133 lines
4.3 KiB
Python
133 lines
4.3 KiB
Python
# pragma pylint: disable=missing-docstring
|
|
|
|
import logging
|
|
import json
|
|
import os
|
|
from typing import Optional, List, Dict
|
|
from pandas import DataFrame
|
|
from freqtrade.exchange import get_ticker_history
|
|
from freqtrade.optimize.hyperopt_conf import hyperopt_optimize_conf
|
|
from freqtrade.analyze import populate_indicators, parse_ticker_dataframe
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
def load_tickerdata_file(pair, ticker_interval):
|
|
"""
|
|
Load a pair from file,
|
|
:return dict OR empty if unsuccesful
|
|
"""
|
|
path = testdata_path()
|
|
file = '{abspath}/{pair}-{ticker_interval}.json'.format(
|
|
abspath=path,
|
|
pair=pair,
|
|
ticker_interval=ticker_interval,
|
|
)
|
|
# The file does not exist we download it
|
|
if not os.path.isfile(file):
|
|
return None
|
|
|
|
# Read the file, load the json
|
|
with open(file) as tickerdata:
|
|
pairdata = json.load(tickerdata)
|
|
return pairdata
|
|
|
|
|
|
def load_data(ticker_interval: int = 5, pairs: Optional[List[str]] = None,
|
|
refresh_pairs: Optional[bool] = False) -> Dict[str, List]:
|
|
"""
|
|
Loads ticker history data for the given parameters
|
|
:param ticker_interval: ticker interval in minutes
|
|
:param pairs: list of pairs
|
|
:return: dict
|
|
"""
|
|
result = {}
|
|
|
|
_pairs = pairs or hyperopt_optimize_conf()['exchange']['pair_whitelist']
|
|
|
|
# If the user force the refresh of pairs
|
|
if refresh_pairs:
|
|
logger.info('Download data for all pairs and store them in freqtrade/tests/testsdata')
|
|
download_pairs(_pairs)
|
|
|
|
for pair in _pairs:
|
|
pairdata = load_tickerdata_file(pair, ticker_interval)
|
|
if not pairdata:
|
|
# download the tickerdata from exchange
|
|
download_backtesting_testdata(pair=pair, interval=ticker_interval)
|
|
# and retry reading the pair
|
|
pairdata = load_tickerdata_file(pair, ticker_interval)
|
|
result[pair] = pairdata
|
|
return result
|
|
|
|
|
|
def preprocess(tickerdata: Dict[str, List]) -> Dict[str, DataFrame]:
|
|
"""Creates a dataframe and populates indicators for given ticker data"""
|
|
return {pair: populate_indicators(parse_ticker_dataframe(pair_data))
|
|
for pair, pair_data in tickerdata.items()}
|
|
|
|
|
|
def testdata_path() -> str:
|
|
"""Return the path where testdata files are stored"""
|
|
return os.path.abspath(os.path.join(os.path.dirname(__file__), '..', 'tests', 'testdata'))
|
|
|
|
|
|
def download_pairs(pairs: List[str]) -> bool:
|
|
"""For each pairs passed in parameters, download 1 and 5 ticker intervals"""
|
|
for pair in pairs:
|
|
try:
|
|
for interval in [1, 5]:
|
|
download_backtesting_testdata(pair=pair, interval=interval)
|
|
except BaseException:
|
|
logger.info('Failed to download the pair: "{pair}", Interval: {interval} min'.format(
|
|
pair=pair,
|
|
interval=interval,
|
|
))
|
|
return False
|
|
return True
|
|
|
|
|
|
def download_backtesting_testdata(pair: str, interval: int = 5) -> bool:
|
|
"""
|
|
Download the latest 1 and 5 ticker intervals from Bittrex for the pairs passed in parameters
|
|
Based on @Rybolov work: https://github.com/rybolov/freqtrade-data
|
|
:param pairs: list of pairs to download
|
|
:return: bool
|
|
"""
|
|
|
|
path = testdata_path()
|
|
logger.info('Download the pair: "{pair}", Interval: {interval} min'.format(
|
|
pair=pair,
|
|
interval=interval,
|
|
))
|
|
|
|
filepair = pair.replace("-", "_")
|
|
filename = os.path.join(path, '{pair}-{interval}.json'.format(
|
|
pair=filepair,
|
|
interval=interval,
|
|
))
|
|
filename = filename.replace('USDT_BTC', 'BTC_FAKEBULL')
|
|
|
|
if os.path.isfile(filename):
|
|
with open(filename, "rt") as fp:
|
|
data = json.load(fp)
|
|
logger.debug("Current Start: {}".format(data[1]['T']))
|
|
logger.debug("Current End: {}".format(data[-1:][0]['T']))
|
|
else:
|
|
data = []
|
|
logger.debug("Current Start: None")
|
|
logger.debug("Current End: None")
|
|
|
|
new_data = get_ticker_history(pair=pair, tick_interval=int(interval))
|
|
for row in new_data:
|
|
if row not in data:
|
|
data.append(row)
|
|
logger.debug("New Start: {}".format(data[1]['T']))
|
|
logger.debug("New End: {}".format(data[-1:][0]['T']))
|
|
data = sorted(data, key=lambda data: data['T'])
|
|
|
|
with open(filename, "wt") as fp:
|
|
json.dump(data, fp)
|
|
|
|
return True
|