You need to install the legacy version of the jesse.
it will install modified version of legacy jesse cli and jesse-tk. You still need to create jesse_db manually as stated in original jesse installation document.
bash <(curl -s https://raw.githubusercontent.com/ysdede/stack-installer/master/jesse-cli-ubuntu20.sh)
Download and install prebuilt ta-lib from https://www.lfd.uci.edu/~gohlke/pythonlibs/
pip install downloaded wheel file
install legacy jesse cli from https://github.com/ysdede/jesse
Clone this repository and install with pip.
pip install .
or
pip install -e git+https://github.com/ysdede/jesse-tk.git#egg=jesse-tk
CLI help explains it:
Usage: jesse-tk import-routes [OPTIONS] [START_DATE]
Import-candles for pairs listed in routes.py Enter start date "YYYY-MM-DD"
If no start date is specified, the system will default to two days earlier.
It's useful for executing script in a cron job to download deltas on a
regular basis.
It's name is confusing, I need to change it. bulk
helps to download candle data faster. It downloads zipped kline csv files from Binance Vision as monthly and daily packets. One disadvantage of it is packed csv files (rarely) has missing data points. Running jesse import-candles
helps to fill gaps.
Usage: jesse-tk bulk [OPTIONS] EXCHANGE SYMBOL START_DATE
Bulk download Binance candles Enter EXCHANGE SYMBOL START_DATE { Optional: --workers n}
jesse-tk bulk Binance btc-usdt 2020-01-01
jesse-tk bulk 'Binance Futures' btc-usdt 2020-01-01
You can use spot or futures keywords instead of full exchange name.
jesse-tk bulk spot btc-usdt 2020-01-01
jesse-tk bulk futures eth-usdt 2017-05-01 --workers 8
Options:
--workers INTEGER The number of workers to run simultaneously. You can use
cpu thread count or x2. [default: 4]
Downloading ADAUSDT-1m-2022-02-05.csv
Candles already exits in DB skipping 1440 datapoints, /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-01.csv
Downloading ADAUSDT-1m-2022-02-06.csv
OK 1440 /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-01.csv
DEBUG: self.exchange Binance Futures, self.symbol ADA-USDT
DEBUG: 1644019200000, 1644105540000, query result: 0, datapoints: 1440
Saving to db: /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-05.csv Size: 152170 bytes, 1440 datapoints. time passed: 37 seconds.
DEBUG: self.exchange Binance Futures, self.symbol ADA-USDT
DEBUG: self.exchange Binance Futures, self.symbol ADA-USDT
DEBUG: 1643932800000, 1644019140000, query result: 1395, datapoints: 1440
Saving to db: /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-04.csv Size: 151526 bytes, 1440 datapoints. time passed: 38 seconds.
Candles already exits in DB skipping 1440 datapoints, /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-03.csv
OK 1440 /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-03.csv
DEBUG: self.exchange Binance Futures, self.symbol ADA-USDT
Candles already exits in DB skipping 1440 datapoints, /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-06.csv
OK 1440 /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-06.csv
OK 1440 /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-04.csv
OK 1440 /tmp/bulkdata/futures/um/daily/klines/ADAUSDT/1m/ADAUSDT-1m-2022-02-05.csv
Completed in 41 seconds.
Usage: jesse-tk bulkpairs [OPTIONS] EXCHANGE START_DATE
Bulk download ALL! Binance Futures candles to Jesse DB. Enter EXCHANGE START_DATE {Optional: --workers n}
jesse-tk bulkpairs 'Binance Futures' 2020-01-01 jesse-tk bulkpairs futures 2017-05-01 -- workers 8
Options:
--workers INTEGER The number of workers to run simultaneously. [default: 2]
--all / --list Get pairs list from api or pairs from file.
- --all / --list: Default is list. Get candle list from user defined
pair.py
file.
If you add--all
it will get pair list from exchange api. Use it with caution. It will take a long time and fill up your disk.
jesse-tk bulkpairs futures 2021-01-01 --workers 4 --all
There's 140 available pairs in Binance Futures:
['BTC-USDT', 'ETH-USDT', 'BCH-USDT', 'XRP-USDT', 'EOS-USDT', 'LTC-USDT', 'TRX-USDT', 'ETC-USDT', 'LINK-USDT', 'XLM-USDT', 'ADA-USDT', 'XMR-USDT', 'DASH-USDT', 'ZEC-USDT', 'XTZ-USDT', 'BNB-USDT', 'ATOM-USDT', 'ONT-USDT', 'IOTA-USDT', 'BAT-USDT', 'VET-USDT', 'NEO-USDT', 'QTUM-USDT', 'IOST-USDT', 'THETA-USDT', 'ALGO-USDT', 'ZIL-USDT', 'KNC-USDT', 'ZRX-USDT', 'COMP-USDT', 'OMG-USDT', 'DOGE-USDT', 'SXP-USDT', 'KAVA-USDT', 'BAND-USDT', 'RLC-USDT', 'WAVES-USDT', 'MKR-USDT', 'SNX-USDT', 'DOT-USDT', 'DEFI-USDT', 'YFI-USDT', 'BAL-USDT', 'CRV-USDT', 'TRB-USDT', 'YFII-USDT', 'RUNE-USDT', 'SUSHI-USDT', 'SRM-USDT', 'EGLD-USDT', 'SOL-USDT', 'ICX-USDT', 'STORJ-USDT', 'BLZ-USDT', 'UNI-USDT', 'AVAX-USDT', 'FTM-USDT', 'HNT-USDT', 'ENJ-USDT', 'FLM-USDT', 'TOMO-USDT', 'REN-USDT', 'KSM-USDT', 'NEAR-USDT', 'AAVE-USDT', 'FIL-USDT', 'RSR-USDT', 'LRC-USDT', 'MATIC-USDT', 'OCEAN-USDT', 'CVC-USDT', 'BEL-USDT', 'CTK-USDT', 'AXS-USDT', 'ALPHA-USDT', 'ZEN-USDT', 'SKL-USDT', 'GRT-USDT', '1INCH-USDT', 'AKRO-USDT', 'CHZ-USDT', 'SAND-USDT', 'ANKR-USDT', 'LUNA-USDT', 'BTS-USDT', 'LIT-USDT', 'UNFI-USDT', 'DODO-USDT', 'REEF-USDT', 'RVN-USDT', 'SFP-USDT', 'XEM-USDT', 'BTCST-USDT', 'COTI-USDT', 'CHR-USDT', 'MANA-USDT', 'ALICE-USDT', 'HBAR-USDT', 'ONE-USDT', 'LINA-USDT', 'STMX-USDT', 'DENT-USDT', 'CELR-USDT', 'HOT-USDT', 'MTL-USDT', 'OGN-USDT', 'NKN-USDT', 'SC-USDT', 'DGB-USDT', '1000SHIB-USDT', 'ICP-USDT', 'BAKE-USDT', 'GTC-USDT', 'BTCDOM-USDT', 'KEEP-USDT', 'TLM-USDT', 'IOTX-USDT', 'AUDIO-USDT', 'RAY-USDT', 'C98-USDT', 'MASK-USDT', 'ATA-USDT', 'DYDX-USDT', '1000XEC-USDT', 'GALA-USDT', 'CELO-USDT', 'AR-USDT', 'KLAY-USDT', 'ARPA-USDT', 'NU-USDT', 'CTSI-USDT', 'LPT-USDT', 'ENS-USDT', 'BTC-USDT', 'ETH-USDT', 'PEOPLE-USDT', 'ANT-USDT', 'ROSE-USDT', 'DUSK-USDT', '1000BTTC-USDT']
There's 80 available pairs in candle database at: 2021-01-01
Start: 2021-01-01 00:00:00 2022-01-01T00:00:00+00:00
temp_dir /tmp
Works same as classic jesse backtest
but it has ability to inject hyperparameters from command line. Useful for batch testing, research.
--dna TEXT Base32 encoded dna string payload [default:None]
--hp TEXT Hyperparameters payload as dict [default: None]
--seq TEXT Fixed width hyperparameters payload [default: None]
- --dna: Jesse encoded dna strings. Supports only integer hyperparameters. eg.
A0@1J
- -- hp: Hyperparameters as json/dict. eg:
{'name': 'boost', 'type': int, 'min': 0, 'max': 150, 'default': 68}
My favourite:
- --seq: Fixed width, zero padded hyperparameters. Last number identifies the width. eg. 730900202
can be encoded as: 73 9 0 20 [2 is width] Jesse-tk tools like random, refine uses seq as default input.
If you pass seq code and--full-reports
at the same time tk prints seq code into quantstats reports to identify results. It has a special use case.
see old tool jesse-picker for details.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.