- new Commission Report highlights where values need to be updated
- parquet capital setup bugs fixed
- ignore weekly GAS_US, LME contracts
- add support for import of split frequency CSV prices
- references to arctic updated in sysinit scripts
- CONTRIBUTING.md updated
- NO LONGER REQUIRES ARCTIC Time series data is stored in parquet, install pyarrow
- But does require newer versions of pandas and python, see requirements.txt
- See the discussion here to see how to switch from arctic to parquet
- added long only constraint to dynamic optimisation
- various bugs fixed
- Further progress in replacing old logging with python logging
- Simplify get recent data frame of ticks code and now also works with delayed data
Stale:
- Stale instruments now removed from all reports except slippage #1211
- Ignore stale strategies #1074
Rolls:
- improved auto rolling facility, with configurable rules #1198
- Clear up roll states, adding new don't open state and redefining close state #931
- change the way roll states are handled to prioritise strategy trades that help us roll #1993
- warn if rolls can't be done because of trade limits #932
No market data:
- Collect data without market data subscriptions #1165
- Allow trading of non subscribed instruments by implementing SNAP order types; replacing MARKET orders #1016
- Allow specific instruments to use specific algos (will be useful when start trading instruments without streaming data) #969
Costs:
- Now cost objects produce SR estimates for you #1026
- SR cost was calculated wrong (multiplied by two) due to confusion about turnover definition #1009
Other:
- Push sent and unsent emails to a file #1132
- Easier access to instrument return correlations from portfolio subsystem #1018
- Added backfilled GBP FX data #1014
- Added warning when priced contract is expiring #1006
- Added order simulator as an optimal replacement for vectorised p&l calculation; prequisite for limit order simulation
- Replace pst logging with python logging
- ignore daily expiries for certain EUREX contracts
- Allow fixed instrument and forecast weights to be specificed as a hierarchy
- Replaced log to database with log to file
- Won't generate roll order if order for instrument already on stack
- Removed warning code for empty spread data
- Moved storage of contract positions, instrument strategy positions, optimal positions from mongo/timedStorage to Arctic announcement
- Multiple instances of same IB symbol in config can now be resolved
- Instrument config and roll config now live purely in .csv
- Seperate database table for spread costs
- Logs have to have one of a limited number of labels
- Positions now use database stored expiries, not live IB expiries
- Portfolio beta in risk reports
- Many additional instruments added to config
- Can apply a long only constraint to instrument positions
- Live p&l now returned as accountCurve object
- Added optional code to run price collection throughout the day, see announcement
- Added ability to pass arguments to methods through control_config.yaml
- Added regions to instrument.csv configuration; use interactive_controls, option 5, option 52 to apply
- Added new run_ processes, breaking out FX, update sampled contracts, and update multiple/adjusted prices from daily price updates. You will need to update your crontab and control_config.yaml, see discussion here
- Refactoring syscore
- Changed handling of expensive instruments, see this discussion
- Added requirement for BLACK see this discussion
- Roll report clean up format discussed here
- Gradual removal of missing_data, missing_contract type objects and replace with exceptions
- Algos won't be allocated just before day end discussed here
- Timezones can now be manually configured discussed here
- Auto copying of instrument config data between db and csv
- Utility to cross check IB configuration multipliers
- Utility to delete instrument data
- Custom private directory
- Utility to modify roll config
- fix for #729. Added hourly and daily per contract prices IMPORTANT YOU MUST RUN /sysinit/futures/create_hourly_and_daily.py from the command line to create hourly and daily data.
- fix for #745. Now correlation estimates for instrument returns are not shrunk, except when doing DO. Note that this will affect risk calculations and risk overlays.
- Changed storage of capital IMPORTANT YOU MUST RUN sysdata/production/TEMP_capital_transfer.py from the command line to transfer your old capital
- "Market movers" report added
- Nicer functionality to get dates into reports interactively
- Added volume limits to auto fill position limits
- Fixed cost calculation for dynamic optimisation (see discussion)
- Private config now always pulled in, even for backtests (see discussion)
- Improved price cleaning (see discussion)
- Started introducing intraday strategy support
- Roll code will now forward fill prices if requested instead of failing
- Report to remove suggested instruments now 'buffers' to avoid too many changes
- A lot of improvements to reporting; new reports, split up existing reports
- Clearer exposition of IB vs my multipliers
- Skipped a few versions to tag it, why not
- Added back risk overlay code
- Added code to get margin
- Post processing takes holding costs into account
- Add buffering into sub systems
- Deal with zero prices and volumes
- Better documentation of instrument handling
- Customised ability to set instrument lists in sim
- Requires python 3.7 to 3.9
- Tweaks to dynamic optimisation
- Dashboard improvements
- Added more flexible instrument exclusion, recommended lists of instruments
- Refactored reporting code so uses common API
- Suggestions for updating slippage cost estimates
- Added buffering to dynamic optimisation code
- Added dashboard (thanks James!)
- Aligned production and backtesting risk calculation code
- More refactoring of production risk code
- Simplified risk report
- Added dynamic optimisation code (thanks Doug!)
- Added leverage constraint to autopopulate limits and trades
- Started reorganising production risk code
- Removed FuturesRawData, just use RawData now
- SR costs now include holding costs
- Can now
allocate_zero_instrument_weights_to_these_instrumentsin config file. - Moved roll config to /data/futures/csvconfig/ directory
- Added intraday bid/ask sampling to order stack handler
- Added liquidity and cost reports
- Can now 'ignore_instruments' in config file.
- improved optimisation code
- bug fixes particularly around roll calendars
- IB positions now summed when returned (IB started returning multiple entries for same contract)
- Added new configuration 'forecast_post_ceiling_cost_SR' which allows us to remove expensive rules after optimisation (so pooling everything works)
- Few bugs in optimisation code
- Added code to get futures data from IB initially, rather than third party
- Speeded up backtesting code after profiling
- Minor bugs
- Added smaller contract sizes to IB config
This update will switch to trading smaller contracts in several instruments. This allows more granular positions to be taken. These smaller contracts still pass my standard tests for volume and trading costs. It won't modify the database history of previously held positions.
If you want to use the small contracts purely for backtesting, and you are using the shipped .csv files for data, you only need to:
- Modify your backtest .yaml file(s). Delete the large instruments and replace with the following new instruments: CRUDE_W_mini, GAS_US_mini, GOLD_micro, KOSPI_mini, NASDAQ_micro, SP500_micro
If you want to use the smaller contracts in production:
- Stop all running processes if any
- The trade stacks should all be clear once you have done this
- Backup all your data
- Pull and install the update
- Run the scripts instruments_csv_mongo.py and roll_parameters_csv_mongo.py to copy the new roll config and instrument config to the database (It will warn you it's deleting your existing config, but unless you have something different from the contents of the shipped .csv files that shouldn't be an issue)
- Run the script clone_large_to_small_contracts.py to create a clone of the individual futures prices, roll calendars, multiple prices, and adjusted prices (clones large contract files to become smaller contract files).
- Modify your production backtest .yaml file(s) so the instrument weight on the larger contracts is zero, and move the instrument weights to the following new instruments: CRUDE_W_mini, GAS_US_mini, GOLD_micro, KOSPI_mini, NASDAQ_micro, SP500_micro. Copy and paste any other relevant parameters, eg position thresholding, forecast diversification multiplier, forecast weights .... Do not delete the parameters for the larger contracts for now.
- If you are using position thresholding, consider removing or modifying the parameters to reflect the smaller contracts.
- Run update_sampled_contracts, interactive_manual_check_historical_prices (for the relevant instruments), update_multiple_adjusted_prices, update_system_backtests, update_strategy_orders; and check all is as expected
- Check that the instrument trade stack contains a series of orders closing your large contract positions (if any), and opening up new positions in the smaller contracts (if any) (there may be other orders if there have been price changes since you last ran your production backtest)
- Restart all running processes
- The system will trade out of one set of positions and into the other (this will happen naturally with the stack_handler process, or you can use interactive_order_stack to control the process manually)
- Once all the positions are closed, to tidy things up you can remove the old instruments entirely from your backtest .yaml files
Note that the large contract prices will continue to be updated (best to keep doing this, in case we decide to switch back), and we'll still have history of prices and positions for the large contracts (which is good, or stuff like P&L and trade reporting will break horribly).
- Fixed a few bugs
- Added Bitcoin data, as everyone wants it
- Is pretty stable so version 1.0 seems appropriate (also my birthday)
- Completely refactored accounting code.
- Completely refactored optimisation code.
- Moved 'quant' type functions to sysquant: robust_vol_calc - you may get error messages - update your config
- Renamed certain functions in systems.rawdata (used by 'fancier' trading rules) - update your config
**WARNING! FROM VERSION 0.85.0 IS A MAJOR UPGRADE. SEE pandas_upgrade BEFORE DOING ANYTHING!
- Upgraded pandas and arctic versions
- Added 'Pause' process status
- Refactored and tidied control code
- Control process now sleep when not needed to save energy
- Removed option to specify machine to run process on
- Finished refactoring of production code (or at least, for now!)
- sysproduction/diagnostics now renamed reporting
- sysproduction/data code now uses generic handler and property methods to access data
- broker API now has proper base classes
- Tinkered with requirements to get running on new machine
- Moved defaults.yaml to /sysdata/config directory
- Removed 'example' strategy from config files - strategies need to be explicit in private yaml config
- Cleaned up configuration code. Production config now accessed through data blob where possible.
- Messed up order database by changing formats; let me know if you have any issues reading your old orders
- Massive refactoring mainly of order code but also IB client structure. Should be backwardly compatible with old saved orders except 'split' orders which are ignored. Read 'journey of an order' in production code for granular detail.
- added remote monitoring
- Split out control configuration from other YAML files (YOU WILL NEED TO CHANGE PRIVATE CONFIG look at the production docs!)
- Refactoring of run and control processes mostly into new syscontrol module
- Added simple monitoring tool
- Added email 'to' option (YOU WILL NEED TO CHANGE PRIVATE CONFIG TO INCLUDE email_to parameter)
- Mostly refactoring and documenting the creation and storage of data
- Essentially 'close' production.md (in as much as anything can be close...)
- Changed data Blobs so now take lists of objects rather than str, easier to see dependencies
- (Done loads of work but forgotten to update the version number or this file. So let's reward ourselves with a 0.20 version bump. The following list is almost certainly incomplete...)
- Done loads of documentation for production
- Added position limits
- Removed broker base classes as redundant
- Minimum tick size used in setting limit orders
- Stopped double counting of volumes when daily/intraday data mixed
- Added startup script
- Fixed issues with time zone mismatches
- Added trades report, strategy report, signals report, risk report, p&l report
- Position locks
- Added interactive diagnostics, interactive order stack, interactive controls
- Execution algo!
- Added capability to trade
- Introduced capital model for production
- Fixed bug in implementation of correlation to covariance
- Added optional code for risk overlay see blog
- Moved fx cross logic out of sim data into fxPricesData
- Strategies now run backtests from configuration file
- Added price 'spike' checker, and manual price checking service
- Removed PIL library (issue 161)
- Fixed ib_insync PIP issue (pull 162)
- MongoDb logs will now try to email user if a critical error is raised
- IB now uses ib_insync, not native IB library
- Cleaned up way defaults and private config files work
- Removed separate mongodb config file
- Added production code to run a system backtest and save optimal position state
- Cleaned up the way path and filename resolution works
- Added production code to backup mongodb to .csv files
- Added production code to get daily futures prices from IB, update sampled contracts, update multiple and adjusted prices.
- Can now get individual futures prices from IB, both historical daily and intraday (with get_prices_at_frequency_for_* methods)
- Added code to deal with VIX weekly expiries - they can now be ignored
- Caching IB contract objects in IB client as rather expensive
- IB client will now avoid pacing violations
- Removed futuresContract.simple() method; you can now just do futuresContract("AUDUSD", "yyyymmdd")
- Cleaned up IB client code, error handling is now consistent
- Added broker_get_contract_expiry_date method to brokerClient and ibClient
- IB connection will now check to see if a clientid is being used even if one is passed. Has '.terminate' method which will try and clear clientid.
- .csv config files are cached in IB price API objects
- Can now pass keyword arguments to data methods when creating a trading rule (Enhancement # 141)
- Fixed bugs relating to building multiple and adjusted prices
- Slight refactoring of futuresContractPrices objects. These only have FINAL, not CLOSE or SETTLE prices now.
- Added more data
- 'resolve_path_and_filename_for_package' can now take absolute as well as relative paths, and can cope with separate file names
- Updated legacy .csv files
- Fixed a few bugs
- Can now get unexpired contracts for a given instrument using 'contractDateWithRollParameters.get_unexpired_contracts_from_now_to_contract_date()'
Now requires python 3.6.0, pandas 0.25.2
- Fixed a few bugs in production functions for FX prices
- Logging now requires an explicit labelling argument, eg `log=logtoscreen("String required here")
- Changed mongodb logging so now indexes on unique ID
- Generally cleaned up logging code
- Moved update fx price logic inside generic fx price object
- Removed dependency on Quandl currency for setting up spot FX, now uses investing.com
- Fixed issues relating to robust vol calc, date offset in roll calendars
- Started documenting 'how to run a production system'
- Created logging to mongo database
- Refactoring of mongo and arctic connections
- Started creating crontab and scripts for various production functions (read and write FX prices)
- Added code to ensure unique client ID for IB
- Added connection code for Interactive Brokers. See connecting pysystemtrade to interactive brokers for more details.
- Implemented data socket for spot FX, getting data from IB
- Added handcrafting optimisation code.
- Added methods to read weight data from csv files
- Put generalised non linear mapping into forecast combination
- Added flag option to use process pools for parallel processing - but not actually used yet
- Cleaned up setup.py file now finds data files recursively
- Fixed bug in getting asset class data from csv config files
- Finished populating configuration files for Quandl and roll configuration.
- Debugged futures.md documentation file.
- See futures documentation for more details.
- New data sources: Quandl. Data storage in mongodb and arctic is now supported.
- Back-adjustment is possible and can be done 'on the fly' or from scratch with new data.
- Further refactoring of sim data objects to support the above.
Massive refactoring of sim data objects, to support alternative data sources and backadjusting
Created classses for individual futures contracts, and included example of how to use Quandl to get them
Updated .csv data and moved to separate section - now stored under Github LFS
Added quandl data (but only for individual futures contracts right now so useless)
Removed uses of old carry function which was deprecated
- Fixed incorrect calculation of returns over weekends
- Forecast scalars now only pool across the set of instruments using a given trading rule
- Changed error handling for empty Rules() objects
- Added TOC to userguide.md file
- Updated to pandas 0.22.0
- Fixed issue #64, #68, #70 and other issues relating to pandas API update breaking correlation matrices
Moved most examples except core to separate git here
- Now supports pandas 0.20.3 (earlier pandas will break)
- Added progress bar (issue 51)
- Stages now have _names and _description defined in init
- log values now passed in when init of stage; hence baseystem.init is much cleaner
- Caching:
- Cache is now accessed via a separate object in system; so system.cache.* rather than system.* for cache methods
- Caching now done through decorators: from systems.system_cache import input, dont_cache, diagnostic, output
- Use protected=True and/or not_cached=True within decorators
- Got rid of 'switching' stages for estimating forecast scalars, forecast weights, instrument weights.
- Explicit import of a Fixed or Estimated version of a class won't work; use the generic version.
- Added separate fields to .yaml file to switch between IDM and FDM estimation or fixed values
- Split ultra-massive accounts.py into multiple files and classes
- Split unwieldy ForecastCombine into several classes
- Added a bunch more unit tests as I went through the above refactoring exercise
- some refactoring of optimisation code - more to come
- fixed up examples and documentation accordingly
- Now requires pandas version > 0.19.0
- Capital correction now works. New methods: system.accounts.capital_multiplier, system.accounts.portfolio_with_multiplier, system.portfolio.get_actual_positon, system.portfolio.get_actual_buffers_with_position, system.accounts.get_buffered_position_with_multiplier. See this blog post and the guide
- Smooth fixed weights as well as variable: removed ewma_span and moved to new config item forecast_weight_ewma_span and same for instruments. Removed override of get_instrument_weights, get_forecast_weights method from estimated classes.
- Added extra methods to support capital scaling, but not implemented yet.
- fixed couple of bugs in getting subsystem p&l to calculate instrument weights
- removed aligned fx method, doesn't speed up and adds complexity
- solved issue #16
- Included option to show account curves as cumulative (compounding): somecurve.cumulative()
- removed percentage options, now a method for account curves: somecurve.percent()
- Incorporated capital into account curves: anycurve.capital
- General clean up of the way capital dealt with in accounting
- More speed up, couple of tweaks...
- Split up optimiser class so can selectively check if need data for equal weights; speed up
- Fixed bugs introduced in last version
- Refactored optimisation with costs code, changed configuration slightly (read this revised blog post for more )
- Introduced method to cope with pooling on both costs and gross returns, so doesn't recalculate several times
- Moved pre-screening for expensive assets to an earlier stage
- New optimisation method "equal_weights" for equal weights; means that eg expensive forecasts can be removed and then take an equal weight on the rest
- Optimisation:
- Replaced slow divide, multiply methods in syscore.pdutils with straightforward division; also means:
- Replaced Tx1 pd.DataFrames with pd.Series except where stricly necessary
- Removed a lot of defensive reindexing code where things should already be on same timestamp
- Replaced remaining reindexing code with pandas native .align methods
- accounting p&l doesn't have to work out trades, then go back to positions, if no trades provided.
- Replaced slow divide, multiply methods in syscore.pdutils with straightforward division; also means:
- Changed / added the following methods to
system.accounts:pandl_for_instrument_forecast_weighted,pandl_for_trading_rule_weighted,pandl_for_all_trading_rules,pandl_for_trading_rule,pandl_for_trading_rule_unweighted,pandl_for_all_trading_rules_unweightedSee Weighted and unweighted account curve groups for more detail. - Added
get_capital_in_rule,get_instrument_forecast_scaling_factorto help calculate these. - fixed error in user guide
- Fixed small bug with shrinkage
- Added references to blog post on costs
- introduced methods for optimisation with costs. See this blog post for more
- made a lot of tweaks to optimisation code; mainly shrinkage now shrinks towards target Sharpe ratio, equalising SR does the same; consistent annualisation
- introduced new parameter for optimisation
ann_target_SR system.combForecast.calculation_of_raw_forecast_weights(estimated version) no longer stores nested weights.
- ability to pickle and unpickle cache (
system.pickle_cache,system.unpickle_cache) - included breakout rule (example is being written)
- separate out weighting calculation so instrument forecast pandl can be cached
- csv data is now daily and updated to present day
- Fixed bug with loading data from private module
- Changed raw cost data so returns dict not tuple
- Added 'flags' to cache identifier to replace horrors like 'portfolio__percentageTdelayfillTroundpositionsT'
- p&l for trading rules now nested in caches rather than using special identifier
- Added method
accounts.pandl_for_instrument_rules
- Renamed method
accounts.pandl_for_instrument_rulestopandl_for_instrument_rules.unweighted - Fixed bug with portfolio and instrument account curves overstating costs by adding cost weightings
- Fixed weighting of account curves and introduced explicit flag for weighting
- Added
pandl_for_trading_rule_unweightedmethod to accounts object.
- Added
pandl_for_trading_rulemethod to accounts object.
- Added t_test method to
accountCurve(and all that inherit from her)
- Added methods to accountCurveGroup.get_stats(): .mean(), .std(), .tstat(), .pvalue()
- Added method to accountCurveGroup stack; stack object can also produce bootstrap
- Added account_t_test(ac1, ac2) to produce a t-test statistic for any two account curve like objects.
- dynamically change class depending on config flag to estimate parameters or not
- add stage description field, and stage.methods() method
- add stage name to cache reference, always pass stage to caching function. Added cache methods to system which understand stages
- Correlation tests failing - fixed up
- Costs SR didn't get turnover - duh! Now fixed. Added a bunch of input methods to accounts object to calculate them
- tweak to account curve grouping to data frame to remove nans
- cost calculation no longer fails if no trades for an instrument
- changed buffering rounding so consistent with my own system
- Introduced maximum cap on IDM and FDM of 2.5, as per the book.
- Correlation cleaning wasn't working as documented - now does.
- cleaning up:
- renamed misleading 'get_instrument_prices_for_position_or_forecast' method to 'get_raw_price', to fix some tests that hadn't realised the difference
- fixed a bunch of tests
- changed the way cross rates are calculated to ensure data isn't lost
- Include buffering / position intertia
- Included cost data and calculations.
- Account curve improvements: generate lists of simulated trades, extend the
accountCurveobject to handle multiple columns, statistics over different periods.
- Fixed bug with bootstrapping with missing values
- Changed clean correlations so it replaces with an average
- Fixed some documentation SNAFU's
- Added reference to latest blog post
- Calculating forecast weights in ForecastCombineEstimated
- Created PortfoliosEstimated
- Calculating instrument weights
- Calculating instrument diversification multiplier
- Added a logging function
- Modified system.get_instruments so will check config.instruments (useful if estimating instrument weights)
- Included daily_prices method in data; raw data method just points to it; replaced most uses of (intraday) data.get_instrument_price with daily prices
- Added some new methods to account stage
- Cleaned up the way pooling works in correlation estimation
- Finished clean_correlation function so now deals with incomplete matricies
- Changed the way defaults feed into config objects
- Added estimation of forecast diversification multiplier to ForecastCombineEstimated
- Changed default forecast correlation estimation period; had to fix up some test output
- Changed way that forecast correlations are cached
- Started using more logical version numbering scheme :-)
- Created ForecastCombineEstimated, with get_forecast_correlation_matrices
- Added get_trading_rule_list and get_all_forecasts to forecast_combine
- Added rule_variations config option
- Added Bund data to test suite; had to fix some tests
- Pooling for forecast scalar doesn't need its own function anymore
- Changed the way config defaults are handled
- Fixed bugs: use of bool to convert str
- Fixed bugs: some test configs had wrong trading rule parameter setup; had to fix slew of tests as a result
- Added rolling estimate of forecast scalars; try
System([rawdata, rules, ForecastScaleCapEstimated()], data, config) - Moved .get_instrument_list from portfolio object to parent system
- Basic backtesting environment with example futures data.