Added support for multi-asset portfolio and general improvements

[matteorosato]

This pull request adds support for multi-asset portfolio while maintaining backward compatibility with the original single-asset behavior. The update introduces a new simulate_portfolio_returns function to simulate investments in portfolios composed of multiple assets, with customizable portfolio composition.

A Jupyter Notebook CustomPortfolio.ipynb has been added to demonstrate how to use the tool with a sample two-fund portfolio.

Key Changes:

1. simulate_portfolio_returns:

   • Introduced a new function to handle multi-asset portfolios, allowing users to specify the composition of the portfolio (asset weights).
   • The portfolio is defined as a dictionary where asset tickers are the keys, and the corresponding portfolio percentages are the values.

2. simulate_multiple_investments:

   • Updated to support the simulation of portfolios. If a portfolio_composition is provided, the function calls simulate_portfolio_returns. If no composition is provided, it defaults to the original behavior of simulating single asset investments.

3. simulate_multiple_durations:

   • Modified to accept and pass the portfolio_composition argument to simulate_multiple_investments, enabling simulations for multiple durations with multi-asset portfolios.

4. New Notebook: Notebooks/CustomPortfolio.ipynb:

   • A new Jupyter Notebook has been added to illustrate how to use the simulation tool with a portfolio composed of two assets. The notebook walks through the steps of simulating and analyzing portfolio returns for a user-defined asset composition.

5. Tests:

   • The unit tests have been rewritten to be compliant with the unittest framework.

6. Improved README:

   • The README.md has been updated to reflect the latest changes, including extra information on installation and usage.

7. Improved import_data function:

   • The import_data function has been improved to handle multiple DataFrames more efficiently.

[matteorosato]

@apalladi this is more a general question: why are we using average capital instead of total one when computing the gross return?