stable/freqtrade/data/dataprovider.py
2022-09-01 23:15:03 -06:00

397 lines
15 KiB
Python

"""
Dataprovider
Responsible to provide data to the bot
including ticker and orderbook data, live and historical candle (OHLCV) data
Common Interface for bot and strategy to access data.
"""
import logging
from collections import deque
from datetime import datetime, timezone
from typing import Any, Dict, List, Optional, Tuple
from pandas import DataFrame
from freqtrade.configuration import TimeRange
from freqtrade.constants import ListPairsWithTimeframes, PairWithTimeframe
from freqtrade.data.history import load_pair_history
from freqtrade.enums import CandleType, RPCMessageType, RunMode
from freqtrade.exceptions import ExchangeError, OperationalException
from freqtrade.exchange import Exchange, timeframe_to_seconds
from freqtrade.rpc import RPCManager
from freqtrade.util import PeriodicCache
logger = logging.getLogger(__name__)
NO_EXCHANGE_EXCEPTION = 'Exchange is not available to DataProvider.'
MAX_DATAFRAME_CANDLES = 1000
class DataProvider:
def __init__(
self,
config: dict,
exchange: Optional[Exchange],
pairlists=None,
rpc: Optional[RPCManager] = None
) -> None:
self._config = config
self._exchange = exchange
self._pairlists = pairlists
self.__rpc = rpc
self.__cached_pairs: Dict[PairWithTimeframe, Tuple[DataFrame, datetime]] = {}
self.__slice_index: Optional[int] = None
self.__cached_pairs_backtesting: Dict[PairWithTimeframe, DataFrame] = {}
self.__producer_pairs_df: Dict[str,
Dict[PairWithTimeframe, Tuple[DataFrame, datetime]]] = {}
self.__producer_pairs: Dict[str, List[str]] = {}
self._msg_queue: deque = deque()
self.__msg_cache = PeriodicCache(
maxsize=1000, ttl=timeframe_to_seconds(self._config.get('timeframe', '1h')))
self._num_sources = len(
self._config.get('external_message_consumer', {}).get('producers', [])
)
self.external_data_enabled = self._num_sources > 0
def _set_dataframe_max_index(self, limit_index: int):
"""
Limit analyzed dataframe to max specified index.
:param limit_index: dataframe index.
"""
self.__slice_index = limit_index
def _set_cached_df(
self,
pair: str,
timeframe: str,
dataframe: DataFrame,
candle_type: CandleType
) -> None:
"""
Store cached Dataframe.
Using private method as this should never be used by a user
(but the class is exposed via `self.dp` to the strategy)
:param pair: pair to get the data for
:param timeframe: Timeframe to get data for
:param dataframe: analyzed dataframe
:param candle_type: Any of the enum CandleType (must match trading mode!)
"""
pair_key = (pair, timeframe, candle_type)
self.__cached_pairs[pair_key] = (
dataframe, datetime.now(timezone.utc))
# For multiple producers we will want to merge the pairlists instead of overwriting
def set_producer_pairs(self, pairlist: List[str], producer_name: str = "default"):
"""
Set the pairs received to later be used.
This only supports 1 Producer right now.
:param pairlist: List of pairs
"""
self.__producer_pairs[producer_name] = pairlist.copy()
def get_producer_pairs(self, producer_name: str = "default") -> List[str]:
"""
Get the pairs cached from the producer
:returns: List of pairs
"""
return self.__producer_pairs.get(producer_name, [])
def emit_df(
self,
pair_key: PairWithTimeframe,
dataframe: DataFrame
) -> None:
"""
Send this dataframe as an ANALYZED_DF message to RPC
:param pair_key: PairWithTimeframe tuple
:param data: Tuple containing the DataFrame and the datetime it was cached
"""
if self.__rpc:
self.__rpc.send_msg(
{
'type': RPCMessageType.ANALYZED_DF,
'data': {
'key': pair_key,
'value': dataframe
}
}
)
def add_external_df(
self,
pair: str,
timeframe: str,
dataframe: DataFrame,
candle_type: CandleType,
producer_name: str = "default"
) -> None:
"""
Add the pair data to this class from an external source.
:param pair: pair to get the data for
:param timeframe: Timeframe to get data for
:param candle_type: Any of the enum CandleType (must match trading mode!)
"""
pair_key = (pair, timeframe, candle_type)
if producer_name not in self.__producer_pairs_df:
self.__producer_pairs_df[producer_name] = {}
self.__producer_pairs_df[producer_name][pair_key] = (dataframe, datetime.now(timezone.utc))
def get_external_df(
self,
pair: str,
timeframe: str,
candle_type: CandleType,
producer_name: str = "default"
) -> Tuple[DataFrame, datetime]:
"""
Get the pair data from the external sources. Will wait if the policy is
set to, and data is not available.
:param pair: pair to get the data for
:param timeframe: Timeframe to get data for
:param candle_type: Any of the enum CandleType (must match trading mode!)
"""
pair_key = (pair, timeframe, candle_type)
if producer_name not in self.__producer_pairs_df:
# We don't have this data yet, return empty DataFrame and datetime (01-01-1970)
return (DataFrame(), datetime.fromtimestamp(0, tz=timezone.utc))
if pair_key not in self.__producer_pairs_df:
# We don't have this data yet, return empty DataFrame and datetime (01-01-1970)
return (DataFrame(), datetime.fromtimestamp(0, tz=timezone.utc))
return self.__producer_pairs_df[producer_name][pair_key]
def add_pairlisthandler(self, pairlists) -> None:
"""
Allow adding pairlisthandler after initialization
"""
self._pairlists = pairlists
def historic_ohlcv(
self,
pair: str,
timeframe: str = None,
candle_type: str = ''
) -> DataFrame:
"""
Get stored historical candle (OHLCV) data
:param pair: pair to get the data for
:param timeframe: timeframe to get data for
:param candle_type: '', mark, index, premiumIndex, or funding_rate
"""
_candle_type = CandleType.from_string(
candle_type) if candle_type != '' else self._config['candle_type_def']
saved_pair = (pair, str(timeframe), _candle_type)
if saved_pair not in self.__cached_pairs_backtesting:
timerange = TimeRange.parse_timerange(None if self._config.get(
'timerange') is None else str(self._config.get('timerange')))
# Move informative start time respecting startup_candle_count
timerange.subtract_start(
timeframe_to_seconds(str(timeframe)) * self._config.get('startup_candle_count', 0)
)
self.__cached_pairs_backtesting[saved_pair] = load_pair_history(
pair=pair,
timeframe=timeframe or self._config['timeframe'],
datadir=self._config['datadir'],
timerange=timerange,
data_format=self._config.get('dataformat_ohlcv', 'json'),
candle_type=_candle_type,
)
return self.__cached_pairs_backtesting[saved_pair].copy()
def get_pair_dataframe(
self,
pair: str,
timeframe: str = None,
candle_type: str = ''
) -> DataFrame:
"""
Return pair candle (OHLCV) data, either live or cached historical -- depending
on the runmode.
Only combinations in the pairlist or which have been specified as informative pairs
will be available.
:param pair: pair to get the data for
:param timeframe: timeframe to get data for
:return: Dataframe for this pair
:param candle_type: '', mark, index, premiumIndex, or funding_rate
"""
if self.runmode in (RunMode.DRY_RUN, RunMode.LIVE):
# Get live OHLCV data.
data = self.ohlcv(pair=pair, timeframe=timeframe, candle_type=candle_type)
else:
# Get historical OHLCV data (cached on disk).
data = self.historic_ohlcv(pair=pair, timeframe=timeframe, candle_type=candle_type)
if len(data) == 0:
logger.warning(f"No data found for ({pair}, {timeframe}, {candle_type}).")
return data
def get_analyzed_dataframe(self, pair: str, timeframe: str) -> Tuple[DataFrame, datetime]:
"""
Retrieve the analyzed dataframe. Returns the full dataframe in trade mode (live / dry),
and the last 1000 candles (up to the time evaluated at this moment) in all other modes.
:param pair: pair to get the data for
:param timeframe: timeframe to get data for
:return: Tuple of (Analyzed Dataframe, lastrefreshed) for the requested pair / timeframe
combination.
Returns empty dataframe and Epoch 0 (1970-01-01) if no dataframe was cached.
"""
pair_key = (pair, timeframe, self._config.get('candle_type_def', CandleType.SPOT))
if pair_key in self.__cached_pairs:
if self.runmode in (RunMode.DRY_RUN, RunMode.LIVE):
df, date = self.__cached_pairs[pair_key]
else:
df, date = self.__cached_pairs[pair_key]
if self.__slice_index is not None:
max_index = self.__slice_index
df = df.iloc[max(0, max_index - MAX_DATAFRAME_CANDLES):max_index]
return df, date
else:
return (DataFrame(), datetime.fromtimestamp(0, tz=timezone.utc))
@property
def runmode(self) -> RunMode:
"""
Get runmode of the bot
can be "live", "dry-run", "backtest", "edgecli", "hyperopt" or "other".
"""
return RunMode(self._config.get('runmode', RunMode.OTHER))
def current_whitelist(self) -> List[str]:
"""
fetch latest available whitelist.
Useful when you have a large whitelist and need to call each pair as an informative pair.
As available pairs does not show whitelist until after informative pairs have been cached.
:return: list of pairs in whitelist
"""
if self._pairlists:
return self._pairlists.whitelist.copy()
else:
raise OperationalException("Dataprovider was not initialized with a pairlist provider.")
def clear_cache(self):
"""
Clear pair dataframe cache.
"""
self.__cached_pairs = {}
self.__cached_pairs_backtesting = {}
self.__slice_index = 0
# Exchange functions
def refresh(self,
pairlist: ListPairsWithTimeframes,
helping_pairs: ListPairsWithTimeframes = None) -> None:
"""
Refresh data, called with each cycle
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
if helping_pairs:
self._exchange.refresh_latest_ohlcv(pairlist + helping_pairs)
else:
self._exchange.refresh_latest_ohlcv(pairlist)
@property
def available_pairs(self) -> ListPairsWithTimeframes:
"""
Return a list of tuples containing (pair, timeframe) for which data is currently cached.
Should be whitelist + open trades.
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
return list(self._exchange._klines.keys())
def ohlcv(
self,
pair: str,
timeframe: str = None,
copy: bool = True,
candle_type: str = ''
) -> DataFrame:
"""
Get candle (OHLCV) data for the given pair as DataFrame
Please use the `available_pairs` method to verify which pairs are currently cached.
:param pair: pair to get the data for
:param timeframe: Timeframe to get data for
:param candle_type: '', mark, index, premiumIndex, or funding_rate
:param copy: copy dataframe before returning if True.
Use False only for read-only operations (where the dataframe is not modified)
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
if self.runmode in (RunMode.DRY_RUN, RunMode.LIVE):
_candle_type = CandleType.from_string(
candle_type) if candle_type != '' else self._config['candle_type_def']
return self._exchange.klines(
(pair, timeframe or self._config['timeframe'], _candle_type),
copy=copy
)
else:
return DataFrame()
def market(self, pair: str) -> Optional[Dict[str, Any]]:
"""
Return market data for the pair
:param pair: Pair to get the data for
:return: Market data dict from ccxt or None if market info is not available for the pair
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
return self._exchange.markets.get(pair)
def ticker(self, pair: str):
"""
Return last ticker data from exchange
:param pair: Pair to get the data for
:return: Ticker dict from exchange or empty dict if ticker is not available for the pair
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
try:
return self._exchange.fetch_ticker(pair)
except ExchangeError:
return {}
def orderbook(self, pair: str, maximum: int) -> Dict[str, List]:
"""
Fetch latest l2 orderbook data
Warning: Does a network request - so use with common sense.
:param pair: pair to get the data for
:param maximum: Maximum number of orderbook entries to query
:return: dict including bids/asks with a total of `maximum` entries.
"""
if self._exchange is None:
raise OperationalException(NO_EXCHANGE_EXCEPTION)
return self._exchange.fetch_l2_order_book(pair, maximum)
def send_msg(self, message: str, *, always_send: bool = False) -> None:
"""
Send custom RPC Notifications from your bot.
Will not send any bot in modes other than Dry-run or Live.
:param message: Message to be sent. Must be below 4096.
:param always_send: If False, will send the message only once per candle, and surpress
identical messages.
Careful as this can end up spaming your chat.
Defaults to False
"""
if self.runmode not in (RunMode.DRY_RUN, RunMode.LIVE):
return
if always_send or message not in self.__msg_cache:
self._msg_queue.append(message)
self.__msg_cache[message] = True