How based-watermelon/pesu_reimagined works
PESU Reimagined is a FastAPI-backed university portal assistant with student, faculty, parent, and guest modes. Users log in to retrieve academic records or ask role-aware questions; guests can ask about public university information. The main flow combines portal data, role filtering, institutional knowledge retrieval, and Gemini-generated answers. The README describes a future portal-replacement goal but gives little architectural detail; links for frontend interactions are therefore not inferred from the HTML files, which were not sampled.
Main entry points: main.py FastAPI routes · ai_handler.py AI response generation · security.py Prompt assembly · data_fetcher.py Student records
Components 15
API and sessions
- FastAPI routes
root/main.py - Session management
root/auth.py - Session data cache
root/main.py
University data
- Student records
root/data_fetcher.py - Faculty records
root/data_fetcher.py - Guest information
root/data_fetcher.py
AI assistant
- Role data filtering
root/security.py - Prompt assembly
root/security.py - Institutional knowledge
root/knowledge - Knowledge retrieval
root/rag.py - AI response generation
root/ai_handler.py
Web interface
- Login and dashboard
root/frontend
Other components
- Portal user
- PESU Academy
- Gemini API
Connections 18
- Portal user → Login and dashboard: uses
- Login and dashboard → FastAPI routes: requests
- FastAPI routes → Session management: creates or checks
- FastAPI routes → Session management: deletes
- FastAPI routes → Student records: fetches records
- FastAPI routes → Faculty records: fetches records
- FastAPI routes → Guest information: loads public data
- Student records → PESU Academy: reads portal data
- Faculty records → PESU Academy: reads portal data
- FastAPI routes → Session data cache: stores and reads
- FastAPI routes → Role data filtering: filters chat data
- Role data filtering → Prompt assembly: supplies filtered data
- Prompt assembly → Knowledge retrieval: retrieves context
- Institutional knowledge → Knowledge retrieval: loads text
- Prompt assembly → AI response generation: builds prompt
- FastAPI routes → AI response generation: submits prompt
- AI response generation → Gemini API: generates answer
- AI response generation → FastAPI routes: returns answer