Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
Report Generator Tests / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
API Tests / build (push) Has been cancelled
* Fix half of the CA1051 warnings
This warning is about not declaring visible instance fields. There are
something about 500 warnings in the solution, mostly in the QuantConnect and QuantConnect.Algorithm.CSharp projects. I aim to fix one of them in this PR and the other half of them in a second one. To fix it, I'm changing the visible instancce fields for properties.
* fix bugs
* Addressing minor reviews
* More minor fixes
---------
Co-authored-by: Martin Molinero <martin.molinero1@gmail.com>
* Add support for FTP notifications
* Add optional parameters for SFTP notifications with private key
* Add optional "secure" parameter for FTP notifications.
Improve FTP notifications constructor and input checks.
* FTP notification improvements.
Add public key property.
Add multiple constructors and methods for FTP and SFTP notifications with either password or SSH keys.
Encode file contents.
* Address peer review
Add support for FTP notification with string contents
* Minor changes
* Minor changes
* Remove public key argument
* Fixes and minor changes
Regression Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
Benchmarks / build (push) Has been cancelled
Build & Test Lean / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
* Refactor alpha statistics
- Refactor alpha statistics, cleaning up and simplifying no longer required calculations and scoring
- Adding new InsightEvaluator abstraction, adding C# & PY regression
algorithms
* Optimization backtest result json converter update
* Address reviews
- Remove IAlphaHandler, move insight storage responsability to IResultHandler
and centralizing insight collection on the QCAlgorithm.Insights to be
reused by the framework models
- Fix portfolio turnover single day backtests and duplicate time
sampling handling. Updating regression algorithms
* Add InsightCollection tests and minor fixes
* Adding more & improved tests
* add data count properties
* 'add history count property
* assert data counts
* update missing override
* consider override/virtual cases
* implement data count
* add message handler for regression tests
* use regression test message handler
* set algorithm manager for regression test message handler
* update data count
* check if stats are present, check if algo manager is not null
* update
* add c# algo
* make same as c# algo
* use new line
* logic shifted to RegressionTestMessageHandler
* cleanup
* auto cleanup
* skip non deterministic data count
* change data count
* use inheritance
* improve stats
* update couht
* add sma indicator to c# and customSMA to python
* call base method before executing further
* skip test
* revert to original
* add duplicate sma
* skip regression test
Regression Tests / build (push) Has been cancelled
Build & Test Lean / build (push) Has been cancelled
* Binance fees deducted from fill quantity accordengly
- For Binance cash accounts while buying, if fees are from the base
currency of leans virtual position, we need to deduct the fee from the
fill quantity, else we can end with a position bigger that it actually
is and not be able to liquidate
* Refactor solution
- Refactor solution into a more generic approach solving fees in base
currency at the BrokerageTransactionHandler level, covering all
brokerages that require it. Adding regression algorithm reproducing
issue.
- Update Bitfinex and Binance fee models to correctly reflact reality
* Log fill quantity adjusment once
Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
- After https://github.com/QuantConnect/Lean/pull/5872 trading API
changes numpy float64 was not converted correctly by pythonNet and
used an int. Reverting API changes and adding regression test. This
should be fixed at pythonNet layer
Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
* Fix SubscriptionDataReader auxiliary data out of order
- Fix for SubscriptionDataReader emitting auxiliary data out of order
due to sparse that and enumerator refresh logic. Adding regression
algorithm
- Minor tweaks for unit tests failing on and off
* Fix for SDR internal subscriptions
- Fix for SubscriptionDataReader enumerator refresh for internal
subscriptions which were ignored. Adding regression test
- Updating custom data regression algorithms affected by the issue
- DropboxBaseDataUniverse was emitting a custom data point being end time
- UnlinkedTraderBarIconicType was emitting a single data point
of the underlying SPY minute data when if should of emitted
all data points
Regression Tests / build (push) Has been cancelled
Build & Test Lean / build (push) Has been cancelled
* Filter out small orders based on Setting
- BuyingPowerModel will filter out small orders based on algorithm
setting, a % of PTV, instead of hard coded 1 share value. Addin unit
and regression tests
- Updating regression algorithms to use new setting, reduce order trades
* Update regression algorithms
* Refactor HasSufficientBuyingPowerForOrder implementations
Adds Sufficient and Insufficient helper methods to HashSufficientbuyingPowerForOrderParameters
enabling syntax like:
return paraeeters.Sufficient()
returnparameters.Insufficient(reason)
The next change will add the initial margin required which will simply require
updating both of these helper methods to accept the value.
* IBuyingPowerModel: Add margin functions Maintenance/Initial/ForOrder
These were originally hidden in an effort to only expose what's necessary
for the engine to perform its work. Additionally, we encapsulated all of
the method arguments into parameters classes to prevent having to break
anyone in the future. Not including these foundational methods turns out to
be an oversight. These methods are not required by the engine, but rather by
other models. Another possible solution here is to add an additional abstraction
and include these methods on this new abstraction. BuyingPowerModel would then
explicitly implement these methods and models that depend on them would require
two code paths, one for when the buying power model implements this interface
and another for when it doesn't.
Tests were additionally updated to remove test model implementations created for
the sole purpose of exposing these private methods.
* Add ConstantBuyingPowerModel
Provides an implementation of IBuyingPowerModel that returns the same
constant value
* Update BuyingPowerModelPythonWrapper to use reflection for method names
Having a bunch of hard-coded strings is a sure fire way for someone to
overlook when changing methods. This change ensures that noone needs to
remember that this code exists :)
Cleans up the syntax around verifying a python object implements a particular
C# interface via the ValidateImplementationOf<T> method by having it return a
value since the only use cases are in constructors when setting the models.
I was initially going to update ALL python wrappers to validate the passed
in models, but such a change could break many things that are 'working' right
now. Such an effort should be saved for its own dedicated PR.
* Add Parameters/Result types for new buying power model methods
* Support computing maintenance margin for arbitrary quantities
The existing GetMaintenanceMargin function assumes that we're only interested
in the maintenance margin for the entirety of the provided security's holdings.
This makes it impossible to perform what-if analysis or to even ask how much
maintenance margin is devoted to a particular subset of the security's holdings.
This change adds the quantity to the MaintenanceMarginParameters class. Futures
and Options models also depend on holdings cost and holdings value, so they have
also been added to the parameters type. Finally, static factory methods were
added to improve discernment of intent: ForCurrentHoldings provides the existing
behavior and then ForQuantityAtCurrentPrice to support what-if scenarios where
we're looking for the change in maintenance margin if we were to execute an order
for the securiy at the current time step. Obviously a constructor is provided to
set all of the values explicitly, using any price metric the caller desires.
* Address review
- Fix BPM xml documentation
- Fix python unit tests and PythonWrapper validate method
* Add SecurityHolding.QuantityChanged event
Adding event handlers will allow us to orchestrate complex
events from distant parts of the codebase through wiring
them up. If we continue down this path, it will move us away
from the current, very 'mechanical' data flows expressed in
LEAN and towards a more modern, event processing based system.
This is but a baby step in that direction and the initial use
case is using this QuantityChanged event to trigger resolution
of the algoritm's positions groups. This is part of an effort
to improve the fidelity of options margin modeling where we'll
model an OptionStrategy as an IPositionGroup. This will allow
us to compute the margin requirements of an OptionStrategy as
a unit instead of computing margin of each security individually
in isolation.
See #4065
* PortfolioManager: Group fields and remove unused field
This codebase generally places fields as the first members, but
this class had some fields at the top, then some properties, and
then some more fields. This change brings all the fields together
at the top of the file and also removes pointless comments placed
directly above some of the fields. Additionally, an unused field
was removed.
* Remove unused _currencyConverter from Security
Looks like at some point the only code using this member variable was removed
and the necessary clean up was overlooked.
* Add Parse.Enum functions
* Support disabling regression algorithms by language via config.json
Adds 'regression-test-languages' to config.json and filters regerssion algorithms to
run based on this value. When cycling on a particular feature, it's nice to be able
to run the entire regression set while ignoring the python algorithms. Once the C#
algorithms are all passing, one can then go back and run C# and Python in a final run,
since 99% of feature work doesn't impact python specifically.
* Implement IComparable in SecurityIdentitfier
This can be used to deterministically sort securities and symbols
* Add .editorconfig to enforce common formatting for json/sh files
* Fix typo in IBuyingPowerModel.GetBuyingPower xml docs
* Add ListEquals/GetListHashCode and OrderDirection.Closes(PositionSide)
ListEquals and GetListHashCode are designed to be used together as they
complement each other according to C#'s requirements for Equals and
GetHashCode functions.
PositionSide.ToOrderDirection() extension simply converts a PositionSide
to its logical equivalent OrderDirection. Long->Buy, Short->Sell, None->Hold
OrderDirection.Closes(PositionSide) determines if a particular OrderDirection
would have the effect of reducing a position's absolute size. This function
greatly improves the readability of buying power functions that must provide
adjustments when an order/contemplated trade reduces/closes an existing position.
OrderDirection.Buy.Closes(PositionSide.Short)
OrderDirection.Sell.Closes(PositionSide.Long)
All other combinations return false
Adds ToArray/ToImmutableArray convenience functions that combine a call
to Select followed by To(Immutable)Array all in one function call.
* Add decimal.DiscretelyRoundBy extension method
Supports rounding a decimal value by an arbitrarily chosen maximum precision,
or 'quanta'
* Update FutureMarginBuyingPowerModelTests to respect the security's lot size
* Add core position group classes and abstractions
* Add initial/maintenance margin support, buying power model consistency tests
* Add SufficientBuyingPower and GetReservedBuyingPower to position group model
Includes update to BrokerageTransactionHandler to use position group BPM for
sufficient buying power checks.
* Resolve position groups on each fill
We need to update the state of our position groups on each fill so that
we can properly handle multiple orders within the same time step. We
also limit the number of positions sent into the resolver by removing
securities without any holdings.
* fixup! Add SufficientBuyingPower and GetReservedBuyingPower to position group model
* Add GetMaximumLotsFor{Target|Delta}BuyingPower
Instead of computing order quantity, these functions compute the
maximum number of position group lots, which is the position group
quantity, and is guaranteed to be a whole number, for the provided
target/delta buying power parameters.
The SecurityPositionGroupBuyingPowerModel delegates to the security's
IBuyingPowerModel by applying a scaling factor equal to the security's
lot size.
This change also updates references to IBuyingPowerModel.GetMaximum...
to use the new position group model methods.
* Convert remaining IBuyingPowerModel call sites to position groups
* Rename PositionManasger.CreateDefaultGroup -> GetOrCreateDefaultGroup
Better describes its behavior
* Add Position Groups readme.md
* Add Option Strategy BuyingPowerModel
- Adding CompositePrositionGroupResolver and
OptionStrategyPositionGroupResolver
- Adding OptionStrategyPositionGroupBuyingPowerModel handling option
strategies based on IBs margin table. Adding regression algorithms
- Few changes so that option strategies executed by multiple orders are
detected
- Adjust OptionStrategyDefinitionMatch to include equity legs in the
matching result
- Minor tweaks fixing previous rebase
- Minor fixes for existing option strategies definitions, adding new
missing strategies.
- Fixing minor bugs in option strategy matcher. Adding more unit tests
* Address self reviews
- Fixing bug in 'PositionGroupCollection'
- Few minor simplificaitons
- Adding BasicTemplateOptionEquityStrategyAlgorithm
* Address reviews
- Improve regression algorithms margin remaining and used assert logic to be exact. Taking into account spread and fees
Co-authored-by: Michael Handschuh <mhandschuh@gmail.com>
* Update OrderListHash to use MD5 as hash instead of hash code
* Update regression algorithm OrderListHash statistic
* Use full MD5 hash as OrderListHash, update regression statistic
* Fixes failing regression tests
- Adding new SerializedOrderEvent and SerializedOrder with new
JsonConverters
- Specifying OrderEvent json converter when storing, streaming data
- Adding unit tests
Can now simply flip a flag in the code at the top of Messaging.cs or
specify via configuration 'regression-update-statistics' and it will
automatically update the code statistics using the statistics from
the current run. Useful when updating factor files and need to update
all statistics :P
The term 'alpha' is used to describe the entire algorithm. Therefore, 'alpha'
produces insights. From this we have things like IAlphaModel, which is the model
defining how insights are produced. We have IAlphaHandler, which defines how the
insights from a single 'alpha' (the algorithm) are managed, analyzed, and stored.
Types closer to the individual prediction level, such as InsightDirection, or
InsightScore relate directly to exactly 1 insight. The distinction between the
two became more clear as we developed the insights API, and from that effort it
was decided to harmonize alpha/insight terminology across the various QC systems.
The IAlphaManagerExtension defines a type that needs to react to events produced
by the AlphaManager. The actual events were removed in favor of a interface to
handle the events. This removes the need to wire events and instead just pass the
extensions to the alph manager and it will handle invoking the extensions at the
appropriate time.
This change removes all charting and statistics aggregation logic from the alpha
handler and moves it into dedicated types, AlphaChartingManagerExtension and
AlphaStatisticsManagerExtension. The resulting types are highly decoupled from the
LEAN ecosystem allowing them to be easily unit tested, whereas before the logic
was embedded in a handler with many many dependencies which would be very hard to
properly unit test.
As part of this change (and in preparation for moving scoring to the alpha thread)
the resolution of SecurityValues was removed from the alpha manager. In this new
pattern, the alpha manager is pushed generated alphas and security values at each
time step.
When the algorithm finishes, the alpha statistics are merged with the backtest
result statistics. This keeps the regression testing pattern of alpha statistics
the same as the existing regular statistics.
Ricardo Andrés Marino Rojas