adelorenzo/stable
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #7842 from wizrds/feat/refactor-emc

Browse Source 
Change to broadcasting single candles in Producer/Consumer
This commit is contained in:
Matthias
2022-12-15 19:47:29 +01:00
committed by GitHub
parent b915872f66 ca2a878b86
commit 2828255435
10 changed files with 345 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
freqtrade/constants.py
View File

@@ -61,6 +61,7 @@ USERPATH_FREQAIMODELS = 'freqaimodels'
TELEGRAM_SETTING_OPTIONS = ['on', 'off', 'silent']
WEBHOOK_FORMAT_OPTIONS = ['form', 'json', 'raw']
FULL_DATAFRAME_THRESHOLD = 100
ENV_VAR_PREFIX = 'FREQTRADE__'

89
freqtrade/data/dataprovider.py
View File

@@ -9,14 +9,16 @@ from collections import deque
from datetime import datetime, timezone
from typing import Any, Dict, List, Optional, Tuple
from pandas import DataFrame
from pandas import DataFrame, to_timedelta
from freqtrade.configuration import TimeRange
from freqtrade.constants import Config, ListPairsWithTimeframes, PairWithTimeframe
from freqtrade.constants import (FULL_DATAFRAME_THRESHOLD, Config, 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.misc import append_candles_to_dataframe
from freqtrade.rpc import RPCManager
from freqtrade.util import PeriodicCache
@@ -120,7 +122,7 @@ class DataProvider:
'type': RPCMessageType.ANALYZED_DF,
'data': {
'key': pair_key,
'df': dataframe,
'df': dataframe.tail(1),
'la': datetime.now(timezone.utc)
}
}
@@ -131,7 +133,7 @@ class DataProvider:
'data': pair_key,
})
def _add_external_df(
def _replace_external_df(
self,
pair: str,
dataframe: DataFrame,
@@ -157,6 +159,85 @@ class DataProvider:
self.__producer_pairs_df[producer_name][pair_key] = (dataframe, _last_analyzed)
logger.debug(f"External DataFrame for {pair_key} from {producer_name} added.")
def _add_external_df(
self,
pair: str,
dataframe: DataFrame,
last_analyzed: datetime,
timeframe: str,
candle_type: CandleType,
producer_name: str = "default"
) -> Tuple[bool, int]:
"""
Append a candle to the existing external dataframe. The incoming dataframe
must have at least 1 candle.
: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!)
:returns: False if the candle could not be appended, or the int number of missing candles.
"""
pair_key = (pair, timeframe, candle_type)
if dataframe.empty:
# The incoming dataframe must have at least 1 candle
return (False, 0)
if len(dataframe) >= FULL_DATAFRAME_THRESHOLD:
# This is likely a full dataframe
# Add the dataframe to the dataprovider
self._replace_external_df(
pair,
dataframe,
last_analyzed=last_analyzed,
timeframe=timeframe,
candle_type=candle_type,
producer_name=producer_name
)
return (True, 0)
if (producer_name not in self.__producer_pairs_df
or pair_key not in self.__producer_pairs_df[producer_name]):
# We don't have data from this producer yet,
# or we don't have data for this pair_key
# return False and 1000 for the full df
return (False, 1000)
existing_df, _ = self.__producer_pairs_df[producer_name][pair_key]
# CHECK FOR MISSING CANDLES
timeframe_delta = to_timedelta(timeframe) # Convert the timeframe to a timedelta for pandas
local_last = existing_df.iloc[-1]['date'] # We want the last date from our copy
incoming_first = dataframe.iloc[0]['date'] # We want the first date from the incoming
# Remove existing candles that are newer than the incoming first candle
existing_df1 = existing_df[existing_df['date'] < incoming_first]
candle_difference = (incoming_first - local_last) / timeframe_delta
# If the difference divided by the timeframe is 1, then this
# is the candle we want and the incoming data isn't missing any.
# If the candle_difference is more than 1, that means
# we missed some candles between our data and the incoming
# so return False and candle_difference.
if candle_difference > 1:
return (False, candle_difference)
if existing_df1.empty:
appended_df = dataframe
else:
appended_df = append_candles_to_dataframe(existing_df1, dataframe)
# Everything is good, we appended
self._replace_external_df(
pair,
appended_df,
last_analyzed=last_analyzed,
timeframe=timeframe,
candle_type=candle_type,
producer_name=producer_name
)
return (True, 0)
def get_producer_df(
self,
pair: str,

18
freqtrade/misc.py
View File

@@ -301,3 +301,21 @@ def remove_entry_exit_signals(dataframe: pd.DataFrame):
dataframe[SignalTagType.EXIT_TAG.value] = None
return dataframe
def append_candles_to_dataframe(left: pd.DataFrame, right: pd.DataFrame) -> pd.DataFrame:
"""
Append the `right` dataframe to the `left` dataframe
:param left: The full dataframe you want appended to
:param right: The new dataframe containing the data you want appended
:returns: The dataframe with the right data in it
"""
if left.iloc[-1]['date'] != right.iloc[-1]['date']:
left = pd.concat([left, right])
# Only keep the last 1500 candles in memory
left = left[-1500:] if len(left) > 1500 else left
left.reset_index(drop=True, inplace=True)
return left

3
freqtrade/rpc/api_server/api_ws.py
View File

@@ -91,9 +91,10 @@ async def _process_consumer_request(
elif type == RPCRequestType.ANALYZED_DF:
# Limit the amount of candles per dataframe to 'limit' or 1500
limit = min(data.get('limit', 1500), 1500) if data else None
pair = data.get('pair', None) if data else None
# For every pair in the generator, send a separate message
for message in rpc._ws_request_analyzed_df(limit):
for message in rpc._ws_request_analyzed_df(limit, pair):
# Format response
response = WSAnalyzedDFMessage(data=message)
await channel.send(response.dict(exclude_none=True))

7
freqtrade/rpc/api_server/ws/channel.py
View File

@@ -27,7 +27,8 @@ class WebSocketChannel:
self,
websocket: WebSocketType,
channel_id: Optional[str] = None,
serializer_cls: Type[WebSocketSerializer] = HybridJSONWebSocketSerializer
serializer_cls: Type[WebSocketSerializer] = HybridJSONWebSocketSerializer,
send_throttle: float = 0.01
):
self.channel_id = channel_id if channel_id else uuid4().hex[:8]
self._websocket = WebSocketProxy(websocket)
@@ -41,6 +42,7 @@ class WebSocketChannel:
self._send_times: Deque[float] = deque([], maxlen=10)
# High limit defaults to 3 to start
self._send_high_limit = 3
self._send_throttle = send_throttle
# The subscribed message types
self._subscriptions: List[str] = []
@@ -106,7 +108,8 @@ class WebSocketChannel:
# Explicitly give control back to event loop as
# websockets.send does not
await asyncio.sleep(0.01)
# Also throttles how fast we send
await asyncio.sleep(self._send_throttle)
async def recv(self):
"""

2
freqtrade/rpc/api_server/ws_schemas.py
View File

@@ -47,7 +47,7 @@ class WSWhitelistRequest(WSRequestSchema):
class WSAnalyzedDFRequest(WSRequestSchema):
type: RPCRequestType = RPCRequestType.ANALYZED_DF
data: Dict[str, Any] = {"limit": 1500}
data: Dict[str, Any] = {"limit": 1500, "pair": None}
# ------------------------------ MESSAGE SCHEMAS ----------------------------

120
freqtrade/rpc/external_message_consumer.py
View File

@@ -8,15 +8,17 @@ import asyncio
import logging
import socket
from threading import Thread
from typing import TYPE_CHECKING, Any, Callable, Dict, List, TypedDict
from typing import TYPE_CHECKING, Any, Callable, Dict, List, TypedDict, Union
import websockets
from pydantic import ValidationError
from freqtrade.constants import FULL_DATAFRAME_THRESHOLD
from freqtrade.data.dataprovider import DataProvider
from freqtrade.enums import RPCMessageType
from freqtrade.misc import remove_entry_exit_signals
from freqtrade.rpc.api_server.ws import WebSocketChannel
from freqtrade.rpc.api_server.ws.channel import WebSocketChannel, create_channel
from freqtrade.rpc.api_server.ws.message_stream import MessageStream
from freqtrade.rpc.api_server.ws_schemas import (WSAnalyzedDFMessage, WSAnalyzedDFRequest,
WSMessageSchema, WSRequestSchema,
WSSubscribeRequest, WSWhitelistMessage,
@@ -38,6 +40,10 @@ class Producer(TypedDict):
logger = logging.getLogger(__name__)
def schema_to_dict(schema: Union[WSMessageSchema, WSRequestSchema]):
return schema.dict(exclude_none=True)
class ExternalMessageConsumer:
"""
The main controller class for consuming external messages from
@@ -92,6 +98,8 @@ class ExternalMessageConsumer:
RPCMessageType.ANALYZED_DF: self._consume_analyzed_df_message,
}
self._channel_streams: Dict[str, MessageStream] = {}
self.start()
def start(self):
@@ -118,6 +126,8 @@ class ExternalMessageConsumer:
logger.info("Stopping ExternalMessageConsumer")
self._running = False
self._channel_streams = {}
if self._sub_tasks:
# Cancel sub tasks
for task in self._sub_tasks:
@@ -175,7 +185,6 @@ class ExternalMessageConsumer:
:param producer: Dictionary containing producer info
:param lock: An asyncio Lock
"""
channel = None
while self._running:
try:
host, port = producer['host'], producer['port']
@@ -190,19 +199,21 @@ class ExternalMessageConsumer:
max_size=self.message_size_limit,
ping_interval=None
) as ws:
channel = WebSocketChannel(ws, channel_id=name)
async with create_channel(
ws,
channel_id=name,
send_throttle=0.5
) as channel:
logger.info(f"Producer connection success - {channel}")
# Create the message stream for this channel
self._channel_streams[name] = MessageStream()
# Now request the initial data from this Producer
for request in self._initial_requests:
await channel.send(
request.dict(exclude_none=True)
# Run the channel tasks while connected
await channel.run_channel_tasks(
self._receive_messages(channel, producer, lock),
self._send_requests(channel, self._channel_streams[name])
)
# Now receive data, if none is within the time limit, ping
await self._receive_messages(channel, producer, lock)
except (websockets.exceptions.InvalidURI, ValueError) as e:
logger.error(f"{ws_url} is an invalid WebSocket URL - {e}")
break
@@ -229,11 +240,19 @@ class ExternalMessageConsumer:
# An unforseen error has occurred, log and continue
logger.error("Unexpected error has occurred:")
logger.exception(e)
await asyncio.sleep(self.sleep_time)
continue
finally:
if channel:
await channel.close()
async def _send_requests(self, channel: WebSocketChannel, channel_stream: MessageStream):
# Send the initial requests
for init_request in self._initial_requests:
await channel.send(schema_to_dict(init_request))
# Now send any subsequent requests published to
# this channel's stream
async for request, _ in channel_stream:
logger.debug(f"Sending request to channel - {channel} - {request}")
await channel.send(request)
async def _receive_messages(
self,
@@ -270,19 +289,31 @@ class ExternalMessageConsumer:
latency = (await asyncio.wait_for(pong, timeout=self.ping_timeout) * 1000)
logger.info(f"Connection to {channel} still alive, latency: {latency}ms")
continue
except (websockets.exceptions.ConnectionClosed):
# Just eat the error and continue reconnecting
logger.warning(f"Disconnection in {channel} - retrying in {self.sleep_time}s")
await asyncio.sleep(self.sleep_time)
break
except Exception as e:
# Just eat the error and continue reconnecting
logger.warning(f"Ping error {channel} - {e} - retrying in {self.sleep_time}s")
logger.debug(e, exc_info=e)
await asyncio.sleep(self.sleep_time)
raise
break
def send_producer_request(
self,
producer_name: str,
request: Union[WSRequestSchema, Dict[str, Any]]
):
"""
Publish a message to the producer's message stream to be
sent by the channel task.
:param producer_name: The name of the producer to publish the message to
:param request: The request to send to the producer
"""
if isinstance(request, WSRequestSchema):
request = schema_to_dict(request)
if channel_stream := self._channel_streams.get(producer_name):
channel_stream.publish(request)
def handle_producer_message(self, producer: Producer, message: Dict[str, Any]):
"""
@@ -336,16 +367,45 @@ class ExternalMessageConsumer:
pair, timeframe, candle_type = key
if df.empty:
logger.debug(f"Received Empty Dataframe for {key}")
return
# If set, remove the Entry and Exit signals from the Producer
if self._emc_config.get('remove_entry_exit_signals', False):
df = remove_entry_exit_signals(df)
# Add the dataframe to the dataprovider
self._dp._add_external_df(pair, df,
last_analyzed=la,
timeframe=timeframe,
candle_type=candle_type,
producer_name=producer_name)
logger.debug(f"Received {len(df)} candle(s) for {key}")
did_append, n_missing = self._dp._add_external_df(
pair,
df,
last_analyzed=la,
timeframe=timeframe,
candle_type=candle_type,
producer_name=producer_name
)
if not did_append:
# We want an overlap in candles incase some data has changed
n_missing += 1
# Set to None for all candles if we missed a full df's worth of candles
n_missing = n_missing if n_missing < FULL_DATAFRAME_THRESHOLD else 1500
logger.warning(f"Holes in data or no existing df, requesting {n_missing} candles "
f"for {key} from `{producer_name}`")
self.send_producer_request(
producer_name,
WSAnalyzedDFRequest(
data={
"limit": n_missing,
"pair": pair
}
)
)
return
logger.debug(
f"Consumed message from `{producer_name}` of type `RPCMessageType.ANALYZED_DF`")
f"Consumed message from `{producer_name}` "
f"of type `RPCMessageType.ANALYZED_DF` for {key}")

39
freqtrade/rpc/rpc.py
View File

@@ -1062,15 +1062,26 @@ class RPC:
return self._convert_dataframe_to_dict(self._freqtrade.config['strategy'],
pair, timeframe, _data, last_analyzed)
def __rpc_analysed_dataframe_raw(self, pair: str, timeframe: str,
limit: Optional[int]) -> Tuple[DataFrame, datetime]:
""" Get the dataframe and last analyze from the dataprovider """
def __rpc_analysed_dataframe_raw(
self,
pair: str,
timeframe: str,
limit: Optional[int]
) -> Tuple[DataFrame, datetime]:
"""
Get the dataframe and last analyze from the dataprovider
:param pair: The pair to get
:param timeframe: The timeframe of data to get
:param limit: The amount of candles in the dataframe
"""
_data, last_analyzed = self._freqtrade.dataprovider.get_analyzed_dataframe(
pair, timeframe)
_data = _data.copy()
if limit:
_data = _data.iloc[-limit:]
return _data, last_analyzed
def _ws_all_analysed_dataframes(
@@ -1078,7 +1089,16 @@ class RPC:
pairlist: List[str],
limit: Optional[int]
) -> Generator[Dict[str, Any], None, None]:
""" Get the analysed dataframes of each pair in the pairlist """
"""
Get the analysed dataframes of each pair in the pairlist.
If specified, only return the most recent `limit` candles for
each dataframe.
:param pairlist: A list of pairs to get
:param limit: If an integer, limits the size of dataframe
If a list of string date times, only returns those candles
:returns: A generator of dictionaries with the key, dataframe, and last analyzed timestamp
"""
timeframe = self._freqtrade.config['timeframe']
candle_type = self._freqtrade.config.get('candle_type_def', CandleType.SPOT)
@@ -1091,10 +1111,15 @@ class RPC:
"la": last_analyzed
}
def _ws_request_analyzed_df(self, limit: Optional[int]):
def _ws_request_analyzed_df(
self,
limit: Optional[int] = None,
pair: Optional[str] = None
):
""" Historical Analyzed Dataframes for WebSocket """
whitelist = self._freqtrade.active_pair_whitelist
return self._ws_all_analysed_dataframes(whitelist, limit)
pairlist = [pair] if pair else self._freqtrade.active_pair_whitelist
return self._ws_all_analysed_dataframes(pairlist, limit)
def _ws_request_whitelist(self):
""" Whitelist data for WebSocket """