# load stock-signal.py # load it by manual

# load load-code.py # load it by manual

import StockSignal

import LoadCode

import time

import random

import pandas

import pp

stocks=LoadCode.load_code_ex()

count=stocks[0].count()

data=pandas.DataFrame()

skip_before=''

skip=len(skip_before)

summary=None

result_file='candidates-%s.csv' % pandas.datetime.now().strftime('%Y-%m-%d')

def local_func(stock, code, name):

return None

ppservers = ()

jobs = []

start_time = time.time()

# Total cpus=4, 8 performance the best among [4,8,12,16,24]

job_server = pp.Server(8, ppservers=ppservers)

for i in range(count):

if stocks[0][i].find(skip_before)==0:

skip=0

if skip > 0:

continue;

jobs.append(job_server.submit(local_func, (stocks[0][i], stocks[0][i], stocks[1][i]), (), ("StockSignal", "pandas", )))

for job in jobs:

ret = job()

if isinstance(ret, type(None)):

continue

if not isinstance(summary, type(None)):

summary=pandas.concat([ret, summary])

else:

summary=ret

print "Time elapsed: ", time.time() - start_time, "s"

job_server.print_stats()

if not isinstance(summary, type(None)):

summary=summary.sort(['profit'])

summary.to_csv(result_file)

print summary[['code', 'signal', 'Volume', 'buy', 'sell', 'profit', 'name']].sort_values(['signal'])

# # select those have been sold

# sold=summary2.select(lambda x: True if summary2.loc[x]['signal']==-1 else False)

# holding=summary2.select(lambda x: True if summary2.loc[x]['signal']==1 else False)

# print sold.to_string()

# print holding.to_string()