In [25]:

    

import pandas as pd
import numpy as np
import matplotlib.pylab as plt
import csv
import glob


    

In [2]:

    

#s5s.close()
#h5s = pd.HDFStore('C:\\Users\Michal\Dropbox\IB_data\SPY'+ '.h5s', 'r')
#spy = h5s['SPY']


    

In [5]:

    

def read_data(path):
    all_features=[]
    time_list=[]
    #sym_list = []
    bidP_list = []
    bidS_list=[]
    askP_list=[]
    askS_list=[]
    count=0
    #path = 'C:\\Users\Michal\Dropbox\IB_data'
    csv_file = open(path, 'r')
    content = csv_file.readlines()
    filtered=(line.replace('\n','') for line in content)
    for line in filtered:
        l=line.split(',')
        all_features.append(l)
        time,askPrice,askSize,bidPrice,bidSize = l
        time_list.append(time)
        #sym_list.append(sym)
        bidP_list.append(float(bidPrice))
        bidS_list.append(float(bidSize))
        askP_list.append(float(askPrice))
        askS_list.append(float(askSize))
    return time,bidP_list,bidS_list,askP_list,askS_list
    #return askP_list


    

In [37]:

    

def read_data(path):
    all_features=[]
    time_list=[]
    #sym_list = []
    bidP_list = []
    bidS_list=[]
    askP_list=[]
    askS_list=[]

    allFiles = glob.glob(path)
    frame = pd.DataFrame()
    list_ = []
    for file_ in allFiles:
        df = pd.read_csv(file_,index_col=None, header=0)
        list_.append(df)
        l=list_.split(',')
        all_features.append(l)
        time,askPrice,askSize,bidPrice,bidSize = l
        time_list.append(time)
        #sym_list.append(sym)
        bidP_list.append(float(bidPrice))
        bidS_list.append(float(bidSize))
        askP_list.append(float(askPrice))
        askS_list.append(float(askSize))
    return time,bidP_list,bidS_list,askP_list,askS_list
        #frame = pd.concat(list_)
    #print(list_)


    

In [3]:

    

def fearure_creation(time,bidP_list,bidS_list,askP_list,askS_list,length):
        #Initialising
    close_list=[]
    price_list=[]
    for i in range(1, len(bidS_list)-1):
        close_list=(bidP_list[i]+askP_list[i])/2
        price_list=(bidP_list*bidS_list+askP_list*askS_list)/(bidS_list+askS_list)
    return close_list,price_list


    

In [38]:

    

#path = 'C:\\Users\Michal\Dropbox\IB_data'
path = 'C:\\Users\Michal\Desktop'+ '\SPY31July17.csv'
time,bidP_list,bidS_list,askP_list,askS_list=read_data(path)
#close_list,price_list=fearure_creation(time,bidP_list,bidS_list,askP_list,askS_list,20)


    

    




---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-38-7e03a4b7dc08> in <module>()
      1 #path = 'C:\\Users\Michal\Dropbox\IB_data'
      2 path = 'C:\\Users\Michal\Desktop'+ '\SPY31July17.csv'
----> 3 time,bidP_list,bidS_list,askP_list,askS_list=read_data(path)
      4 #close_list,price_list=fearure_creation(time,bidP_list,bidS_list,askP_list,askS_list,20)

<ipython-input-37-f2326ba024d5> in read_data(path)
     14         df = pd.read_csv(file_,index_col=None, header=0)
     15         list_.append(df)
---> 16         l=list_.split(',')
     17         all_features.append(l)
     18         time,askPrice,askSize,bidPrice,bidSize = l

AttributeError: 'list' object has no attribute 'split'

In [15]:

    

#time,bidP_list,bidS_list,askP_list,askS_list


    

In [27]:

    

#how to avoid \n ?? filtered=(line.replace() for line in myfile)