In [1]:

    

%pylab inline
import sys
sys.path.append("..")
from blotter import blotter
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
pd.options.display.max_rows = 10
blt = blotter.Blotter(prices="./prices", interest_rates="daily_interest_rates.csv", eod_time=None)
blt.connect_market_data()


    

    




Populating the interactive namespace from numpy and matplotlib

In [2]:

    

blt.define_generic("AUDUSD", ccy="USD", margin=0, multiplier=1,
                       commission=2.5, isFX=True)
blt.define_generic("USDJPY", ccy="JPY", margin=0, multiplier=1,
                       commission=2.5, isFX=True)
blt.map_instrument(generic="AUDUSD", instrument="AUDUSD")


    

In [3]:

    

aud = pd.read_csv("AUDUSD-1min.csv", parse_dates=True, index_col=0)


    

In [4]:

    

aud.head()


    

    
Out[4]:






  

    

      

      
bid
      
ask
    
    

      
timestamp
      

      

    
  
  

    

      
2017-01-02 00:00:00
      
0.72096
      
0.72188
    
    

      
2017-01-02 00:01:00
      
0.72096
      
0.72188
    
    

      
2017-01-02 00:02:00
      
0.72096
      
0.72188
    
    

      
2017-01-02 00:03:00
      
0.72096
      
0.72188
    
    

      
2017-01-02 00:04:00
      
0.72096
      
0.72188
    
  

In [5]:

    

aud.loc[:, "mid"] = (aud.loc[:, "bid"] + aud.loc[:, "ask"]) / 2


    

In [6]:

    

aud.loc[:, "mid"].plot(title="Mid Price", figsize=(12, 8))


    

    
Out[6]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f132ab27e80>






    

In [7]:

    

aud = aud.loc["2017-01-02T04:00:00":]


    

In [8]:

    

rets = np.log(aud.loc[:, "mid"] / aud.loc[:, "mid"].shift(20))


    

In [9]:

    

rets.plot()


    

    
Out[9]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f132ab27668>






    

In [10]:

    

rets[np.abs(rets) < 0.0005] = 0
aud_alpha = np.sign(rets).fillna(value=0)
aud_alpha = aud_alpha.resample("5T").last()
aud_trd = (aud_alpha - aud_alpha.shift(1)).fillna(value=0)

signal = pd.concat([aud_alpha, aud_trd], axis=1)
signal.columns = ["Alpha", "Trade"]
signal.plot(title="Alpha and Trades for AUDUSD", figsize=(12, 8))


    

    
Out[10]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f132776c5f8>






    

In [11]:

    

signal.tail()


    

    
Out[11]:






  

    

      

      
Alpha
      
Trade
    
    

      
timestamp
      

      

    
  
  

    

      
2017-01-02 23:35:00
      
1.0
      
0.0
    
    

      
2017-01-02 23:40:00
      
0.0
      
-1.0
    
    

      
2017-01-02 23:45:00
      
-1.0
      
-1.0
    
    

      
2017-01-02 23:50:00
      
-1.0
      
0.0
    
    

      
2017-01-02 23:55:00
      
-1.0
      
0.0
    
  

In [12]:

    

def get_price(prices, trade, ts):
    if trade > 0:
        return prices.loc[ts, "ask"]
    if trade < 0:
        return prices.loc[ts, "bid"]
    if trade == 0:
        return None


    

In [13]:

    

for ts, trd in signal.loc[:, "Trade"].iteritems():
    price = get_price(aud, trd, ts)
    blt.trade(ts, "AUDUSD", trd*1000000, price)
    # save intraday PnL calculations
    blt.dispatch_events(blt.create_events(ts, "PNL"))


    

In [14]:

    

blt.get_instruments()


    

    
Out[14]:





AUDUSD   -1000000.0
dtype: float64

In [15]:

    

pos = blt.get_instruments()
trd = -float(pos)
ts = pd.Timestamp("2017-01-02T23:59:00")
price = get_price(aud, trd, ts)
blt.trade(ts, "AUDUSD", trd, price)


    

In [16]:

    

pnls = blt.get_pnl_history()
pnls["USD"]


    

    
Out[16]:






  

    

      

      
pnl
      
closed pnl
      
open pnl
    
  
  

    

      
2017-01-02 04:55:00
      
-312.50000
      
-2.50000
      
-310.0
    
    

      
2017-01-02 05:00:00
      
-625.00000
      
-625.00000
      
0.0
    
    

      
2017-01-02 05:05:00
      
-625.00000
      
-625.00000
      
0.0
    
    

      
2017-01-02 05:10:00
      
-625.00000
      
-625.00000
      
0.0
    
    

      
2017-01-02 05:15:00
      
-625.00000
      
-625.00000
      
0.0
    
    

      
...
      
...
      
...
      
...
    
    

      
2017-01-02 23:35:00
      
-2612.53637
      
-2962.53637
      
350.0
    
    

      
2017-01-02 23:40:00
      
-2685.03637
      
-2685.03637
      
0.0
    
    

      
2017-01-02 23:45:00
      
-2737.53637
      
-2687.53637
      
-50.0
    
    

      
2017-01-02 23:50:00
      
-2392.53637
      
-2687.53637
      
295.0
    
    

      
2017-01-02 23:55:00
      
-2242.53637
      
-2687.53637
      
445.0
    
  

229 rows × 3 columns

In [17]:

    

pnls["USD"].plot(title="USD PnL", figsize=(12,8))


    

    
Out[17]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f13278004e0>






    

In [18]:

    

positions = blt.get_holdings_history()
aud_pos = positions["USD"]["AUDUSD"]
aud_pos.reindex(aud.index, method="ffill").plot(title="AUDUSD Position", figsize=(12,8))


    

    
Out[18]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f1325d2d4e0>