224 lines
7.4 KiB
Python
224 lines
7.4 KiB
Python
# pragma pylint: disable=missing-docstring
|
|
|
|
import gzip
|
|
import json
|
|
import logging
|
|
import os
|
|
from typing import Optional, List, Dict, Tuple, Any
|
|
import arrow
|
|
|
|
from freqtrade import misc, constants
|
|
from freqtrade.exchange import get_ticker_history
|
|
|
|
from user_data.hyperopt_conf import hyperopt_optimize_conf
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
def trim_tickerlist(tickerlist: List[Dict], timerange: Tuple[Tuple, int, int]) -> List[Dict]:
|
|
if not tickerlist:
|
|
return tickerlist
|
|
|
|
stype, start, stop = timerange
|
|
|
|
start_index = 0
|
|
stop_index = len(tickerlist)
|
|
|
|
if stype[0] == 'line':
|
|
stop_index = start
|
|
if stype[0] == 'index':
|
|
start_index = start
|
|
elif stype[0] == 'date':
|
|
while start_index < len(tickerlist) and tickerlist[start_index][0] < start * 1000:
|
|
start_index += 1
|
|
|
|
if stype[1] == 'line':
|
|
start_index = len(tickerlist) + stop
|
|
if stype[1] == 'index':
|
|
stop_index = stop
|
|
elif stype[1] == 'date':
|
|
while stop_index > 0 and tickerlist[stop_index-1][0] > stop * 1000:
|
|
stop_index -= 1
|
|
|
|
if start_index > stop_index:
|
|
raise ValueError(f'The timerange [{start},{stop}] is incorrect')
|
|
|
|
return tickerlist[start_index:stop_index]
|
|
|
|
|
|
def load_tickerdata_file(
|
|
datadir: str, pair: str,
|
|
ticker_interval: str,
|
|
timerange: Optional[Tuple[Tuple, int, int]] = None) -> Optional[List[Dict]]:
|
|
"""
|
|
Load a pair from file,
|
|
:return dict OR empty if unsuccesful
|
|
"""
|
|
path = make_testdata_path(datadir)
|
|
pair_file_string = pair.replace('/', '_')
|
|
file = os.path.join(path, '{pair}-{ticker_interval}.json'.format(
|
|
pair=pair_file_string,
|
|
ticker_interval=ticker_interval,
|
|
))
|
|
gzipfile = file + '.gz'
|
|
|
|
# If the file does not exist we download it when None is returned.
|
|
# If file exists, read the file, load the json
|
|
if os.path.isfile(gzipfile):
|
|
logger.debug('Loading ticker data from file %s', gzipfile)
|
|
with gzip.open(gzipfile) as tickerdata:
|
|
pairdata = json.load(tickerdata)
|
|
elif os.path.isfile(file):
|
|
logger.debug('Loading ticker data from file %s', file)
|
|
with open(file) as tickerdata:
|
|
pairdata = json.load(tickerdata)
|
|
else:
|
|
return None
|
|
|
|
if timerange:
|
|
pairdata = trim_tickerlist(pairdata, timerange)
|
|
return pairdata
|
|
|
|
|
|
def load_data(datadir: str,
|
|
ticker_interval: str,
|
|
pairs: Optional[List[str]] = None,
|
|
refresh_pairs: Optional[bool] = False,
|
|
timerange: Optional[Tuple[Tuple, int, int]] = None) -> Dict[str, List]:
|
|
"""
|
|
Loads ticker history data for the given parameters
|
|
: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 %s', datadir)
|
|
download_pairs(datadir, _pairs, ticker_interval, timerange=timerange)
|
|
|
|
for pair in _pairs:
|
|
pairdata = load_tickerdata_file(datadir, pair, ticker_interval, timerange=timerange)
|
|
if pairdata:
|
|
result[pair] = pairdata
|
|
else:
|
|
logger.warn('No data for pair %s, use --update-pairs-cached to download the data', pair)
|
|
|
|
return result
|
|
|
|
|
|
def make_testdata_path(datadir: str) -> str:
|
|
"""Return the path where testdata files are stored"""
|
|
return datadir or os.path.abspath(
|
|
os.path.join(
|
|
os.path.dirname(__file__), '..', 'tests', 'testdata'
|
|
)
|
|
)
|
|
|
|
|
|
def download_pairs(datadir, pairs: List[str],
|
|
ticker_interval: str,
|
|
timerange: Optional[Tuple[Tuple, int, int]] = None) -> bool:
|
|
"""For each pairs passed in parameters, download the ticker intervals"""
|
|
for pair in pairs:
|
|
try:
|
|
download_backtesting_testdata(datadir,
|
|
pair=pair,
|
|
tick_interval=ticker_interval,
|
|
timerange=timerange)
|
|
except BaseException:
|
|
logger.info(
|
|
'Failed to download the pair: "%s", Interval: %s',
|
|
pair,
|
|
ticker_interval
|
|
)
|
|
return False
|
|
return True
|
|
|
|
|
|
def load_cached_data_for_updating(filename: str,
|
|
tick_interval: str,
|
|
timerange: Optional[Tuple[Tuple, int, int]]) -> Tuple[
|
|
List[Any],
|
|
Optional[int]]:
|
|
"""
|
|
Load cached data and choose what part of the data should be updated
|
|
"""
|
|
|
|
since_ms = None
|
|
|
|
# user sets timerange, so find the start time
|
|
if timerange:
|
|
if timerange[0][0] == 'date':
|
|
since_ms = timerange[1] * 1000
|
|
elif timerange[0][1] == 'line':
|
|
num_minutes = timerange[2] * constants.TICKER_INTERVAL_MINUTES[tick_interval]
|
|
since_ms = arrow.utcnow().shift(minutes=num_minutes).timestamp * 1000
|
|
|
|
# read the cached file
|
|
if os.path.isfile(filename):
|
|
with open(filename, "rt") as file:
|
|
data = json.load(file)
|
|
# remove the last item, because we are not sure if it is correct
|
|
# it could be fetched when the candle was incompleted
|
|
if data:
|
|
data.pop()
|
|
else:
|
|
data = []
|
|
|
|
if data:
|
|
if since_ms and since_ms < data[0][0]:
|
|
# the data is requested for earlier period than the cache has
|
|
# so fully redownload all the data
|
|
data = []
|
|
else:
|
|
# a part of the data was already downloaded, so
|
|
# download unexist data only
|
|
since_ms = data[-1][0] + 1
|
|
|
|
return (data, since_ms)
|
|
|
|
|
|
def download_backtesting_testdata(datadir: str,
|
|
pair: str,
|
|
tick_interval: str = '5m',
|
|
timerange: Optional[Tuple[Tuple, int, int]] = None) -> None:
|
|
|
|
"""
|
|
Download the latest ticker intervals from the exchange for the pairs passed in parameters
|
|
The data is downloaded starting from the last correct ticker interval data that
|
|
esists in a cache. If timerange starts earlier than the data in the cache,
|
|
the full data will be redownloaded
|
|
|
|
Based on @Rybolov work: https://github.com/rybolov/freqtrade-data
|
|
:param pairs: list of pairs to download
|
|
:param tick_interval: ticker interval
|
|
:param timerange: range of time to download
|
|
:return: None
|
|
|
|
"""
|
|
|
|
path = make_testdata_path(datadir)
|
|
filepair = pair.replace("/", "_")
|
|
filename = os.path.join(path, f'{filepair}-{tick_interval}.json')
|
|
|
|
logger.info(
|
|
'Download the pair: "%s", Interval: %s',
|
|
pair,
|
|
tick_interval
|
|
)
|
|
|
|
data, since_ms = load_cached_data_for_updating(filename, tick_interval, timerange)
|
|
|
|
logger.debug("Current Start: %s", misc.format_ms_time(data[1][0]) if data else 'None')
|
|
logger.debug("Current End: %s", misc.format_ms_time(data[-1][0]) if data else 'None')
|
|
|
|
new_data = get_ticker_history(pair=pair, tick_interval=tick_interval, since_ms=since_ms)
|
|
data.extend(new_data)
|
|
|
|
logger.debug("New Start: %s", misc.format_ms_time(data[0][0]))
|
|
logger.debug("New End: %s", misc.format_ms_time(data[-1][0]))
|
|
|
|
misc.file_dump_json(filename, data)
|