Qstrader

* Fixed bug involving NaN at Timestamp in sixty_forty example.
* Removed support for python 3.7 and 3.8
* Updated the python package requirements to work with matplotlib 3.8, numpy 1.26 and pandas 2.2.0

* Updated the python package requirements to work with matplotlib 3.4, numpy 1.21 and pandas 1.3
* Removed support for python 3.6
* Added a Tactical Asset Allocation monthly momentum strategy to the examples

* Added link to full documentation at [https://www.quantstart.com/qstrader/](https://www.quantstart.com/qstrader/)
* Fixed bug where burn-in period was still allowing portfolio rebalances and trade executions
* Added QSTrader Dockerfiles for various Linux distributions
* Removed support for Python 3.5 and added support for Python 3.9
* Increased minimum supported Pandas version to 1.1.5 from 0.25.1
* Modified end-to-end backtest integration test to check for approximate equality of results to fix differences across Pandas versions
* Disallowed Matplotlib 3.3.3 temporarily to avoid deprecated functionality from causing errors
* Event print messages during backtests can now be disabled through a boolean setting

* Added VolatilitySignal class to calculate rolling annualised volatility of returns for an asset
* Removed errors for orders that exceed cash account balance in SimulatedBroker and Portfolio. Replaced with console warnings.

* Significant overhaul of Position, PositionHandler, Portfolio, Transaction and SimulatedBroker classes to correctly account for short selling of assets
* Addition of LongShortLeveragedOrderSizer to allow long/short leveraged portfolios
* Added a new long/short leveraged portfolio example backtest
* Added some unit and integration tests to improve test coverage slightly

* Added ValueError with more verbose description for NaN pricing data when backtest start date too early
* Removed usage of 'inspect' library for updating attributes of Position within PositionHandler
* Added unit tests for Cash asset, StaticUniverse, DynamicUniverse and string colour utility function
* Added two more statistics to the JSON statistics calculation