Skip to content

Latest commit

 

History

History
105 lines (72 loc) · 4.82 KB

CHANGELOG.md

File metadata and controls

105 lines (72 loc) · 4.82 KB
Raw

0.3.0

  • Updates dependencies to use numpy v2.0.0.
  • Updates simulated_broker.py to change np.NaN to np.nan
  • Updates backtest_data_handler.py to change np.NaN to np.nan
  • Updates daily_bar_csv.py to change np.NaN to np.nan
  • Updates tests

0.2.9

  • Updates requirements file to use numpy v1.26.4 or lower. This is the last version of QSTrader that supports numpy<2.0.0.

0.2.8

  • Updates BacktestTradingSession.get_target_allocations() to use burn_in_dt.date() instead of burn_in_dt Timestamp. Previous method compared a Timestamp to a datetime.date.
  • Adds an integration test to check that target allocations match the expected output, including a date index.

0.2.7

  • Updates the execution handler to update final orders ensuring an execution order is created in the event of a single submission without a further rebalance.
  • Updates rebalance_buy_and_hold to check if the start_dt is a business day If start_dt is a business day rebalance_dates = [start_dt] If start_dt is a weekend rebalance_dates = [next business day]
  • Adds a unit test to check that the buisness day calculation is correct
  • Adds an integration test to check that a backtest using buy_and_hold_rebalance generates execution orders on the correct dates

0.2.6

  • Removed get_portfolio_total_non_cash_equity and get_account_total_non_cash_equity from broker/broker.py abstract base class. These methods are not implemented.
  • Added save option to TearsheetStatistics class in statistics/tearsheet.py. The tearsheet output can now be saved to a given filename by passing the optional filename parameter as a string when calling the plot_results function.

0.2.5

  • Moved build-backend system to Hatchling from setuptools
  • Updated the python package requirements to work with click 8.1
  • Updated ReadMe and ChangeLog.

0.2.4

  • Fixed bug involving NaN at Timestamp in sixty_forty example.
  • Removed support for python 3.7 and 3.8
  • Updated the python package requirements to work with matplotlib 3.8, numpy 1.26 and pandas 2.2.0

0.2.3

  • Updated the python package requirements to work with matplotlib 3.4, numpy 1.21 and pandas 1.3
  • Removed support for python 3.6
  • Added a Tactical Asset Allocation monthly momentum strategy to the examples

0.2.2

  • Added link to full documentation at https://www.quantstart.com/qstrader/
  • Fixed bug where burn-in period was still allowing portfolio rebalances and trade executions
  • Added QSTrader Dockerfiles for various Linux distributions
  • Removed support for Python 3.5 and added support for Python 3.9
  • Increased minimum supported Pandas version to 1.1.5 from 0.25.1
  • Modified end-to-end backtest integration test to check for approximate equality of results to fix differences across Pandas versions
  • Disallowed Matplotlib 3.3.3 temporarily to avoid deprecated functionality from causing errors
  • Event print messages during backtests can now be disabled through a boolean setting

0.2.1

  • Added VolatilitySignal class to calculate rolling annualised volatility of returns for an asset
  • Removed errors for orders that exceed cash account balance in SimulatedBroker and Portfolio. Replaced with console warnings.

0.2.0

  • Significant overhaul of Position, PositionHandler, Portfolio, Transaction and SimulatedBroker classes to correctly account for short selling of assets
  • Addition of LongShortLeveragedOrderSizer to allow long/short leveraged portfolios
  • Added a new long/short leveraged portfolio example backtest
  • Added some unit and integration tests to improve test coverage slightly

0.1.4

  • Added ValueError with more verbose description for NaN pricing data when backtest start date too early
  • Removed usage of 'inspect' library for updating attributes of Position within PositionHandler
  • Added unit tests for Cash asset, StaticUniverse, DynamicUniverse and string colour utility function
  • Added two more statistics to the JSON statistics calculation

0.1.3

  • Fixed bug involving DynamicUniverse not adding assets to momentum and signal calculation if not present at start of backtest
  • Modified MomentumSignal and SMASignal to allow calculation if available prices less than lookbacks
  • Added daily rebalancing capability
  • Added some unit tests to improve test coverage slightly

0.1.2

  • Added RiskModel class hierarchy
  • Modified API for MomentumSignal and SMASignal to utilise inherited Signal object
  • Added SignalsCollection entity to update data for derived Signal classes
  • Removed unnecessary BufferAlphaModel
  • Added some unit tests to improve test coverage slightly

0.1.1

  • Removed the need to specify a CSV data directory as an environment variable by adding a default of the current working directory of the executed script
  • Addes CI support for Python 3.5, 3.6 and 3.8 in addition to 3.7
  • Added some unit tests to improve test coverage slightly

0.1.0

  • Initial relase of QSTrader to PyPI