https://pdf2image.readthedocs.io/en/latest/installation.html
• Run streamlit run app.py from the root directory and open PORT 8501

Deployment

In order to deploy this application, I have compiled all the steps in a blog

https://namyalg.medium.com/deploy-a-streamlit-app-on-heroku-with-ocr-c0ec2a3dab2e