How bongsangkim/mijung works
Project type and purpose
• This is a full-stack, data-driven web application (“미정(味定)”) providing ingredient price trends, related ingredients, and recipe recommendations.
• It combines a React-based frontend, two backend services (Java/Spring Boot and Python/FastAPI), a MySQL database, big-data analysis (Hadoop/Spark), and CI/CD (Docker, Jenkins, Nginx, AWS).
File structure
Top-level directories:
• analysis: Python scripts & Jupyter notebooks for time-series forecasting (ARIMA, LSTM, Prophet), word2vec embedding, FPGrowth, data preprocessing.
• backend: Java/Spring Boot project (controllers, services, repositories) exposing REST APIs for ingredients, recipes, carts.
• fastapi: Python FastAPI microservice handling embedding, cosine similarity, scheduling jobs with APScheduler.
• frontend: React/TypeScript SPA (Zustand, Tanstack Query, D3.js/Sigma.js), static assets, routing, components.
• exec: Dockerfiles and docker-compose for backend, FastAPI, Jenkins, Nginx; certbot folder for SSL.
• hadoop: scripts and README for setting up Hadoop/Spark cluster.
• jenkins, nginx, docker-compose.yml: CI/CD pipelines and reverse-proxy configuration.
README insights
• Technology stack: React/Tailwind, Java Spring Boot, Python (data & FastAPI), MySQL, Hadoop/Spark, Docker, Jenkins, Nginx, AWS EC2.
• Features: time-series analysis, network graph (cosine similarity), association rules (FPGrowth), scheduled jobs, CI/CD.