def main():
from dpark import DparkContext, optParser
#optParser.set_default('master', 'mesos')
optParser.add_option('-e', '--query', type='string', default='',
help='execute the SQL qeury then exit')
options, args = optParser.parse_args()
load_history()
if options.query:
execute(options.query)
sys.exit(0)
shell()
# -*- coding: utf-8 -*-
import uuid
import inspect
from dpark import DparkContext, optParser
from sql.parser import parse
from collections import OrderedDict
from models import Model
optParser.add_option("-s") # "option used for py.test"
optParser.add_option("-x")
class Table(object):
dialect = "excel"
columns = ()
def __init__(self, name, paths=None, columns=None, query=None):
self.name = name
self.columns = OrderedDict(columns or self.__class__.columns)
self.paths = paths or []
self.query = query
def index(self, field):
return self.columns.keys().index(field)
def rdd(self, dpark=None):
if self.query:
return self.query.rdd
f.write('\n'.join(comb_ad_context))
dtest_sample_temp = xgb.DMatrix(temp_dir + '.libsvm')
ypred.append(' '.join(dp.parallelize(bst.predict(dtest_sample_temp)).map(lambda x:str(x)).collect()))
with open(mix_dir + '.txt', 'w') as f:
f.write('\n'.join(ypred))
check_call('rm -rf %s' % temp_dir + '.libsvm', shell=True)
## plot histogram
#plt.hist(ypred_sample_temp,10)
#plot_path = '/home2/songsiyu/data/models_%s/%s' % (options.feature_domain, model_date_str) + '/plots'
#if not os.path.exists(plot_path):
# check_call('mkdir %s' % plot_path, shell=True)
#plt.savefig(plot_path + '/ad%d' % curr_ad)
if __name__ == '__main__':
optParser.add_option('--model_version', dest='model_version') ##input=yesterday
optParser.add_option('--feature_domain', dest='feature_domain')
options, _ = optParser.parse_args()
dp = DparkContext()
if not options.model_version:
model_date_str = (datetime.today() - timedelta(1)).strftime('%Y%m%d') ##yestoday
else:
model_date_str = options.model_version
# train
logger.info('mixing %s' % options.feature_domain)
_mix(dp, options.feature_domain, model_date_str)
logger.info('mix.py done!')
if s.lower().startswith(c):
arg = s[len(c):].strip()
getattr(self,c)(arg)
self.sql = ''
continue
if not self.sql.rstrip().endswith(';'):
continue
self.run_sql()
except Exception, e:
import traceback; traceback.print_exc()
self.sql = ''
if __name__ == '__main__':
from dpark import optParser
optParser.set_default('master', 'flet6')
optParser.add_option('-e', '--query', type='string', default='',
help='execute the SQL qeury then exit')
optParser.add_option('-s', '--script', type='string', default='',
help='execute the SQL script file then exit')
options, args = optParser.parse_args()
console = Console()
if options.query:
console.run_script(options.query)
elif options.script:
with open(options.script) as f:
console.run_script(f.read())
else:
console.run()
f.seek(0)
f2 = open(outfile, 'w')
for line in f:
l = line.strip('\n').split(sp)
f2.write(l[0])
f2.write(sp)
l = l[1:]
for i in xrange(len(minlst)):
if maxlst[i] == minlst[i]:
stf = maxlst[i]
f2.write(str(stf))
if i != len(minlst) - 1:
f2.write(sp)
else:
stf = (float(l[i]) - minlst[i]) / (maxlst[i] - minlst[i])
f2.write(str(stf))
if i != len(minlst) - 1:
f2.write(sp)
f2.write('\n')
if __name__ == '__main__':
from dpark import optParser
optParser.add_option('--in', dest='inputf')
optParser.add_option('--out', dest='outputf')
options, args = optParser.parse_args()
inf = options.inputf
outf = options.outputf
norm(inf, outf)
import os, sys
from subprocess import check_call, call
from datetime import timedelta, datetime
from dpark import DparkContext, optParser
import random
import math
from config import PLAN_SETTINGS, logger
import ctr_rdds
from common.util import GeneralMap, is_spider
from db_tools import DBTools
sys.path.append('%s/user_profile' % os.path.dirname(os.path.realpath(__file__)))
optParser.add_option("--date", dest="date")
optParser.add_option("--collection", dest="collection")
options, _ = optParser.parse_args()
dp = DparkContext()
def feature_gen(current_date, collection):
db_tools = DBTools()
if collection == 'CPC':
AD_TO_TAGS = dp.broadcast(db_tools.get_ad_to_tags())
elif collection == 'market':
AD_TO_TAGS = dp.broadcast(db_tools.get_market_ad_to_features())
else:
raise Exception
AD_TO_ITEM = dp.broadcast(db_tools.get_ad_to_item()) ## extract from db
AD_TO_ORDER = dp.broadcast(db_tools.get_ad_to_order())
AD_TO_ACCOUNT = dp.broadcast(db_tools.get_ad_to_account())
AD_TO_UNIT = dp.broadcast(db_tools.get_ad_to_units())
f.seek(0)
f2 = open(outfile, 'w')
for line in f:
l = line.strip('\n').split(sp)
f2.write(l[0])
f2.write(sp)
l = l[1:]
for i in xrange(len(minlst)):
if maxlst[i] == minlst[i]:
stf = maxlst[i]
f2.write(str(stf))
if i != len(minlst) - 1:
f2.write(sp)
else:
stf = (float(l[i]) - minlst[i]) / (maxlst[i] - minlst[i])
f2.write(str(stf))
if i != len(minlst) - 1:
f2.write(sp)
f2.write('\n')
if __name__ == '__main__':
from dpark import optParser
optParser.add_option('--in', dest='inputf')
optParser.add_option('--out', dest='outputf')
options, args = optParser.parse_args()
inf = options.inputf
outf = options.outputf
norm(inf, outf)
