Portfolio Modeling

A comprehensive multi-currency portfolio tracking and analytics system built with FastAPI, SQLAlchemy, and Streamlit.

Features

• 📊 Multi-currency portfolio tracking
• 📈 Financial metrics calculation (Sharpe ratio, time-weighted returns, etc.)
• 🔄 Automatic market data updates via Yahoo Finance
• 🌐 RESTful API with FastAPI
• 📱 Web interface with Streamlit
• 💾 SQLite database with SQLAlchemy ORM

Structure

• config.py: Configuration settings with environment variable support
• startup.py: CLI tool for database initialization and server management
• main.py: FastAPI application entry point
• requirements.txt: Project dependencies
• database/:
  • access.py: Database access and operations
  • entities.py: Database entities/models defined by SQLAlchemy ORM
  • init_db.py: Database initialization script
• services/:
  • action_service.py: Manages trading actions
  • metric_service.py: Calculates financial metrics
  • portfolio_service.py: Manages portfolios
  • yfinance_service.py: Interacts with Yahoo Finance API
• routers/: FastAPI route handlers
  • action_router.py: Action-related endpoints
  • metric_router.py: Metrics and analytics endpoints
  • portfolio_router.py: Portfolio management endpoints
• frontend/: Streamlit web interface
• utils/: Utility functions

Quick Start

1. Install Dependencies

    pip install -r requirements.txt
    ```
2. Initialize the database:
    ```bash
    python startup.py startup
    ```
    
## TODO
 - ~~read and insert actions needs to be able to interpret USD as currency~~
 - ~~action service to read excel data, clean/interpret, insert/update database~~
 - ~~portfolio service updates portfolio and portfolioholdings tables~~
 - ~~change DatabaseAccess methods to return session queries, not detached data objects.  Like done "in get_portfolio_asset_actions()". Session lifecycle moves to business logic layer.~~
 - ~~working on update_holding_time_series() in DatabaseAccess~~
 - ~~time series for holdings in kind - separate table or dataframe in ram. - separate table~~
 - ~~transform in-kind holdings to currency denominated holdings test performing in ram with pandas vs sql querying~~
 - ~~simple plotting for proof of concept~~
 - ~~want to be able to view metrics for combinations of portfolios and assets~~
- ~~backend services as fastapi~~
 - adding core metric calcualtion functions - ~~value invested~~ ~~cost basis~~, ~~unrealized gain/loss~~, ~~time-weighted return~~, ~~sharpe ratio~~, realized gain/loss, get all
 - summary statistics for metrics - for each portfolio, total value, total invested, total unrealized gain/loss, ytd return, 1 year return, 30 day return, sharpe ratio
 - tax calculations, FIFO method (USA), cost basis method (Iceland)
 - add enpoints for all data outputs
 - frontend
 - dividends robust
 - action defined on precise datetime (currently data aggregates on day, we lose fidelity of inter day transactions)