def testRFC822(self):
# rfc822 conversion
dt = DateTime('2002-05-02T08:00:00+00:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 +0000')
dt = DateTime('2002-05-02T08:00:00+02:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 +0200')
dt = DateTime('2002-05-02T08:00:00-02:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 -0200')
# Checking that conversion from local time is working.
dt = DateTime()
dts = dt.rfc822().split(' ')
times = dts[4].split(':')
_isDST = time.localtime(time.time())[8]
if _isDST:
offset = time.altzone
else:
offset = time.timezone
self.assertEqual(dts[0], dt.aDay() + ',')
self.assertEqual(int(dts[1]), dt.day())
self.assertEqual(dts[2], dt.aMonth())
self.assertEqual(int(dts[3]), dt.year())
self.assertEqual(int(times[0]), dt.h_24())
self.assertEqual(int(times[1]), dt.minute())
self.assertEqual(int(times[2]), int(dt.second()))
self.assertEqual(dts[5], "%+03d%02d" % divmod((-offset / 60), 60))
def testRFC822(self):
# rfc822 conversion
dt = DateTime('2002-05-02T08:00:00+00:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 +0000')
dt = DateTime('2002-05-02T08:00:00+02:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 +0200')
dt = DateTime('2002-05-02T08:00:00-02:00')
self.assertEqual(dt.rfc822(), 'Thu, 02 May 2002 08:00:00 -0200')
# Checking that conversion from local time is working.
dt = DateTime()
dts = dt.rfc822().split(' ')
times = dts[4].split(':')
_isDST = time.localtime(time.time())[8]
if _isDST:
offset = time.altzone
else:
offset = time.timezone
self.assertEqual(dts[0], dt.aDay() + ',')
self.assertEqual(int(dts[1]), dt.day())
self.assertEqual(dts[2], dt.aMonth())
self.assertEqual(int(dts[3]), dt.year())
self.assertEqual(int(times[0]), dt.h_24())
self.assertEqual(int(times[1]), dt.minute())
self.assertEqual(int(times[2]), int(dt.second()))
self.assertEqual(dts[5], "%+03d%02d" % divmod((-offset / 60), 60))
def len_month(self, date):
""" fetches a monthname and returns translation plus lenght """
if type(date)== StringType:
date = DateTime(date)
month = date.aMonth().lower()
label = 'month_'+month
tmonth = pmf(label)
return (tmonth, len(tmonth))
def get_time(self):
""" Get timestamp at the initiation of task
Returns:
[string]: [time at task initiation]
"""
current_time = DateTime(time.time(), "US/Eastern")
self.timestamp = (str(current_time.year()) +
str(current_time.aMonth()) +
str(current_time.day()) + str(current_time.h_24()) +
str(current_time.minute()) + str(time.time())[:2])
def init(dirt,
task,
runlist,
timelist,
foldlist,
rep,
task_token,
fitmetrics=False):
current_time = DateTime(time.time(), "US/Eastern")
if task == "bt":
taskname = 'binaryTarget'
if not fitmetrics:
fitmetrics = autosklearn.metrics.log_loss
elif task == "bre":
taskname = 'binaryRareEvent'
if not fitmetrics:
fitmetrics = autosklearn.metrics.log_loss
elif task == "it":
taskname = 'intervalTarget'
if not fitmetrics:
fitmetrics = autosklearn.metrics.mean_squared_error
sasdatalist = sorted(glob.glob(dirt + taskname + "/*sas7bdat"),
key=os.path.getsize)
print(sasdatalist)
sasdatalist = remove_dirt(sasdatalist, dirt + "/" + taskname + "/")
print(sasdatalist)
sasdatalist = [i[:-9] for i in sasdatalist]
print("sas datalist\n", sasdatalist)
#sasdatalist = sorted(sasdatalist)
csvdatalist = sorted(glob.glob(dirt + taskname + "/*.csv"),
key=os.path.getsize)
csvdatalist = remove_dirt(csvdatalist, dirt + "/" + taskname + "/")
csvdatalist = [i[:-4] for i in csvdatalist]
print("csv datalist\n", csvdatalist)
#csvdatalist = sorted(csvdatalist)
#glob.glob(dirt + taskname+"/*sas7bdat")
metalist = glob.glob(dirt + "/tmp_metadata/*meta.csv")
metalist = remove_dirt(metalist, dirt + "/tmp_metadata/")
metalist = [i[:-9] for i in metalist]
print("working dirt\t", dirt)
print("metadatalit\n", metalist)
#metalist = sorted(metalist)
timestamp = (str(current_time.year()) + str(current_time.aMonth()) +
str(current_time.day()) + str(current_time.h_24()) +
str(current_time.minute()) + str(time.time())[:2])
logfile = open(
"results/log_" + str(len(runlist)) + "dataset_" + str(timelist[0]) +
"s_" + str(foldlist[0]) + "f_rep" + str(rep) + "_task_" +
str(task_token) + ".txt",
"w",
)
return dirt, taskname, logfile, csvdatalist, sasdatalist, metalist, timestamp, fitmetrics,
def fromLineFrom(self,email,date):
"""
Generate a conformant mbox From line from email and date strings.
(unless date is unparseable, in which case we omit that part)
"""
# "email" is in fact a real name or zwiki username - adapt it
email = re.sub(r'\s','',email) or 'unknown'
try:
d = DateTime(date)
return 'From %s %s %s %d %02d:%02d:%02d %s %d\n' % (
email,d.aDay(),d.aMonth(),d.day(),d.hour(),
d.minute(),d.second(),d.timezone(),d.year())
except (DateTimeSyntaxError,AttributeError,IndexError):
return 'From %s\n' % email
warnings.simplefilter("ignore")
current_time = DateTime(time.time(), 'US/Eastern')
###################################################################
# Use sklearn to holdout, read in original data
#######################################################################
framework = 'autosklearn'
datasetn = 'bankmarketing'
foldn = 3
timeforjob = foldn * 2 * 1800
prepart = True
ncore = 4
dirt = '/root/data/'
############################################################################################################
resultfile = str(datasetn)+str(foldn) +"fold"+ str(timeforjob) + "seconds" + str(ncore)+"core"+\
str(current_time.year()) + str(current_time.aMonth())+ str(current_time.day()) + \
str(current_time.h_24()) + str(current_time.minute()) + str(time.time())[:2] + str(framework)+'prepart.txt'
dataset = "uci_bank_marketing_pd"
numeric_features = []
categorical_features = []
def prep(dataset,
dirt,
numeric_features,
categorical_features,
delim=',',
indexdrop=False):
index_features = ['_dmIndex_', '_PartInd_']
data = pd.read_csv(dirt + dataset + '.csv',
delimiter=delim) # panda.DataFrame
