How ravingupta/budgeting works
Project type and purpose
• This is a lightweight, full-stack, monolithic budgeting web application.
• Purpose: Allow users to upload bank CSV statements, auto-categorize transactions via built-in parsers and ML logic, store everything locally in a database, and present a dashboard UI.
File structure
• Top-level scripts:
• main.py – Boots the FastAPI/Uvicorn server
• migration.py – Initializes database schema
• setup.sh / start.sh – Environment and launch scripts
• app/ – FastAPI route definitions (accounts.py, transactions.py, dashboard.py)
• models/ – SQLModel ORM definitions (Account, Category, Transaction, Base)
• parsers/ – CSV parser framework (parser.py) and specific bank parser (scotia.py)
• workers/ – Background analyzer for advanced categorization logic (uses Pandas, Transformers, TensorFlow-Keras)
• scripts/ – Data-feed helpers for seeding accounts/categories
• templates/ – Jinja2 HTML templates (accounts, transactions, dashboard) styled with Tailwind CSS
• config.py & constants/apis.py – Configuration and constant definitions for APIs, file paths, DB connection strings
README for additional insights
• Highlights:
• FastAPI + Uvicorn for backend
• SQLModel for local DB
• Pandas + ML stack for transaction categorization
• Jinja2 + Tailwind CSS for frontend
• No external bank APIs—everything runs locally on CSV input.