- We will now check if python selection method returned `Universe.Unchanged`
- Removing `ToList()` call on fine and coarse data before sending it to
the python algorithm
- Adding regression algorithms
- Adding new `InsightWeightingPortfolioConstructionModel` that will
generate percent `Targets` based on the latest active `Insight` `Weight` per
`Symbol`.
- Will ignore `Insights` that have no `Weight`.
- If the sum of all the last active `Insight` per `Symbol` is bigger than 1, it
will factor down each target percent holdings proportionally so the sum is 1.
- Adding unit tests
- Adding a new regression test framework algorithm
- Note most of the code, including tests, are reused from the
`EqualWeightingPortfolioConstructionModel`
- Removing `using QCAlgorithmFramework = QuantConnect.Algorithm.QCAlgorithm`
- Removing `QCAlgorithmFrameworkBridge`
- Removing `IsFrameworkAlgorithm`
- Making `EmitInsightBasedOnFill` private. Adding new
`IOrderEventProvider` exposing an `event` to which `QCAlgorithm` will
subscribe.
- `AccountType.Cash` algorithms will be allowed to manually trade and
emight insights manually or with alpha model.
- This commit is related to PR 19 in QC/pythonnet
- C# decimal will be cast to C# double and converted into python
float
- Adding new `decimal.py` into the python algorithm project. This is
required for backwards compatibility with users performing operations
over expected decimal types (like `Price`)
- Updating two python regression test algorithms using custom python
execution models to be aware and ignore floating point precision errors
when handling order sizing.
`RsiAlphaModel` and `BlackLittermanOptimizationPortfolioConstructionModel` didn't have a key check after a history request. If a history request retuns no data for a given symbol, trying to access the pandas dataframe results in a `KeyError`.
`RsiAlphaModel` and `BlackLittermanOptimizationPortfolioConstructionModel` didn't have a key check after a history request. If a history request retuns no data for a given symbol, trying to access the pandas dataframe results in a `KeyError`.
Securities from different types may have different timezones. In this case, daily resolution data is split in two slices in a history request, so these slices are grouped together to determined whether we have data from all the symbols in the same date.
- Moves `PearsonCorrelationPairsTradingAlphaModel` class to its own file in order to make it available as a framework model.
- Follows existing pattern design that alpha models receive a lookback and a `Resolution` object.
- Algorithm implements `IRegressionAlgorithmDefinition`.
- Since it will be used as a base class for other pairs trading models, it was ranamed as BasePairsTradingAlphaModel
- Use a tuple of symbols are key of BasePairsTradingAlphaModel._pairs dictionary.
Instead of using a single, pre-defined, pair set in the class constructor, the pair is defined when securities are changed, therefore depending on the universe selection model.
We are not using python lists instead of generator (yield) because we get better exception information in this case. The aim is to lead users to avoid using generators and/or know its limitations.
The exception in ignored because the generator isn't closed until it is being deleted (automatically in this case, when Python exits); the generator __del__ handler closes the generator, which triggers an exception of there is one.
The PairsTradingAlphaModel is a simple example of defining an insight
grouping. Insights that are grouped together are assigned a unique
group-id that can be used by the portfolio construction model.
Updates were made to the CommonAlphaModelTests to give more control to
derived types. Some changes are still needed here to give securities
unique prices. I would recommend using a psuedo-random walk approach
by using Random with a constant seed value.
Stefano Raggi