* Improves stability and fixes various bugs
* Adds unit tests covering changes
* Adds COVID-19 crisis plots
* Adjusts styling of crisis plots for more pleasant viewing
* Fixes bug where null BacktestResult caused application to crash
* Order JSON bug fixes and stability improvements
* MaxDrawdownReportElement now produces results for Live
* Replaced Estimated Capacity w/ Days Live
* Added Live marker to sharpe ratio
* Added support for MOO and MOC orders in PortfolioLooper
* Address review: adds new unit tests and cleans up code
* Bug fix: use LastFillTime instead of Order.Time for MOO and MOC
* Address review: Fixes tests and cleans up code
* Deletes NullAlphaHandler, NullLeanManager, NullSynchronizer
* Calculate the backtest and live PointInTimePortfolios only once now
* Refactor Metrics calculations
* Add missing license headers to some files
* Reverts accessibility of AddToUserDefinedAlgorithm to private
* Other misc. fixes and cleanup
* Can handle null Result packets
* Created utility files
* Added various helper methods to PortfolioLooper
* Fixes build issue by removing System.Collections.Immutable
* Updates plots to show "Insufficient data" when it can't be created
* Hides empty crisis page
* Fixes wkhtmltopdf display bug
* Removes Calculations.cs
* Modifies accessibility of AddToUserDefinedUniverse in QCAlgorithm
* Add null value handling in OrderJsonConverter
* Various bug fixes
* Fixes broken ReportChartTests.py
* Adds leverage to PointInTimePortfolio
We now will calculate all of the metrics needed to directly plot the
data in C# instead of doing calculations in Python.
This has added benefits in that the code can be easily reused to generate
plots from Result packets and makes the underlying plotting library we
use more extensible and replaceable.
* Adds Calculations.cs
* Refactors some of ReportCharts.py
* Adds Drawdown classes
* Adds Deedle and MathNet.Numerics packages
* Completes calcuations in C# side instead of Python
Side note: I think it's a little funny that we were previously
calculating the plots via Pandas, which uses a C backend. Effectively, we
were transferring data between at least 2 separate FFI
boundaries: C# -> Py -> C -> ? (whatever BLAS is written in)
Gerardo Salazar