def dataframe(self, timeframe, seed, dformat, dbs_extra, newdata=None):
"""Form a dataframe from various CMS data-providers"""
dtypes, stypes, rtypes, tiers = self.data_types()
pop_datasets = 0
dbs_datasets = 0
popdb_results = [
r for r in self.popdb.dataset_stat(timeframe[0], timeframe[1])
]
if dformat == 'csv':
dataset = None
for row in popdb_results:
if len(row['dataset'].split(
'/')) == 4: # dataset with 3 slashes
dataset = row['dataset']
break
if not dataset:
raise Exception(
"Unable to find valid dataset name in popdb output")
target = dict(naccess=row['naccess'],
nusers=row['nusers'],
totcpu=row['totcpu'],
rnaccess=row['rnaccess'],
rnusers=row['rnusers'],
rtotcpu=row['rtotcpu'])
# seed dataset to determine headers of the dataframe
rows = self.dataset_info(timeframe, seed, dtypes, stypes, rtypes,
tiers, 'headers', target)
headers = [r for r in rows][0]
yield ','.join(headers)
tstamp = time.strftime("%Y-%m-%d %H:%M:%S GMT", time.gmtime())
if newdata: # request new dataset
if self.verbose:
print("Generate dataframe for new datasets", tstamp)
n_days = 7
if timeframe:
n_days = ndays(yyyymmdd(timeframe[0]), yyyymmdd(timeframe[1]))
new_datasets = self.dbs.new_datasets(n_days)
target = dict(naccess=0,
nusers=0,
totcpu=0,
rnaccess=0,
rnusers=0,
rtotcpu=0)
for row in new_datasets:
dataset = row['dataset']
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
for row in rows:
yield row
return
# get list of popular datasets in certain time frame
# popdb_results = [r for r in self.popdb.dataset_stat(timeframe[0], timeframe[1])]
popdb_datasets = {} #
for row in popdb_results:
dataset = row['dataset']
if not DATASET_PAT.match(dataset):
continue
if self.verbose:
print("Generate dataframe for %s, timeframe: %s, %s" \
% (dataset, timeframe, tstamp))
target = dict(naccess=row['naccess'],
nusers=row['nusers'],
totcpu=row['totcpu'],
rnaccess=row['rnaccess'],
rnusers=row['rnusers'],
rtotcpu=row['rtotcpu'])
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
popdb_datasets[dataset] = row
for row in rows:
yield row
pop_datasets += 1
# get list of datasets from DBS and discard from this list
# those who were presented in popdb
all_dbs_datasets = self.dbs.datasets()
dbsdatasets = [
d for d in all_dbs_datasets if d not in popdb_datasets.keys()
]
target = dict(naccess=0,
nusers=0,
totcpu=0,
rnaccess=0,
rnusers=0,
rtotcpu=0)
for dataset in random.sample(dbsdatasets, dbs_extra):
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
for row in rows:
yield row
dbs_datasets += 1
if self.verbose:
print("DBS datasets : %s" % dbs_datasets)
print("PopDB datasets: %s out of %s" %
(pop_datasets, len(popdb_results)))
def dataframe(self, timeframe, seed, dformat, dbs_extra, newdata=None):
"""Form a dataframe from various CMS data-providers"""
dtypes, stypes, rtypes, tiers = self.data_types()
pop_datasets = 0
dbs_datasets = 0
popdb_results = [r for r in self.popdb.dataset_stat(timeframe[0], timeframe[1])]
if dformat == 'csv':
row = popdb_results[0]
dataset = row['dataset']
target = dict(naccess=row['naccess'],nusers=row['nusers'],totcpu=row['totcpu'],
rnaccess=row['rnaccess'],rnusers=row['rnusers'],rtotcpu=row['rtotcpu'])
# seed dataset to determine headers of the dataframe
rows = self.dataset_info(timeframe, seed, dtypes, stypes, rtypes,
tiers, 'headers', target)
headers = [r for r in rows][0]
yield ','.join(headers)
tstamp = time.strftime("%Y-%m-%d %H:%M:%S GMT", time.gmtime())
if newdata: # request new dataset
if self.verbose:
print("Generate dataframe for new datasets", tstamp)
n_days = 7
if timeframe:
n_days = ndays(yyyymmdd(timeframe[0]), yyyymmdd(timeframe[1]))
new_datasets = self.dbs.new_datasets(n_days)
target = dict(naccess=0,nusers=0,totcpu=0,
rnaccess=0,rnusers=0,rtotcpu=0)
for row in new_datasets:
dataset = row['dataset']
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
for row in rows:
yield row
return
# get list of popular datasets in certain time frame
# popdb_results = [r for r in self.popdb.dataset_stat(timeframe[0], timeframe[1])]
popdb_datasets = {} #
for row in popdb_results:
dataset = row['dataset']
if not DATASET_PAT.match(dataset):
continue
if self.verbose:
print("Generate dataframe for %s, timeframe: %s, %s" \
% (dataset, timeframe, tstamp))
target = dict(naccess=row['naccess'],nusers=row['nusers'],totcpu=row['totcpu'],
rnaccess=row['rnaccess'],rnusers=row['rnusers'],rtotcpu=row['rtotcpu'])
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
popdb_datasets[dataset] = row
for row in rows:
yield row
pop_datasets += 1
# get list of datasets from DBS and discard from this list
# those who were presented in popdb
all_dbs_datasets = self.dbs.datasets()
dbsdatasets = [d for d in all_dbs_datasets if d not in popdb_datasets.keys()]
target = dict(naccess=0,nusers=0,totcpu=0,
rnaccess=0,rnusers=0,rtotcpu=0)
for dataset in random.sample(dbsdatasets, dbs_extra):
rows = self.dataset_info(timeframe, dataset, dtypes, stypes, \
rtypes, tiers, dformat, target)
for row in rows:
yield row
dbs_datasets += 1
if self.verbose:
print("DBS datasets : %s" % dbs_datasets)
print("PopDB datasets: %s out of %s" % (pop_datasets, len(popdb_results)))
def test_ndays(self):
"Test ndays function"
time1, time2 = '20141120', '20141124'
result = ndays(time1, time2)
expect = 4
self.assertEqual(expect, result)
