OptReturn is a simplified return instance which mimics a strategy, to reduce the overhead of message passing from worker processes to the main process.
See: Docs - Optimization Improvements
Plotting has never been tested with optimization (not an expected use case)
Old comments are available at: https://disqus.com/home/forum/backtrader/
@viktor-w said in Cheat_on_open true: how to execute the order on another price than the next open price:
For now my code looks like this:
class TestStrategy(bt.Strategy): params = ( ('maperiod', 60), ) def log(self, txt, dt=None): ''' Logging function for this strategy''' dt = dt or self.datas[0].datetime.date(0) print('%s, %s' % (dt.isoformat(), txt)) def __init__(self): # Keep a reference to the "close" line in the data[0] dataseries self.dataclose = self.datas[0].close self.data_cpa = self.datas[0].high self.data_cpb = self.datas[0].low self.order = None self.buyprice = None self.buycomm = None self.sma = bt.indicators.SimpleMovingAverage( self.datas[0], period=self.params.maperiod) def notify_order(self, order): if order.status in [order.Submitted, order.Accepted]: # Buy/Sell order submitted/accepted to/by broker - Nothing to do return # Check if an order has been completed # Attention: broker could reject order if not enough cash if order.status in [order.Completed]: if order.isbuy(): self.log('BUY EXECUTED, %.8f' % order.executed.price) elif order.issell(): self.log('SELL EXECUTED, %.8f' % order.executed.price) self.bar_executed = 0 elif order.status in [order.Canceled, order.Margin, order.Rejected]: self.log('Order Canceled/Margin/Rejected') # Write down: no pending order self.order = None def notify_trade(self, trade): if not trade.isclosed: return self.log('OPERATION PROFIT, GROSS %.8f, NET %.f' % (trade.pnl, trade.pnlcomm)) def next(self): # Simply log the closing price of the series from the reference self.log('Close, %.8f' % self.dataclose[0]) if self.order: return # Check if we are in the market if not self.position: if self.dataclose[0] > self.sma[0]: # BUY, BUY, BUY!!! (with all possible default parameters) self.log('BUY CREATE, %.8f' % self.dataclose[0]) # Keep track of the created order to avoid a 2nd order self.order = self.buy(price= self.dataclose[0]) else: if self.dataclose[0] < self.sma[0]: # SELL, SELL, SELL!!! (with all possible default parameters) self.log('SELL CREATE, %.8f' % self.dataclose[0]) # Keep track of the created order to avoid a 2nd order self.order = self.sell(price= self.dataclose[0]) if __name__ == '__main__': # Create a cerebro entity cerebro = bt.Cerebro(cheat_on_open=True) # Add a strategy cerebro.addstrategy(TestStrategy) # retrieve the csv file modpath = os.path.dirname(os.path.abspath(sys.argv[0])) datapath = os.path.join(modpath, './datas/zrxeth_sample.txt') # Create a Data Feed data = bt.feeds.GenericCSVData( dataname = datapath, fromdate=datetime.datetime(2018, 9, 28), todate=datetime.datetime(2018, 12, 3), nullvalue=0.0, dtformat=('%Y-%m-%d %H:%M:%S'), datetime=0, high=1, low=2, open=3, close=4, volume=5, openinterest=-1 ) # Add the Data Feed to Cerebro cerebro.adddata(data) # Set our desired cash start cerebro.broker.setcash(100.0) # Print out the starting conditions print('Starting Portfolio Value: %.8f' % cerebro.broker.getvalue()) # Run over everything cerebro.run() # Print out the final result print('Final Portfolio Value: %.8f' % cerebro.broker.getvalue())From my understanding of the library, when you use the parameter cheat_on_open=True, you can authorize the order the order to be executed directly and not the day after. However, the executed order takes as price the next open price, how is it possible to choose which price the order will apply? Thanks!
@tw00000 said in Simple buy-and-hold strategy:
My only guess is that either self.buy() or self.close() is not working as intended here.
Sorry, but with a statement like that I don't really feel like writing a proper answer. Please have a look at your own code.
If If wrote this code
import queue
class MyQueue(queue.Queue):
def __init__(self, **kwargs):
print('doing something')
Would you expect MyQueue to be inside the queue module?
The same applies to something you define in your own script named Seasonals. It's not going to be inside bt.feeds (or backtrader.feeds without aliasing)
@viktor-w said in Misunderstanding of the process: Order created but not executed at the same price:
I used the code Quickstart in the documentation but with my own data (which is on a minute basis)
No you haven't. You are using something different. Your code
@viktor-w said in Misunderstanding of the process: Order created but not executed at the same price:
# Keep track of the created order to avoid a 2nd order self.order = self.buy(price= self.dataclose[0])
Code from: https://www.backtrader.com/docu/quickstart/quickstart.html#do-not-only-buy-but-sell
# Keep track of the created order to avoid a 2nd order self.order = self.buy()
There is a clear obvious difference. When you code, small differences can mean a huge difference.
@viktor-w said in Misunderstanding of the process: Order created but not executed at the same price:
I would like now to understand why my orders are created but not executed at the same price and days later... thanks! I am really a beginner with the platform, so any help is more than welcome.
The code is apparently producing logs in the console. Why don't you paste them and point out an example of what you think is wrong? This is a rhetorical question to point out, that if you think there is something wrong, we cannot know it, because there is no information as to what yo think is wrong.
Furthermore, you only show the strategy, but believe me: the most obvious mistake you are making it's not in that part of the code. It is in the code you aren't showing.
In any case let me point you to this:
and look at what a Market order is. Let me explain it here: it is an order for which YOU DON'T specify the price. The Market price is used. No, not the market price you currently see, but the one which you still haven't seen, i.e.: the next incoming price tick (if you use daily bars, that's already the next trading day)
@franklili said in multiprocessing.pool.MaybeEncodingError:
multiprocessing.pool.MaybeEncodingError: Error sending result: '[<main.EmaMeanReversionStrategy object at 0x7f4ab7de6b38>]'. Reason: 'PicklingError("Can't pickle <class'main.EmaMeanReversionStrategy'>: it's not the same object as__main__.EmaMeanReversionStrategy")'
This multiprocessing error has to do with scoping in Python, hence the ("it's not the same ...") It is a well-known pickling problem.
My guess is that you have to take the core out of if __name___ ... and put it into a function of its own.
@FaiqS said in Multiple Datafeeds and multiple Timeframes:
Which is not convenient to iterate through in next function of strategy.
It may not be convenient for you. Don't mix them in such a way then. Add first the lower timeframes and later the higher timeframes. You will have two clear partitions.
Or iterate using steps other than 1 and with different starting indices.
@FaiqS said in Multiple Datafeeds and multiple Timeframes:
The straightforward it is referring to is to add "cerebro.adddata(data)"
No. It is referring to resampledata(data, ...). Using any other approach won't change the order in which things are in the system and where.
@RandyT There is no magic. The problem is that the initial values of the filter are orders of magnitude greater than the ones when the filter has settled.
Question number 1:
SharpeRatio analyzer)Question number 2:
Question number 3 (set of questions)
Question number 4:
365 in your code? That's not what people use.@Suraj-Thorat said in Pandas Dataframe issue (int object has no attribute to_pydatetime):
Anyone has got a solution to this?
Yes. Read the documentation.
The error is self-explaining.
int object has no attribute to_pydatetime
@Suraj-Thorat said in Pandas Dataframe issue (int object has no attribute to_pydatetime):
datetime open high low close volume 0 2019-09-03 15.50 15.50 14.30 14.45 681 1 2019-09-04 14.20 15.45 14.10 14.90 5120
And you have an index which is made up of int values.
@Suraj-Thorat said in Pandas Dataframe issue (int object has no attribute to_pydatetime):
- EVEN IF I add this line to set index same error is given.
Basic engineering principles:
The PandasData has a dedicated documentation page, which shows the parameters available for the data feed, and the following is stated
Link: Docs - PandasData - https://www.backtrader.com/docu/pandas-datafeed/pandas-datafeed/
...
params = (
# Possible values for datetime (must always be present)
# None : datetime is the "index" in the Pandas Dataframe
# -1 : autodetect position or case-wise equal name
# >= 0 : numeric index to the colum in the pandas dataframe
# string : column name (as index) in the pandas dataframe
('datetime', None),
...
You probably want to use a Store, rather than a fork (which had no good reason to be). Someone took the right approach extracting the code from the fork.
Use coupling: Docs - Mixing Timeframes in Indicators
This functionality allows to calculate an indicator on a timeframe an automatically apply the indicator to another lower timeframe.
@ab_trader said in AttributeError: 'str' object has no attribute '_name' - The error is clear - but for the life of me I can't see why it's happening:
I can attach the code if that would be helpful
Everybody here can read other people minds, sure no script is needed.
@east1992 said in AttributeError: 'str' object has no attribute '_name' - The error is clear - but for the life of me I can't see why it's happening:
@ab_trader As you like
I think @ab_trader was just trying a prank ...
In any case ... let me wonder how you write such a script, yet miss the top of the forum
For code/output blocks: Use ``` (aka backtick or grave accent) in a single line before and after the block. See: http://commonmark.org/help/
Without the three backticks your code is basically rendered in a form which cannot be properly read.
@east1992 said in AttributeError: 'str' object has no attribute '_name' - The error is clear - but for the life of me I can't see why it's happening:
File "C:\ProgramData\Anaconda3\envs\finance\lib\site-packages\backtrader\broker.py", line 80, in getcommissioninfo if data._name in self.comminfo: AttributeError: 'str' object has no attribute '_name'
The problem is clear: you have created a position for something which is a string and not a data feed. Hence data (which is a str) has no attribute _name
From what one can read, it is unclear why you need to define a `get_data_by_name', given that a method to retrieve data feeds by name does already exist
getdatabynameBut playing with the names is for sure the root of the problem.
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
The current implementation of this indicator does not works: it has these two problems:
Sorry, but it does work.
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
- both the H1 and the D1 charts are shown. I only want the H1 chart.
You have not told the system that you only want the H1 chart. It cannot be a problem. You have not configured things to work the way you want.
See - Docs - Plotting and use plotinfo.plot=False for whatever you don't want plotted.
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
- the indicator shows the right values in the D1 chart, but it is not plotted in the H1 chart
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
data_h1 = bt.feeds.GenericCSVData(**{**csv_opts, **h1_opts}) data_d1 = bt.feeds.GenericCSVData(**{**csv_opts, **d1_opts})
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
self.levels = Levels(self.data1, period=self.p.period)
Well, you load d1 as the second feed in the system (hence self.data1) and you apply the indicator to self.data1. It should be no surprise that the indicator plots on d1 and not on h1. It would be a surprise if it did otherwise.
Again. See - Docs - Plotting and use plotinfo.plotmaster=YOUR_DESIRED_DATA to plot on your chosen target.
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
See? No indicator in the H1 chart and a superfluous D1 chart.
See above. There is nothing superfluous. You have to configure what you want and what you don't want and where you want it and where you don't want it.
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
As you can see in the following code, at the top level I first load the H1 data and then the D1 data. . This should make H1 the default data source, right?
No idea what you mean with "default data source". There is no such thing in backtrader. The only thing you have achieved is that h1 can be referred as self.data0 or self.datas[0] or self.data. This is described here: Docs - Strategy in the section Member Attributes
@fino said in Multi timeframe indicator plotted twice and in the timeframe:
Then, in the strategy, I set up the indicator to use self.data1. This should link its calculations to the D1 timeframe, shouldn't it?
Yes. And that's why the indicator obviously plots on d1
Indicators can take anything as a data feed, including other indicators and temporary line buffers created out of (for example) arithmetic operations, which of course immediately forbids something like getdatabyname
There isn't at the moment
@michael172 said in Extending Datafeeds GenericCSV Attribute Error:
AttributeError: module 'backtrader.feeds' has no attribute 'GenericCSV_PE'
@michael172 said in Extending Datafeeds GenericCSV Attribute Error:
if __name__ == '__main__': # Create a cerebro entity cerebro = bt.Cerebro() data = btfeeds.GenericCSV_PE( # data = btfeeds.GenericCSVData(
@michael172 said in Extending Datafeeds GenericCSV Attribute Error:
class GenericCSV_PE(GenericCSVData): # Add a 'pe' line to the inherited ones from the base class lines = ('pe_ttm',)
This is something you define.
The original answer does still apply
@backtrader said in Extending Datafeeds GenericCSV Attribute Error:
@Osofuego said in Extending Datafeeds GenericCSV Attribute Error:
data = btfeeds.GenericCSVData(You are not using your own extended data feed.