Hi @backtrader ,
Just a follow up on this wonderful idea mixing backtrader+ influxDB +Grafana using Analyzer, here's what I did:
Pre-req:
1.You setup the grafana/influxdb connection, basic stuff like installed Json/influxdb-python blah blah blah.
- You created a influxdb client and passed to strategy using params
(i.e. self.strategy.params['client'] <==this is the influxdb client object - You already created a database called "simulation" to store the simulation result.
- For the Grafana OHLC bar please take reference here
Feel free to discuss and let me know if you have any great idea on it.
(Btw InfluxDB will have a high cardinality issue if talking about recording a lot of ohlc simulations, talking about 10k~ up with 16gb ram will be overflow, im thinking should i remove the candlestick instead...)
from __future__ import (absolute_import, division, print_function,
unicode_literals)
import backtrader as bt
import pandas as pd
from influxdb import DataFrameClient
class GrafanaAn(bt.Analyzer):
def create_analysis(self):
self.tradeBegV=0
self.tradeEndV=0
def notify_trade(self, trade):
t=self.datas[0].datetime.date(0).strftime('%Y-%m-%dT%H:%M:%SZ')
if trade.justopened:
self.tradeBegV=self.strategy.broker.getvalue()
self.tradeBegTime=t
self.tradeBegP=float(trade.price)
self.tradeSize=float(trade.size)
#Send a buy indicator to Grafana
jsonBody=[{"measurement":"portfolio",
"tags":{
"stock":self.data._name,
"market":self.strategy.params['market'],
"ind":self.strategy.indName
},
"time":t,
"fields":{
"BuyPrice":float(trade.price),
"Size":float(trade.size),
}
}]
ar=self.strategy.params['client'].write_points(jsonBody,database='simulation')
assert ar
elif trade.status == trade.Closed:
self.tradeEndV = self.strategy.broker.getvalue()
ret=(self.tradeEndV/self.tradeBegV-1)*100
t=self.datas[0].datetime.date(0).strftime('%Y-%m-%dT%H:%M:%SZ')
jsonBody=[{"measurement":"trade",
"tags":{
"stock":self.data._name,
"market":self.strategy.params['market'],
"ind":self.strategy.indName
},
"time":t,
"fields":{
"commission":float(trade.commission),
"return":float(ret),
"pnl":float(trade.pnl),
"pnlcomm":float(trade.pnlcomm),
"tradingTime":int(trade.barlen),
"beginTime":self.tradeBegTime,
"endTime":t
}
}]
ar=self.strategy.params['client'].write_points(jsonBody,database='simulation')
assert ar
jsonBody=[{"measurement":"portfolio",
"tags":{
"stock":self.data._name,
"market":self.strategy.params['market'],
"ind":self.strategy.indName
},
"time":t,
"fields":{
"SellPrice":float(trade.pnl/self.tradeSize+self.tradeBegP),
"Size":float(trade.size),
}
}]
ar=self.strategy.params['client'].write_points(jsonBody,database='simulation')
assert ar
def start(self):
self._i=0
self._dict={}
def next(self):
_value=self.strategy.broker.getvalue()
_pos=self.strategy.getposition().size
self._dict[self._i]={'portv':_value,'pos':_pos}
self._i+=1
#for k,ind in self.strategy.ind.items():
def stop(self):
portVFrame=pd.DataFrame.from_dict(self._dict,"index")
portVFrame.columns=['portv','pos']
priceFrame=self.data._dataname
portVFrame.set_index(priceFrame.index,inplace=True)
priceFrame['portv']=portVFrame['portv']
priceFrame['pos']=portVFrame['pos']
priceFrame=priceFrame.astype(float)
client=DataFrameClient(database='simulation')
ar=client.write_points(priceFrame,"portfolio",{"ind":self.strategy.indName,"market":self.strategy.params['market'],"stock":self.data._name},database='simulation')
assert ar
#Print Monthly return
_mr=self.strategy.analyzers.TR.get_analysis()
MonthlyReport=pd.DataFrame.from_dict(_mr,"index")
ar=client.write_points(MonthlyReport,"monthlyReturn",{"ind":self.strategy.indName,"market":self.strategy.params['market'],"stock":self.data._name},database='simulation')
assert ar
def get_analysis(self):
return None
'''