How vishvas-dubey/ai_event_bot works
Project type and purpose
• This is a full-stack web application (“AI_EVENT_BOT”) that provides a conversational AI assistant for workshop participants.
• It uses a Python Flask backend, a simple HTML/CSS/JavaScript frontend, and integrates with Google Gemini for natural language understanding, resume processing, recommendations, and feedback collection.
File structure
Top level:
• app.py / main.py / newapp.py – Flask entry points and routing
• utils.py, audio.py – helper modules (e.g., audio handling, shared utilities)
• style.css – frontend styling
• resume/ – uploaded PDF resumes
• *.json / *.csv (registrations.csv, participants.json, agenda.json, location.json, confirmed_users.json) – flat-file data stores
• requirements.txt, Dockerfile – build & deployment
README for architecture insights
• Backend system broken into modules: Authentication, Resume Processing, Event Info, Recommendation Engine, Feedback Collection
• Frontend: chat UI, quick-access buttons, session display, feedback form
• AI integration: Google Gemini for language understanding and resume analysis
Main components
a. Users & Clients
• Web browser or mobile browser (UI)
b. Frontend Layer
• HTML/CSS/JavaScript chat interface
• Quick-access buttons & feedback form
c. Backend Layer (Flask Application)
• Authentication Service (verifies against registrations.csv/confirmed_users.json)
• Resume Processing Module (uploads to resume/, invokes Google Gemini API)
• Event Information Service (reads agenda.json, location.json, time logic)
• Recommendation Engine (matches extracted skills to sessions)
• Feedback Collection Service (stores star ratings and comments into CSV/JSON)
• Utility Services (audio processing, logging, configuration)
d. Data Stores
• Flat files: CSV/JSON for registration, participants, agendas, locations, feedback
• Resume file store (resume/*.pdf)
e. External Services
• Google Gemini API (NLP, resume analysis, context-aware responses)
Relationships and interactions
• User interacts via the Frontend → sends HTTP requests (AJAX/WebSocket) → Flask Backend
• Backend Authentication Service reads CSV/JSON → responds with access grant or denial
• On resume upload, Backend calls Google Gemini → parses response → stores skill profile
• Event Info Service monitors server time → replies to queries (current session, countdown, locations)
• Recommendation Engine reads skill profiles + agenda → returns recommended sessions
• Feedback Module captures ratings/comments → appends to feedback CSV/JSON
• All modules communicate internally via function calls or shared utils
Architectural patterns & design principles
• Layered architecture (Presentation / Application / Data)
• Modular services within the Flask app for separation of concerns
• Stateless HTTP endpoints, with state persisted in flat files (could evolve to a database)
• External AI service orchestration for NLP and resume analysis