An application to assist immigration attorneys and refugee representatives in advocating for clients in asylum cases by identifying patterns in judicial decisions and predicting possible outcomes based on those patterns.
Human Rights First (HRF) is a non-profit, nonpartisan, 501(c)(3), international human rights organization based in New York, Washington D.C., Houston, and Los Angeles. HRF works to link immigration attorneys and advocates with asylum seekers and provide those attorneys with resources to represent their clients. Our application leverages historical data to inform advocates better of a judge's past decisions. The hope is that advocates for asylum seekers can use our tools to tailor their arguments before a particular judge and maximize their client's chances of receiving asylum.
This diagram shows the current state of the architecture.
This diagram shows the flow of data between frontend, backend, data science, and the database.
Front-End
Uses NodeJS to create the web-based user interface for uploading case documents, managing users, and viewing data in the form of tables and visualizations.
Back-End
Uses Javascript and Postgres to manage databases containing tables for users, judges, and cases.
Data Science
Uses Python and Tesseract for optical character recognition (OCR) to convert pdf images into text data that can be searched via natural language processing (NLP) techniques. Key data, which we refer to as structured fields, are extracted from the text data and sent to the back-end for storage.
See this team's history here.
The app folder contains the FastAPI application which creates an endpoint for processing a single .pdf file stored in the AWS S3 bucket.
- VISUALIZATIONS_INFO contains information pertaining to the visualizations.py as well as the main.py file and how they interact with each other.
- SCRAPER_INFO contains important information about the status of each function used to extract fields from documents.
- READMEdev contains AWS CLI2 installation instruction
- KnownDefects contains information about potential issues to be aware of.
- NOTEBOOKS contains helpful information about the notebooks and scripts used to explore solutions for scrapper.
- visualizations/README contains helpful information about notebooks, scripts, and data used to explore visualizations.
- Pytesseract - package for converting pdfs to text
- FastAPI - API for endpoint that processes pdfs and data visualizations
- AWS Elastic Beanstalk - Service for app deployment
After cloning the repository, in your command line run the following commands: Replace with a name of your choice.
docker build . -t <name>
docker run -it -p 5000:5000 <name> uvicorn app.main:app --host=0.0.0.0 --port=5000
NOTE: An error may be thrown when trying to run the app if you have not added the .env file with aws credentials
NOTE: env file should have 4 env variables namely ACCESS_KEY,SECRET_KEY, BUCKET_NAME, DB_URL
NOTE: Uncomment COPY .env .env in dockerfile before run docker locally
Then open http://0.0.0.0:5000 in your browser. The application should be running.
If 0.0.0.0 does not work, try http://localhost:5000
Ivan Mihailov
Emily A. Frener
Mathew Mahoney
Azamatkhuja Jalilov
Lucas Chatham
Trevor Dewalt
Azamatkhuja Jalilov
Matthew Blackburn
Zunera Bokhari
Joel Hoover
Lee Rowe
Minh Nguyen
Waqas Khwaja
Jason Young
Andrew Lee
Michael Carrier
Dylan Sivori
Frank Howd
Malachi Ivey
Jacob Bohlen
Kevin Weatherwalks
River Bellamy
Patrick Raborn
Ricardo Rodriguez
Filipe Collares
Nicholas Major
Evan Grinalds
Joe Sasson
Michael Kolek
Francis LaBounty
Jennifer Faith
Brett Doffing
Daniel Fernandez
Kevin Weatherwalks
Alex Krieger
Bharath Gogineni
Jace Hambrick
Nicholas Adamski
Rassamy J. Soumphonphakdy
Rebecca Duke Wiesenberg
Reid Harris
Lucas Petrus
Noah Caldwell
Tomas Phillips
Sean Byrne
Henry Gultom
RJ Proctor
Liam Cloud Hogan
Steven Lee
Edwina Palmer
Tristan Brown
This project is licensed under the terms of the MIT license.