Pip install PyAlgoTrade 0.13

When running the Optimizing examples from tutorial, code crashes systems.

First ran as server and client on same machine. started server first, then client. Code ran for a bit,then system crashes... There is no logging for post mortem of code.

Second, Ran single file started up, but again crashed entire system.

Running OpenSuse 12.2, The both CPU were pinned @100%. No issue with memory.

Just did the getting started tutorial, and here are the errors that I got because of using Python 3. Unless I'm using the 2to3 tool wrong, there may be a few syntactic issues in the code (for Python 3 atleast). tools/yahoofinance.py:

Line 21: Current: import urllib2 Error: urllib2 module cannot be found Reason: Since I'm using Python 3, urllib2 is no apart of urllib, and urllib is split up into different modules, specifically urllib.request and urllib.errors. My change: import urllib.request as urllib2

Line 130: Current:except Exception, e Error: invalid syntax Reason: Python 3 semantics My change: except Exception as e

Line 48: Current: while not buff[0].isalnum() Error: 'int' object does not have attribute isalnum() Reason: isalnum() is only applicable to strings. My change: while not str(buff[0]).isalbum()

Line 67: Current: f.write(bars) Error: TypeError: must be str, not bytes Reason: See error My change: f.write(str(bars))

After that compiled, I went over to the actual strategy code. Here's my list of changes to the errors received:

utils/csvutils.py

line 30: Current: self.__fieldNames = self.reader.next() Error: AttributeError: '_csv.reader' object has no attribute 'next' Reason: Python 3 changes My change: ```self.__fieldNames = next(self.reader())

More to come later...

Hi,

If a I have a file of intraday data that spans several days. What I would like to do is to calculate sma on each day, and only start trading when sma for one day start to have value. In short, I would like to trade in each day seperately on sma, but I would like to keep the trading statistics in the same Strategy object.

Right now, if I supply a feed with many days' intraday data into feed

  feed = ninjatraderfeed.Feed(Frequency.MINUTE)
  feed.addBarsFromCSV(instrument, 'index_f.csv')

then I do this in the Strategy object,

   self.slow_ma = ma.SMA(feed[instrument].getCloseDataSeries(), slow_window)
   self.fast_ma = ma.SMA(feed[instrument].getCloseDataSeries(), fast_window)

I would like to confirm if this SMA calculation starts each day fresh? My experiment shows it doent. If it doesnt, how do I make it calculate SMA each day fresh, so if my fast_window = 5, each day I will have 4 None values at the beginning of the DataSeries self.fast_ma

I used "pip install Quandl" This what I get:

from pyalgotrade.tools import quandl

ImportError: cannot import name quandl

I already install Quandl and pyalgotrade. I am using windows 8. Why is not working?

I execute this command: $ python -c "from pyalgotrade.tools import yahoofinance; yahoofinance.download_daily_bars('orcl', 2000, 'orcl-2000.csv')" And receive error:

Traceback (most recent call last): File "", line 1, in File "/usr/local/lib/python2.7/site-packages/pyalgotrade/tools/yahoofinance.py", line 54, in download_daily_bars bars = download_csv(instrument, datetime.date(year, 1, 1), datetime.date(year, 12, 31), "d") File "/usr/local/lib/python2.7/site-packages/pyalgotrade/tools/yahoofinance.py", line 40, in download_csv return csvutils.download_csv(url) File "/usr/local/lib/python2.7/site-packages/pyalgotrade/utils/csvutils.py", line 59, in download_csv response.raise_for_status() File "/usr/local/lib/python2.7/site-packages/requests/models.py", line 929, in raise_for_status raise HTTPError(http_error_msg, response=self) requests.exceptions.HTTPError: 404 Client Error: Not Found for url: https://ichart.finance.yahoo.com/table.csv?s=orcl&a=0&b=1&c=2000&d=11&e=31&f=2000&g=d&ignore=.csv

How's that going?

Hi, Nice to see you, gbeced! Thanks for you great project which helps me work faster.

Sometimes we need to give the man some initial "btc" to do the trade, not only the cash. So I made this PR : )

Best Regards, winkidney

I'm confused about sample code in http://gbeced.github.io/pyalgotrade/docs/v0.18/html/bitstamp_example.html

usually, when I use Frequency.DAY and SMA(10) it means 10 days average price. What does SMA(10) mean when using Frequency.TRADE？

and how to use daily SMA in a barfeed with Frequency.TRADE?

Here we split broker_backtesting_test module into two separate files, one with tests specific the backtesting broker and another with tests can be used with different broker implementations.

What is the best way / proper way to get minute tickdata in as csv. Any special rules on the time format or colums placments?

Is there any specific way to pull in a numpy array instead?

Hi, I started experimenting with pyalgotrade today. I used the SMA crossover strategy sample posted here: http://gbeced.github.io/pyalgotrade/docs/v0.16/html/sample_sma_crossover.html I added a print pyalgotrade.__version__(the 0.16 output below). Plots are similar to your web site on my end. I receive the following RuntimeWarning however and I am therefore wondering what I am doing wrong?

My guess is that data is missing in the early periods of the MA. For reference, this is a Win7 PC with python 2.7.

Best regards, Matthias

C:\dvl\pybacktest_comparison>python pyalgo_sample.py
0.16
Sharpe ratio: 1.12
C:\Anaconda\lib\site-packages\matplotlib\transforms.py:644: RuntimeWarning: invalid value encountered in sign
  dx0 = np.sign(vertices[:, 0] - x0)
C:\Anaconda\lib\site-packages\matplotlib\transforms.py:645: RuntimeWarning: invalid value encountered in sign
  dy0 = np.sign(vertices[:, 1] - y0)
C:\Anaconda\lib\site-packages\matplotlib\transforms.py:646: RuntimeWarning: invalid value encountered in sign
  dx1 = np.sign(vertices[:, 0] - x1)
C:\Anaconda\lib\site-packages\matplotlib\transforms.py:647: RuntimeWarning: invalid value encountered in sign
  dy1 = np.sign(vertices[:, 1] - y1)

Hey:

Restarted working on this again and using the .getXXX and .setXXX getting in the way of sorts. Are you planning or were you considering removing the method leader of "get" and "set." Its not pythonic, its more java-ish...

Using snippets from tutorial: Now: def onStart(self): print "Initial portfolio value: $%.2f" % self.getBroker().getEquity()

New could be: def onStart(self): print "Initial portfolio value: $%.2f" % self.broker().equity()

Now: def onEnterOk(self, position): execInfo = position.getEntryOrder().getExecutionInfo() print "%s: BUY at $%.2f" % (execInfo.getDateTime(), execInfo.getPrice())

New could be: def onEnterOk(self, position): execInfo = position.entryOrder().executionInfo() print "%s: BUY at $%.2f" % (execInfo.dateTime(), execInfo.price())

A way to accomlish this would be: to create new 'defs' in the current classes with the shorten names. and then wrapper orginal functions.

or better yet. Create new def's with shorter names then call new function with old name.

Example

def richQuick(parms); 'do something that makes money'

def getRichQuick(parms): richQuick(parms) print 'Future Waring: get and set methods will be deprecated soon use shorter names''

What do you think??

I understand that this runs in python accesses information in a rather open way, but this is just a headless or "text on screen" and memorize all the commands type of script.

There is no user menu or txt based menu?

There are plots as shown at the bottom of this page http://gbeced.github.io/pyalgotrade/docs/v0.20/html/sample_quandl.html

Quandl it seems has been acquired so, can we still access the data or can i point the software to a different data source like SQL, csv or a pandas dataframe?

I've a PostgreSQL database with a lot of candle data that I want to use for backtesting. I want to create a PostgreSQL feed in order to use those data directly, instead of exporting them into a CSV and then using the CSV itself. But I'm encountering some difficulties to inherit BaseBarFeed because there are abstract methods like join, peekDateTime and so on that I don't know how to override.

Is there somethere a clear tutorial on about to create a custom feed?

Hello, I have a question.

I am trying backtesting for cryptocurrencies.

To do so I defined my own strategy that is inherited from BacktestingStrategy and ran it.

All is done but whenever I put decimal point value that is less than zero (e.g, 0.025) to quantity parameter of enterLong or enterShort, it seems that zero value is passed instead.

How do I make an order with quantity in decimal point unit?

I developed this script to download 1m chart of ADA-USD:

#!/usr/bin/env python3

import yfinance
import os
from datetime import date, timedelta

start_1m = date.today() - timedelta(days=7)

# Getting 1m charts
data = yfinance.download( 'ADA-USD', start=start_1m, end=today, interval="1m")
data.to_csv("data/ADA-USD_1mn.csv")

and when getting bars of this csv file, I got this exception from membf.py:

    raise Exception("Duplicate bars found for %s on %s" % (list(ret.keys()), smallestDateTime))

Exception: Duplicate bars found for ['ADA-USD'] on 2021-09-10 00:00:00

But there is no duplication in the CSV file. I checked this with the Linux uniq cmd.

Can someone help me ti understand this error PLZ?

I was trying to use my own csv data feed to import the data into the strategy using tutorial from the documentation. However, when onBars function is called it yields the following error:AttributeError: 'dict' object has no attribute 'getInstruments'

When I've used quandlfeed csv files which were downloaded as a result of quandl.build_feed and used guide from http://gbeced.github.io/pyalgotrade/docs/v0.20/html/sample_quandl.html it still yields the same error. It looks like somethinhg is wrong with csv data import. Thank you