Skip to content

kjdarthvader/ai-ta-backend

 
 

Repository files navigation

AI TA Backend for UIUC's Course Assistant Chatbot

A Flask application hosting endpoints for AI TA backend.

🛠️ Technical Architecture

Hosted (mostly for free) on Railway. Architecture diagram of Flask + Next.js & React hosted on Vercel. Architecture diagram

Documentation

Automatic API Reference

📣 Development

  1. Rename .env.template to .env and fill in the required variables
  2. Install Python requirements pip install -r requirements.txt
  3. Start the server for development (with live reloads) cd ai_ta_backend then flask --app ai_ta_backend.main:app --debug run --port 8000

The docs are auto-built and deployed to our docs website on every push. Or you can build the docs locally when writing:

  • mkdocs serve

Course metadata structure

'text': doc.page_content,
'readable_filename': doc.metadata['readable_filename'],
'course_name ': doc.metadata['course_name'],
's3_path': doc.metadata['s3_path'],
'pagenumber': doc.metadata['pagenumber_or_timestamp'], # this is the recent breaking change!!
# OPTIONAL properties
'url': doc.metadata.get('url'), # wouldn't this error out?
'base_url': doc.metadata.get('base_url'),

About

API backend for UIUC AI Teaching Assistant.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 96.7%
  • Jupyter Notebook 3.1%
  • Other 0.2%