How ssa-d/ai-tutor works
AI Tutor is a personal study and career assistant. In the implemented web workflow, a user submits a question to the FastAPI endpoint, which checks configuration and calls the LLM adapter for a DeepSeek reply. The Feishu workflow receives chat messages, handles log commands and confirmations, and routes ordinary conversation through persistent chat memory; it can also summarize that day’s saved logs. The README describes Canvas reminders, recommendations, résumé support, and a 3D interface as future plans, not implemented components. The map follows documented behavior and sampled source; planned capabilities have no verified runtime wiring.
Main entry points: llm.py LLM Adapter · feishu.py Message Handler · main.py FastAPI Routes · memory.py Chat Memory
Components 14
Web Q&A
- FastAPI Routes
app/main.py - Configuration
app/config.py - LLM Adapter
app/llm.py
Feishu Assistant
- Message Handler
app/feishu.py - Daily Log Flow
app/feishu.py
Memory and Logs
- Chat Memory
app/memory.py - SQLite Storage
app/storage.py
Planned Capabilities
- Canvas Reminders
- Competition and Planning
- Résumé and Jobs
- 3D Interface
Other components
- User
- DeepSeek API
- Feishu Platform
Connections 15
- User → FastAPI Routes: submits question
- FastAPI Routes → Configuration: checks key
- FastAPI Routes → LLM Adapter: asks question
- LLM Adapter → Configuration: reads settings
- LLM Adapter → DeepSeek API: sends prompt
- DeepSeek API → LLM Adapter: returns answer
- LLM Adapter → FastAPI Routes: returns reply
- Feishu Platform → Message Handler: delivers message
- Message Handler → Feishu Platform: replies via IM
- Message Handler → Daily Log Flow: routes log intent
- Daily Log Flow → SQLite Storage: reads and writes logs
- Daily Log Flow → LLM Adapter: requests summary
- Message Handler → Chat Memory: updates conversation
- Chat Memory → SQLite Storage: loads and saves history
- Message Handler → LLM Adapter: requests conversation