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 engine defines securities for each universe to properly track them within
the data feed. These securities are not tradable and have no price data associated
with them, and as such, we should not be sending history requests for these symbols.
This change removes all universe symbols from history requests.
NOTE: Requests made directly to the history provider are not filtered out, as the
filtering happens within the QCAlgorithm implementation.
- Adds log to display the python version the algorithm is using.
- Fixes python algorithms that were failing because of small subtleties
like leading zeroes.
- Updates pythonnet with a version compiled with python 3.6 flags
- Changes in DockerfileFoundation: we now use miniconda to manage the python
environment.
- Took the opportunity to add NTLK (#1349), Tensorforce (#1369) and
PyTorch/Pyro (#1385).
- Changes readme in Algorithm.Python to show steps to install miniconda
This was added to confirm no changes of substance happened as a result of
altering the subscription synchronization code. Leaving these in as debug
aids in case of further failures w/ this algorithm.
In this update, methods overloads with decimal parameters accept python float.
- Fixes FractionalQuantityRegressionAlgorithm:
With the pythonnet update we can pass a python float where a decimal is required.
In this PR we are disabling the default security seeding (automatically getting the last price for a security when added to the algorithm) for a couple reasons, both when using large universes:
- In live trading, these history requests are sent to a history server, potentially causing timeouts
- In backtesting, depending on the algorithm this could also cause slowdowns up to 30%
Requiring a period here forces signal models to place a time frame on
when their signal is valid. This also allows consumers of signals to
have some expectation of when a prediction should come to fruition.
Michael Handschuh