Merge remote-tracking branch 'robcarver17/master' into missing-data

Author: todd

CHANGELOG.md

@@ -481,7 +481,7 @@ Moved most examples except core to separate git [here](https://github.com/robcar
 
 * Added methods to accountCurveGroup.get_stats(): .mean(), .std(), .tstat(), .pvalue()
 * Added method to accountCurveGroup stack; stack object can also produce bootstrap
-* Added account_test(ac1, ac2) to produce a t-test statistic for any two account curve like objects.
+* Added account_t_test(ac1, ac2) to produce a t-test statistic for any two account curve like objects.
 
 ## Version 0.6.0
 

CONTRIBUTING.md

@@ -1,7 +1,7 @@
-# pysystemtrade contributing guide
+# Pysystemtrade contributing guide
 
 
-## unit tests
+## Unit tests
 This project has a few unit tests. They get run automatically when any PR is 
 submitted. You'll see the result of the run in the PR page. To run the tests
 yourself locally, before submitting, you'll need `pytest` installed. Then run:
@@ -27,7 +27,7 @@ pytest --runslow
 ```
 
 
-## lint
+## Lint / Black
 
 This project keeps its code pretty with 
 [Black](https://black.readthedocs.io/en/stable/). Black gets automatically run 
@@ -47,3 +47,40 @@ Note for pycharm users: The blackd plugin requires a blackd daemon to be running
 Or, configure your local git install to automatically check and fix your code
 as you commit. Configuration instructions 
 [here](https://black.readthedocs.io/en/stable/integrations/source_version_control.html)
+
+## General code guidelines (INCOMPLETE)
+
+These guidelines are aspirations, and do not describe the system as it stands. The project has been written over a period of several years, and it is only quite recently I've decided to set out some guidelines. 
+
+In general, we try and follow the original texts: [PEP 8](https://peps.python.org/pep-0008/) and [clean code](https://gist.github.com/wojteklu/73c6914cc446146b8b533c0988cf8d29).
+
+### General
+
+- Unless there is a single parameter, parameters should be explicit.
+- Type hints should be used, with Unions if required `from typing import Union`
+
+
+### Naming conventions
+
+- For classes, I prefer mixedCase to CamelCase, but single word names should always be Camels.
+- Common methods are `get`, `calculate`, `read`, `write`.
+- There is a specific procedure for naming objects which form part of the data heirarchy, see [here](https://github.com/robcarver17/pysystemtrade/blob/master/docs/data.md#part-2-overview-of-futures-data-in-pysystemtrade). If this is not followed, then the [automated abstraction of data inside Data 'blob' instances](https://github.com/robcarver17/pysystemtrade/blob/master/docs/data.md#data-blobs) won't work.
+- Although arguably redundant, I am a fan of describing eg objects that inherit from dicts with a dict_ prefix. This gives hints as to how they behave without having to look at their code.
+
+
+### Error handling
+
+- Production code should not throw an error unless things are completely unrecoverable; if it does throw an error it must also log.critical which will email the user
+
+
+### Caching
+
+FIXME This is a bit of a mess - Update when a unified cache system setup
+
+
+### Testing
+
+Doc tests should be removed from class methods, since they often require a lot of setup, and make the code harder to read. Unit tests are preferable.
+Doc tests make more sense for seperate, standalone, functions.
+
+Test coverage is extremely sparse. 

docs/backtesting.md

@@ -3693,7 +3693,7 @@ returns the two sided t-test statistic and p-value for a null hypothesis of a
 zero mean.
 
 Sometimes you might want to compare the performance of two systems, instruments
-or trading rules. The function `from syscore.accounting import account_test`
+or trading rules. The function `from syscore.accounting import account_t_test`
 can be used for this purpose. The two parameters can be anything that looks
 like an account curve, no matter where you got it from.
 

syscore/algos.py

@@ -1,29 +1,40 @@
 """
+WARNING DO NOT REFACTOR NAMES OR LOCATION OF THIS CODE AS IT IS USED IN .YAML FILES, eg
+
+capital_multiplier:
+   func: syscore.capital.fixed_capital
+
 Functions to calculate capital multiplier
 
-ALl return Tx1 pd.Series
+ALl return Tx1 pd.Series, where a value of 1 is the original capital
+
+See https://qoppac.blogspot.com/2016/06/capital-correction-pysystemtrade.html for more details
+
 """
 from copy import copy
 import pandas as pd
 import numpy as np
+from systems.basesystem import System
 
 
-def fixed_capital(system, **ignored_args):
+def fixed_capital(system: System, **ignored_args) -> pd.Series:
     multiplier = copy(system.accounts.portfolio().percent)
     multiplier[:] = 1.0
 
     return multiplier
 
 
-def full_compounding(system, **ignored_args):
+def full_compounding(system: System, **ignored_args) -> pd.Series:
     pandl = system.accounts.portfolio().percent
     multiplier = 1.0 + (pandl / 100.0)
     multiplier = multiplier.cumprod().ffill()
 
     return multiplier
 
 
-def half_compounding(system, **ignored_args):
+def half_compounding(system: System, **ignored_args) -> pd.Series:
+
+    ## remove any nans
     pandl = system.accounts.portfolio().percent.curve().ffill().diff()
     multiplier = 1.0
     multiplier_list = []