def test_histData():
# one symbol
hData = yf.getHistoricData(symbols[0], dumpDest="debug")
hData.pct_change(
) # try a calculation. Will result in an error if data is rotten.
assert hData.shape[1] == 6
assert hData.shape[0] > 100
hData = yf.getHistoricData(symbols)
def SMA(symb,intvs,symbec='STK',eday=None):
if eday==None:
eday = datetime.now(timezone('US/Eastern'))
bday=eday-timedelta(days=365)
symbDict={'INDU':'^DJI','COMP':'^IXIC','SPX':'^GSPC','RUT':'^RUT'}
if symb in symbDict.keys():
symb=symbDict[symb]
if symb=='^DJI':
Data = Quandl.get("YAHOO/INDEX_DJI", authtoken='977FGLXLZ1wsdKt8DgUH',
trim_start="%d-%02d-%02d" %(bday.year,bday.month,bday.day),
trim_end="%d-%02d-%02d" %(eday.year,eday.month,eday.day))
sums=list()
for n in intvs:
if len(Data.index)<n: #no enough data
sums.append(-1)
else:
sums.append(round(Data['Adjusted Close'][-n-1:-1].mean(),1))
else:
Data = yf.getHistoricData(symb,bday.timetuple()[0:3],eday.timetuple()[0:3])[['adj_close']] # get data from yahoo finance
print Data.tail()
sums=list()
for n in intvs:
if len(Data.index)<n: #no enough data
sums.append(-1)
else:
sums.append(round(Data['adj_close'][-n-1:-1].mean(),1))
return sums
def getTradingDates():
''' get list of trading dates from yahoo, one year back'''
startDate = (datetime.date.today() -
datetime.timedelta(days=365)).timetuple()[:3]
df = yf.getHistoricData('SPY', sDate=startDate)
dates = [d.date() for d in df.index]
return dates
def Pivots(symb,symbec='STK',eday=None):
if eday== None:
eday = datetime.now(timezone('US/Eastern'))
bday=eday-timedelta(days=10)
symbDict={'INDU':'^DJI','COMP':'^IXIC','SPX':'^GSPC','RUT':'^RUT'}
if symb in symbDict.keys():
symb=symbDict[symb]
if symb=='^DJI':
Data = Quandl.get("YAHOO/INDEX_DJI", authtoken='977FGLXLZ1wsdKt8DgUH',
trim_start="%d-%02d-%02d" %(bday.year,bday.month,bday.day),
trim_end="%d-%02d-%02d" %(eday.year,eday.month,eday.day))[0:1]
H=Data['High'].values[-1]
L=Data['Low'].values[-1]
C=Data['Close'].values[-1]
else:
Data = yf.getHistoricData(symb,bday.timetuple()[0:3],eday.timetuple()[0:3])[['high','low','close','adj_close']] # get data from yahoo finance
H=Data['high'].values[-1]
L=Data['low'].values[-1]
C=Data['close'].values[-1]
P = (H + L + C) / 3
#P = (H + L + C + C) / 4
#P = (H + L + O + O) / 4
R1 = 2*P-L
S1 = 2*P-H
R2 = P+(H-L)
S2 = P-(H-L)
R3 = R1+(H-L)
S3 = S1-(H-L)
return [round(R3,1),round(R2,1),round(R1,1),round(P,1),round(S1,1),round(S2,1),round(S3,1)]
def Pivots(symb, symbec='STK', eday=None):
if eday == None:
eday = datetime.now(timezone('US/Eastern'))
bday = eday - timedelta(days=10)
symbDict = {'INDU': '^DJI', 'COMP': '^IXIC', 'SPX': '^GSPC', 'RUT': '^RUT'}
if symb in symbDict.keys():
symb = symbDict[symb]
if symb == '^DJI':
Data = Quandl.get(
"YAHOO/INDEX_DJI",
authtoken='977FGLXLZ1wsdKt8DgUH',
trim_start="%d-%02d-%02d" % (bday.year, bday.month, bday.day),
trim_end="%d-%02d-%02d" % (eday.year, eday.month, eday.day))[0:1]
H = Data['High'].values[-1]
L = Data['Low'].values[-1]
C = Data['Close'].values[-1]
else:
Data = yf.getHistoricData(symb,
bday.timetuple()[0:3],
eday.timetuple()[0:3])[[
'high', 'low', 'close', 'adj_close'
]] # get data from yahoo finance
H = Data['high'].values[-1]
L = Data['low'].values[-1]
C = Data['close'].values[-1]
P = (H + L + C) / 3
#P = (H + L + C + C) / 4
#P = (H + L + O + O) / 4
R1 = 2 * P - L
S1 = 2 * P - H
R2 = P + (H - L)
S2 = P - (H - L)
R3 = R1 + (H - L)
S3 = S1 - (H - L)
return [
round(R3, 1),
round(R2, 1),
round(R1, 1),
round(P, 1),
round(S1, 1),
round(S2, 1),
round(S3, 1)
]
@author: krisan
"""
#the following script aim to investigate overnight trading
#https://code.google.com/p/trading-with-python/source/browse/trunk/lib/functions.py
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import neighbors
import tradingWithPython.lib.yahooFinance as yf
import tradingWithPython.lib as lb1
#https://www.tradingview.com/v/EVXQPaR9/
# we first need ot get data.
ohlc=yf.getHistoricData('SPY')
ohlc['cc']=(ohlc['close']/ohlc['open'])-1
ohlc['co']=(ohlc['open']/ohlc['close'].shift(1))-1
#ret1=np.cumsum(ohlc['cc']).plot()
##ret2=np.cumsum(ohlc['co']).plot()
plt.plot(ohlc['co'].index,np.cumsum((ohlc['cc'])),'r')
plt.plot(ohlc['co'].index,np.cumsum(ohlc['co']),'b')
train=pd.DataFrame(columns=ohlc.columns)
test=pd.DataFrame(columns=ohlc.columns)
train=train.append(ohlc.iloc[0:1000])
test=test.append(ohlc.iloc[1000:])
today = dt.date.today().strftime('%m/%d/%Y')
now = dt.datetime.now()
tz = os.environ['TZ'] = 'America/New_York' # local timezone
stock_count = len(symbols)
stock_size = int(stock_count / 2)
# create series of all tradings days within range
us_bd = offsets.CustomBusinessDay(calendar=holiday.USFederalHolidayCalendar())
bdays = pd.Series(pd.date_range(start, today, freq=us_bd, name='Date'))
print(tz + '\n' + now.strftime('%I:%M:%S%p') + ' ' + today + '\n\n' +
'Downloading Data for %s Stocks...\n' % stock_count)
try:
stocks = {s: yf.getHistoricData(s, start) for s in symbols}
stocks = {s: tech.analyze(stocks[s]) for s in symbols}
except Exception as e:
pass
elapsed = dt.datetime.now() - now
print(('\n' 'Completed Successfully in %s') % elapsed)
for s in symbols:
stocks[s]["Date"] = stocks[s].index.date
stocks[s].set_index("Date", drop=True, inplace=True)
close = pd.DataFrame({s: stocks[s]['adj_close'] for s in symbols})
first_close = {s: stocks[s]['adj_close'].head(1).transpose() for s in symbols}
last_close = close.tail(1).transpose()
daily_return = close.apply(lambda x: x / x[0])
'VZ', 'T', 'S', 'TMUS', 'WFC', 'ORCL', 'JNJ', 'MSFT', 'DIS', 'WMT', 'MCD',
'TGT', 'COKE', 'FB', 'AAPL', 'SNAP', 'INTC', 'HK'
]
# symbols = fdb.symbols[8:16] # pull symbols directly from database
start = 2000 / 1 / 1 # set starting date for historic data
tz = os.environ['TZ'] = 'America/New_York' # set local timezone
today = dt.date.today() # current day
now = dt.datetime.now() # current time
count = len(symbols) # ticker list size
print("Python: {}\n\n".format(sys.version), tz, '\n',
now.strftime('%I:%M:%S%p'), ' ', today.strftime('%m/%d/%Y'), '\n\n',
'Downloading Data for %s Stocks...\n' % count)
stocks = {s: yf.getHistoricData(s, start)
for s in symbols} # fetch price data for given tickers
stocks = {s: tech.analyze(stocks[s])
for s in symbols} # append technical indicators
for s in symbols: # format date (index) column
stocks[s]["Date"] = stocks[s].index.date
stocks[s].set_index("Date", drop=True, inplace=True)
elapsed = dt.datetime.now() - now # calculate total execution time
print(('\n' 'Completed Successfully in %s\n') % elapsed)
close = pd.DataFrame({s: stocks[s]['adj_close']
for s in symbols}) # all close data
price = {s: close.loc[:, s]
for s in symbols} # series of all close data for each stock
stock_count = len(symbols)
stock_size = int(stock_count / 2)
# create series of all tradings days within range
us_bd = offsets.CustomBusinessDay(calendar=holiday.USFederalHolidayCalendar())
bdays = pd.Series(pd.date_range(start, today, freq=us_bd, name='Date'))
print(tz + '\n' + time + ' ' + today.strftime('%m/%d/%Y') + '\n\n' +
'Downloading Data for %s Stocks...\n' % stock_count)
stocks = {}
close_p = {}
first_close = {}
for s in symbols:
try:
stocks.update({s: yf.getHistoricData(s, start)})
stocks.update({s: tech.analyze(stocks[s])})
stocks[s]["Date"] = stocks[s].index.date
stocks[s].set_index("Date", drop=True, inplace=True)
close = pd.DataFrame(close_p.update({s: stocks[s]['adj_close']}))
first_close.update({s: stocks[s]['adj_close'].head(1).transpose()})
except:
pass
#stocks = {s: yf.getHistoricData(s, start) for s in symbols}
#stocks = {s: tech.analyze(stocks[s]) for s in symbols}
#close = pd.DataFrame({s: stocks[s]['adj_close'] for s in symbols})
#first_close = {s: stocks[s]['adj_close'].head(1).transpose() for s in symbols}
closefill = close.fillna(1)
last_close = close.tail(1).transpose()
def lastTradingDate():
""" determine last historic data date by downloading from yahoo finance"""
data = yf.getHistoricData('SPY', sDate=(2016, 1, 1))
return data.index[-1]
import matplotlib
import matplotlib.pyplot as plt
import quadprog as qp
import numpy as np
start_date = tuple([int(strftime("%Y"))-2,int(strftime("%m")),int(strftime("%d"))])
tran_days = 1 #interval between tranch dates
tran_look = 20 #how far back to tranch
equity_list = ['VTI','VGK', 'VPL','VWO','VNQ', 'VNQI']
bond_list = ['BIV','BLV','BNDX','VWOB','VTIP', 'BSV']
symbol_list = equity_list + bond_list
# Get historical data information
prices = yf.getHistoricData(symbol_list,sDate=start_date)
# Calculate number of dates have
num_pts = len(prices.major_axis)
# Create rankings based on tranch settings
ranking = pd.DataFrame()
for tranch in range(0,tran_look+1,tran_days):
ranking[str(tranch)] = (prices.iloc[:,num_pts-1].loc['adj_close']
- prices.iloc[:,num_pts-1].loc['adj_close'])
for num in range(126,256,1):
current_price = prices.iloc[:,num_pts-1-tranch].loc['adj_close']
past_price = prices.iloc[:,num_pts-1-num-tranch].loc['adj_close']
ranking[str(tranch)] += (current_price - past_price)/past_price
# Find highest x funds and the largest bond fund
! pip install tradingWithPython
import pandas as pd
import tradingWithPython as twp
assert twp.__version__ >= '0.0.12' , 'Please update your twp module '
import tradingWithPython.lib.yahooFinance as yahoo # yahoo finance module
price = yahoo.getHistoricData('SPY')['adj_close'][-2000:]
price.plot()
figsize(10,6)
# sumulate two trades
signal = pd.Series(index=price.index) #init signal series
# while setting values this way, make sure that the dates are actually in the index
signal[pd.datetime(2008,1,3)] = -10000 # go short
signal[pd.datetime(2009,1,5)] = 10000 # go long 10k$ on this day
signal[pd.datetime(2013,1,8)] = 0 # close position
bt = twp.Backtest(price,signal,initialCash=0)
bt.plotTrades()
figure()
bt.pnl.plot()
title('pnl')
figure()
bt.data.plot()
title('all strategy data')