def rank(self, stocks):
"""Ranks a collection of stocks given an end date and a start date
based on a certain objective function.
Args:
stocks: list of ticker symbols
Returns:
an array of ordered stocks according to rank with some information
as to why they were ranked a certain way.
"""
def compareF(item1, item2):
return int((item1['info']['total_return'] * 100) - (item2['info']['total_return'] * 100))
stockInfos = []
for stock in stocks:
tbl = YahooStockTable.make_data_table(stock, start_date, end_date)
#print tbl.table
try:
infoTuple = self.getInformation(tbl)
except Exception:
#print tbl.table
continue
sInfo = {}
sInfo['info'] = infoTuple
sInfo['stock_name'] = stock
stockInfos.append(sInfo)
stockInfos.sort(compareF)
return stockInfos
def main_fundamentals(stocks, start_year):
start_date = DateUtils.getDateStr(start_year, 1, 2)
end_date = DateUtils.getDateStr(start_year, 1, 3)
for stock in stocks:
dt = YahooStockTable.make_data_table(stock, start_date, end_date)
print dt.header
print dt.table
return ''
def get_return(self, stock):
"""Gets the return on a particular stock for the particular
start_date and end_date.
If one had a 5% gain, this will return 1.05
If one had a 10% loss, this will return 0.9
"""
ytable = YahooStockTable.make_data_table(stock, self.start_date, self.end_date)
closing_values = ytable.getAllClosing()
return StockUtils.getTotalReturn(closing_values)
def makeMonthlyTables(ticker_symbol, year):
"""Provide monthly (12 months) for a particular year for a particular
ticker symbol.
Ex: GOOG, 2012 -> tables for Jan 2012, Feb 2012, ..., Dec 2012
"""
months = [month for month in range(1, 13)]
monthly_tables = []
for month in months:
# Assume we go from the 1st of the month, and then 30 days later.
start_date = DateUtils.getDateStr(year, month, 1)
end_date = DateUtils.getEndDate(start_date, 30)
tbl = YahooStockTable.make_data_table(ticker_symbol, start_date, end_date)
monthly_tables.append(tbl)
return monthly_tables
def get_correlation(self, stocks):
"""Returns a correlation matrix defined by numpy.corrcoef.
Basically if you pass in 3 stocks (0, 1, 2)
Then corr[0][2] or corr[2][0] will tell the correlation of 0 and 2.
Correlation is: Var(A * B) / Stdev(A) * Stdev(B)
"""
lists_of_changes = []
for stock in stocks:
tbl = YahooStockTable.make_data_table(stock, self.start_date, self.end_date)
closing = tbl.getAllClosing()
pct_changes = StockUtils.getPercentChanges(closing)
lists_of_changes.append(pct_changes)
for changes in lists_of_changes:
logging.info(self.fformat(changes))
corr = numpy.corrcoef(lists_of_changes)
return corr # correlation matrix
def get_rebal_values(self, stock, freq_days):
"""What are my closing values on rebalancing days.
"""
ytable = YahooStockTable.make_data_table(stock, self.start_date, self.end_date)
closing_values = ytable.getAllClosing()
total_days = len(closing_values)
print 'Total num days: %d' % total_days
# Construct closing values after each end of frequency days.
num_realloc = total_days / freq_days
# Construct closing_values for only the days you start investing and when you reallocate
# your money.
realloc_day = 0
rebal_values = []
for ind in range(0, num_realloc + 1):
# To ensure we stay within bounds of days.
day = min(realloc_day, total_days-1)
print 'Reallocating at day: ' + str(day)
rebal_values.append(closing_values[day])
realloc_day += freq_days
return rebal_values
def makeYearlyTable(ticker_symbol, year): """Provide year data.""" start_date = DateUtils.getDateStr(year, 1, 1) end_date = DateUtils.getDateStr(year, 12, 30) tbl = YahooStockTable.make_data_table(ticker_symbol, start_date, end_date) return tbl
