Release Notes¶
Version 0.5.7¶
Release Date: 2018-03-29
Build¶
- Data Marketplace deployed on mainnet.
- Added progress indicators for publishing data, and made the data publishing synchronous to provide feedback to the publisher.
Version 0.5.6¶
Release Date: 2018-03-22
Build¶
- Data Marketplace: ensures compatibility across wallets, now fully supporting
ledger,trezor,keystore,private key. Partial support formetamask(includes sign_msg, but not sign_tx). Current support forDigital Bitboxis unknown, but believed to be supported. - Data Marketplace: Switched online provider from MyEtherWallet to MyCrypto.
- Data Marketplace: Added progress indicator for data ingestion.
Version 0.5.5¶
Release Date: 2018-03-19
Version 0.5.4¶
Release Date: 2018-03-14
Build¶
- Switched Data Marketplace from Ropstein testnet to Rinkeby testnet after incorporating changes resulting from the marketplace contract audit
- Several usability improvements of the Data Marketplace that make the –dataset parameter optional. If it is not included in the command line, will list available datasets, and let you choose interactively.
Bug Fixes¶
- Fix Binance requirement of symbol to be included in the cancelled order #204
- Fix notenoughcasherror when an open order is filled minutes later #237
- Properly handle of empty candles received from exchanges #236
- Added a function to reduce open orders amount from calculated target/amount for target orders #243
- Fix missing file in live trading mode on date change #252, #253
- Upgraded Data Marketplace to Web3==4.0.0b11, which was breaking some functionality from prior version 4.0.0b7 #257
- Always request more data to avoid empty bars and always give the exact bar number #260
Version 0.5.3¶
Release Date: 2018-02-09
Version 0.5.2¶
Release Date: 2018-02-08
Version 0.4.7¶
Release Date: 2018-01-19
Version 0.4.6¶
Release Date: 2018-01-18
Bug Fixes¶
- Fixed some Python3 issues
- Reading the trade log to get executed order prices on exchanges like Binance (#151)
- Fixed issue with market order executing price (#150 and #111)
- Implemented standardized symbol mapping (#157)
- Improved error handling for unsupported timeframes (#159)
- Using Bitfinex instead of Poloniex to fetch btc_usdt benchmark (#161)
Build¶
- Added a context.state dict to keep arbitrary state values between runs
- Added ability to stop live algo at specified end date
Version 0.4.5¶
Release Date: 2018-01-12
Version 0.4.4¶
Release Date: 2018-01-09
Version 0.4.0¶
Release Date: 2017-12-12
Bug Fixes¶
Build¶
Version 0.3.10¶
Release Date: 2017-11-28
Bug Fixes¶
- Fixed issue with fetching assets with daily frequency
Version 0.3.8¶
Release Date: 2017-11-14
Bug Fixes¶
- Fixed a warning filter issue introduced with the latest release
Version 0.3.6¶
Release Date: 2017-11-4
Version 0.3.5¶
Release Date: 2017-11-4
Version 0.3.4¶
Release Date: 2017-11-2
Bug Fixes¶
- Fixed issue with auto-ingestion of minute data (#47)
- Fixed issue with sell orders in backtesting
- Fixed data frequency issues with data.history() in backtesting
- Fixed an issue with can_trade()
- Reduced the commission and slippage values to account for lower volume transactions
Build¶
- Added more unit tests
Documentation¶
- Improved installation notes for Windows C++ compiler and Conda
- Addition of Jupyter Notebook guide
- Addition of Live Trading page
- Addition of Videos page
- Addition of Resources page
- Addition of Development Guidelines
- Addition of Release Notes
- Updated code docstrings
Version 0.3.3¶
Release Date: 2017-10-26
Bug Fixes¶
- Fix missing -x in ingest-exchange
- Fix issue with daily chunks end date (data bundles)
- Fix issue in the prepare_chunk logic (data bundles)
Build¶
- Added data validation unit tests
Version 0.3.2¶
Release Date: 2017-10-25
Bug Fixes¶
- Fix to work with empty data bundles
- Fix Windows path of
$HOME/.catalystfolder - Fix
etc/python2.7-environment.ymlfor Windows Conda install - Fix hash method to create sid numbers compatible across platforms
- Fix an issue with asset date in chunks
Build¶
- Python3 adjustments
- Added method to clean bundle folders, and remove symbols.json
- Implemented and improved unit tests
Version 0.3.1¶
Release Date: 2017-10-22
Bug Fixes¶
- Fixed OS-dependent path issue in data bundle
- Changed handling of empty
auth.json, instead of throwing an error for missing file - Updated
etc/python2.7-environment.ymlto work with Catalyst version 0.3 - Updated
catalyst/examples/buy_and_hodl.pyandcatalyst/examples/buy_low_sell_high.pyto work with Catalyst version 0.3
Version 0.3¶
Release Date: 2017-10-20
- Standardized live and backtesting syntax
- Added a repository for historical data
- Added supported for multiple exchanges per algorithm
- Added a standardized dictionary of symbols for each exchange
- Added auto-ingestion of bundle data while backtesting
- Bug fixes
Version 0.2.dev5¶
Release Date: 2017-10-03
- Fixes bug in data.history function that was formatting ‘volume’ data as integers, now they are returned as floats with up to 9 decimals of precision. Data bundles redone.
Version 0.2.dev4¶
Release Date: 2017-09-20
- Fixes bug in the pricing resolution of 1-minute data, now set to 8 decimal places. Pricing resolution of daily data remains set to 9 decimal places.
- The current data bundle takes 340MB compressed for download, and 460MB uncompressed on disk for Catalyst to use.
Version 0.2.dev3¶
Release Date: 2017-09-20
- 1-minute resolution OHLCV data bundle for backtesting from Poloniex exchange
- Implementation of trading of fractional crypto assets (i.e. 0.01 BTC)
- Minimum trade size of a coin can be configured on a per-coin basis, defaults to 0.00000001 in backtesting (most exchanges set the minimum trade to larger amounts, which will impact live trading)
- Increased pricing resolution from 3 to 9 decimal places
- The current data bundle takes 40MB compressed for download, and 99MB uncompressed on disk for Catalyst to use.
Version 0.2.dev1¶
Release Date: 2017-09-03
- Implementation of live trading:
- Comprehensive trading functionality against exchanges Bitfinex and Bittrex.
- Support for all trading pairs available on each exchange.
- Multiple algorithms can trade simultaneously against a single exchange using the same account.
- Each algorithm has a persisted state (i.e. algorithm can be stopped and restarted preserving the state without data loss) that tracks all open orders, executed transactions and portfolio positions.
- Minute by minute portfolio performance metrics.
- Daily summary performance statistics compatible with pyfolio, a Python library for performance and risk analysis of financial portfolios
Version 0.1.dev9¶
Release Date: 2017-08-28
- Retrieval of crypto benchmark from bundle, instead of hitting Poloniex exchange directly
- Change of bundle storage provider from Dropbox to AWS
- Fix issue with 1/1000 scaling issue of prices in bundle