def add_datasets(self, input_prim_proc_acq_tier_config, number=1):
"Add blocks to a given primary dataset"
idict = deepcopy(input_prim_proc_acq_tier_config)
prim_val = idict['prim_ds']
proc_ver = idict['processing_era']
acq_era = idict['acquisition_era']
tier = idict['tier']
config = idict['configs']
func = lambda x: isinstance(x, dict) and [x] or x
prim_val = func(prim_val)
proc_ver = func(proc_ver)
acq_era = func(acq_era)
tier = func(tier)
config = func(config)
output = []
for item_prim, item_proc, item_acq, item_tier, item_config \
in zip(prim_val, proc_ver, acq_era, tier, config):
prim = item_prim['primary_ds_name']
acq = item_acq['acquisition_era_name']
tier = item_tier['data_tier_name']
proc_ver = item_proc['processing_version']
attrs = {'prim':prim,
'processing_version':proc_ver,
'acquisition_era_name':acq,
'tier':tier,
'output_configs':[item_config]}
res = self.datasets(number, **attrs)
for row in res:
output.append(row['dataset'])
idict['dataset'] = output
return idict
def add_datasets(self, input_prim_proc_acq_tier_config, number=1):
"Add blocks to a given primary dataset"
idict = deepcopy(input_prim_proc_acq_tier_config)
prim_val = idict['prim_ds']
proc_ver = idict['processing_era']
acq_era = idict['acquisition_era']
tier = idict['tier']
config = idict['configs']
func = lambda x: isinstance(x, dict) and [x] or x
prim_val = func(prim_val)
proc_ver = func(proc_ver)
acq_era = func(acq_era)
tier = func(tier)
config = func(config)
output = []
for item_prim, item_proc, item_acq, item_tier, item_config \
in zip(prim_val, proc_ver, acq_era, tier, config):
prim = item_prim['primary_ds_name']
acq = item_acq['acquisition_era_name']
tier = item_tier['data_tier_name']
proc_ver = item_proc['processing_version']
attrs = {
'prim': prim,
'processing_version': proc_ver,
'acquisition_era_name': acq,
'tier': tier,
'output_configs': [item_config]
}
res = self.datasets(number, **attrs)
for row in res:
output.append(row['dataset'])
idict['dataset'] = output
return idict
def add_files(self, input_dataset, number_of_files=1):
"Add files to a given dataset"
record = deepcopy(input_dataset)
block = record['block']
if not isinstance(block, list):
block = [block]
for rec in block:
_, prim, proc, tier = rec['block_name'].split('#')[0].split('/')
attrs = {'prim':prim, 'proc':proc, 'tier':tier, 'block_name':rec['block_name'], 'output_configs':record['configs']}
res = self.files(number_of_files, **attrs)
return res
def add_blocks(self, input_dataset, number_of_blocks=1):
"Add blocks to a given dataset"
dataset = deepcopy(input_dataset)
name = dataset['dataset']['name']
res = self.blocks(number_of_blocks)
for row in res:
buid = generate_block_uid()
row['block']['name'] = '%s#%s' % (name, buid)
if dataset['dataset']['is-open'] == 'y':
blocks = dataset['dataset']['blocks']
blocks += res
return dataset
def add_blocks(self, input_dataset, number_of_blocks=1):
"Add blocks to a given dataset"
dataset = deepcopy(input_dataset)
name = dataset['dataset']['name']
res = self.blocks(number_of_blocks)
for row in res:
buid = generate_block_uid()
row['block']['name'] = '%s#%s' % (name, buid)
if dataset['dataset']['is-open'] == 'y':
blocks = dataset['dataset']['blocks']
blocks += res
return dataset
def add_blocks(self, input_dataset, number=1):
"Add blocks to a given dataset"
idict = deepcopy(input_dataset)
datasets = idict['dataset']
if isinstance(datasets, dict):
datasets = [datasets]
output = []
for item in datasets:
_, prim, proc, tier = item['dataset'].split('/')
attrs = {'prim': prim, 'proc': proc, 'tier': tier}
res = self.blocks(number, **attrs)
for row in res:
output.append(row['block'])
idict['block'] = output
return idict
def add_blocks(self, input_dataset, number=1):
"Add blocks to a given dataset"
idict = deepcopy(input_dataset)
datasets = idict['dataset']
if isinstance(datasets, dict):
datasets = [datasets]
output = []
for item in datasets:
_, prim, proc, tier = item['dataset'].split('/')
attrs = {'prim':prim, 'proc':proc, 'tier':tier}
res = self.blocks(number, **attrs)
for row in res:
output.append(row['block'])
idict['block'] = output
return idict
def dataset(self):
"return dataset object"
if not hasattr(self, '_dataset'):
self.generate_dataset()
phedex_data = {'dbs_name': self.dbs_name}
dataset = deepcopy(self._dataset)
dataset.update({'is-open': self.dataset_is_open})
for block in dataset['blocks']:
#update block information
size = sum([f['file']['bytes'] for f in block['block']['files']])
block['block'].update({"nfiles": len(block['block']['files']),
"size": size})
phedex_data.update(dict(dataset=dataset))
return phedex_data
def add_files(self, input_dataset, number_of_files=1):
"Add files to a given dataset"
record = deepcopy(input_dataset)
block = record['block']
if not isinstance(block, list):
block = [block]
for rec in block:
_, prim, proc, tier = rec['block_name'].split('#')[0].split('/')
attrs = {
'prim': prim,
'proc': proc,
'tier': tier,
'block_name': rec['block_name'],
'output_configs': record['configs']
}
res = self.files(number_of_files, **attrs)
return res
def gen_runs(self, file_record, number_of_runs=1):
"Generate run/lumis for a given file record"
if not isinstance(file_record, dict) or \
not file_record.has_key('file') or \
not file_record['file'].has_key('logical_file_name'):
msg = 'To generate run/lumis please provide valid file record/JSON file'
raise Exception(msg)
row = deepcopy(file_record)
records = []
for idx in range(0, number_of_runs):
run = random.randint(100000, 200000)
for _ in range(0, random.randint(1, 10)):
lumi = random.randint(1, 100)
rec = {'run_num': str(run), 'lumi_section_num': str(lumi)}
records.append(rec)
row['file']['file_lumi_list'] = records
return row
def gen_runs(self, file_record, number_of_runs=1):
"Generate run/lumis for a given file record"
if not isinstance(file_record, dict) or \
not file_record.has_key('file') or \
not file_record['file'].has_key('logical_file_name'):
msg = 'To generate run/lumis please provide valid file record/JSON file'
raise Exception(msg)
row = deepcopy(file_record)
records = []
for idx in range(0, number_of_runs):
run = random.randint(100000, 200000)
for _ in range(0, random.randint(1, 10)):
lumi = random.randint(1, 100)
rec = {'run_num': str(run), 'lumi_section_num': str(lumi)}
records.append(rec)
row['file']['file_lumi_list'] = records
return row
def dataset(self):
"return dataset object"
if not hasattr(self, '_dataset'):
self.generate_dataset()
phedex_data = {'dbs_name': self.dbs_name}
dataset = deepcopy(self._dataset)
dataset.update({'is-open': self.dataset_is_open})
for block in dataset['blocks']:
#update block information
size = sum([f['file']['bytes'] for f in block['block']['files']])
block['block'].update({
"nfiles": len(block['block']['files']),
"size": size
})
phedex_data.update(dict(dataset=dataset))
return phedex_data
def add_files(self, input_dataset, number_of_files=1):
"Add files to a given dataset"
dataset = deepcopy(input_dataset)
for block in dataset['dataset']['blocks']:
if block['block']['is-open'] != 'y':
continue
block_name = block['block']['name']
_, prim, proc, tier = block_name.split('#')[0].split('/')
attrs = {'prim':prim, 'proc':proc, 'tier':tier}
res = self.files(number_of_files, **attrs)
size = 0
for row in res:
size += row['file']['bytes']
if block['block']['files']:
block['block']['files'] += res
block['block']['size'] += size
block['block']['nfiles'] += len(res)
else:
block['block']['files'] = res
block['block']['size'] = size
block['block']['nfiles'] = len(res)
return dataset
def add_files(self, input_dataset, number_of_files=1):
"Add files to a given dataset"
dataset = deepcopy(input_dataset)
for block in dataset['dataset']['blocks']:
if block['block']['is-open'] != 'y':
continue
block_name = block['block']['name']
_, prim, proc, tier = block_name.split('#')[0].split('/')
attrs = {'prim': prim, 'proc': proc, 'tier': tier}
if "tags" in dataset['dataset']:
attrs['tags'] = dataset['dataset']['tags']
res = self.files(number_of_files, **attrs)
size = 0
for row in res:
size += row['file']['bytes']
if block['block']['files']:
block['block']['files'] += res
block['block']['size'] += size
block['block']['nfiles'] += len(res)
else:
block['block']['files'] = res
block['block']['size'] = size
block['block']['nfiles'] = len(res)
return dataset
def workflow(fin, fout, verbose=None):
"LifeCycle workflow"
initial_payload = None # initial payload, should be provided by LifeCycle
new_payload = [] # newly created payloads will be returned by LifeCycle
with open(fin, 'r') as source:
initial_payload = json.load(source)
if verbose:
print "\n### input workflow"
print pprint.pformat(initial_payload)
### read inputs from payload
workflow = initial_payload['workflow']
# check if input are read from configuration file
try:
cfg = workflow['DataProviderCfg']
except KeyError:
#No configuration, try to use values provided in the workflow
#for backward compatibility
#values using get are optional
cdict = { 'process' :
{'NumberOfDatasets' : workflow['NumberOfDatasets'],
'NumberOfBlocks' : workflow['NumberOfBlocks'],
'NumberOfFiles' : workflow['NumberOfFiles'],
'NumberOfRuns' : workflow['NumberOfRuns'],
'NumberOfLumis' : workflow['NumberOfLumis']},
'dbs' :
{'DBSSkipFileFail': workflow.get('DBSSkipFileFail', None),
'DBSChangeCksumFail': workflow.get('DBSChangeCksumFail', None),
'DBSChangeSizeFail': workflow.get('DBSChangeSizeFail', None)},
'phedex' :
{'PhedexSkipFileFail' : workflow.get('PhedexSkipFileFail', None),
'PhedexChangeCksumFail' : workflow.get('PhedexChangeCksumFail', None),
'PhedexChangeSizeFail' : workflow.get('PhedexChangeSizeFail', None),
'PhedexDBSName' : workflow['PhedexDBSName']}
}
else:
cdict = read_configparser(cfg)
process_cfg = cdict['process']
dbs_cfg = cdict['dbs']
phedex_cfg = cdict['phedex']
phedex_dbs_name = phedex_cfg.get('PhedexDBSName')
number_of_datasets = int(process_cfg.get('NumberOfDatasets'))
number_of_blocks = int(process_cfg.get('NumberOfBlocks'))
number_of_files = int(process_cfg.get('NumberOfFiles'))
number_of_runs = int(process_cfg.get('NumberOfRuns'))
number_of_lumis = int(process_cfg.get('NumberOfLumis'))
try:
phedex_file = float(phedex_cfg.get('PhedexSkipFileFail'))
phedex_cksum = float(phedex_cfg.get('PhedexChangeCksumFail'))
phedex_size = float(phedex_cfg.get('PhedexChangeSizeFail'))
dbs_file = float(dbs_cfg.get('DBSSkipFileFail'))
dbs_cksum = float(dbs_cfg.get('DBSChangeCksumFail'))
dbs_size = float(dbs_cfg.get('DBSChangeSizeFail'))
# if value is None, the cast will fail, which means no failures are used
except TypeError:
failure_rates = None
else:
failure_rates = dict(PhedexSkipFileFail=phedex_file)
failure_rates.update(PhedexChangeCksumFail=phedex_cksum)
failure_rates.update(PhedexChangeSizeFail=phedex_size)
failure_rates.update(DBSSkipFileFail=dbs_file)
failure_rates.update(DBSChangeCksumFail=dbs_cksum)
failure_rates.update(DBSChangeSizeFail=dbs_size)
print failure_rates
phedex_provider = PhedexProvider(dbs_name=phedex_dbs_name, failure_rates=failure_rates)
dbs_provider = DBSProvider(failure_rates=failure_rates)
for _ in xrange(number_of_datasets):
#clone initial payload
payload = deepcopy(initial_payload)
phedex_provider.generate_dataset()
phedex_provider.add_blocks(number_of_blocks)
phedex_provider.add_files(number_of_files)
payload['workflow']['Phedex'] = [phedex_provider.dataset()]
payload['workflow']['DBS'] = dbs_provider.block_dump(number_of_runs,
number_of_lumis)
phedex_provider.reset()
new_payload.append(payload)
with open(fout, 'w') as output:
json.dump(new_payload, output)
if verbose:
print "\n### output workflow"
print pprint.pformat(new_payload)
def workflow(fin, fout, verbose=None):
"LifeCycle workflow"
initial_payload = None # initial payload, should be provided by LifeCycle
new_payload = [] # newly created payloads will be returned by LifeCycle
with open(fin, 'r') as source:
initial_payload = json.load(source)
if verbose:
print "\n### input workflow"
print pprint.pformat(initial_payload)
### read inputs from payload
workflow = initial_payload['workflow']
# check if input are read from configuration file
try:
cfg = workflow['DataProviderCfg']
except KeyError:
#No configuration, try to use values provided in the workflow
#for backward compatibility
#values using get are optional
cdict = {
'process': {
'NumberOfDatasets': workflow['NumberOfDatasets'],
'NumberOfBlocks': workflow['NumberOfBlocks'],
'NumberOfFiles': workflow['NumberOfFiles'],
'NumberOfRuns': workflow['NumberOfRuns'],
'NumberOfLumis': workflow['NumberOfLumis']
},
'dbs': {
'DBSSkipFileFail': workflow.get('DBSSkipFileFail', None),
'DBSChangeCksumFail': workflow.get('DBSChangeCksumFail', None),
'DBSChangeSizeFail': workflow.get('DBSChangeSizeFail', None)
},
'phedex': {
'PhedexSkipFileFail':
workflow.get('PhedexSkipFileFail', None),
'PhedexChangeCksumFail':
workflow.get('PhedexChangeCksumFail', None),
'PhedexChangeSizeFail':
workflow.get('PhedexChangeSizeFail', None),
'PhedexDBSName':
workflow['PhedexDBSName']
}
}
else:
cdict = read_configparser(cfg)
process_cfg = cdict['process']
dbs_cfg = cdict['dbs']
phedex_cfg = cdict['phedex']
phedex_dbs_name = phedex_cfg.get('PhedexDBSName')
number_of_datasets = int(process_cfg.get('NumberOfDatasets'))
number_of_blocks = int(process_cfg.get('NumberOfBlocks'))
number_of_files = int(process_cfg.get('NumberOfFiles'))
number_of_runs = int(process_cfg.get('NumberOfRuns'))
number_of_lumis = int(process_cfg.get('NumberOfLumis'))
try:
phedex_file = float(phedex_cfg.get('PhedexSkipFileFail'))
phedex_cksum = float(phedex_cfg.get('PhedexChangeCksumFail'))
phedex_size = float(phedex_cfg.get('PhedexChangeSizeFail'))
dbs_file = float(dbs_cfg.get('DBSSkipFileFail'))
dbs_cksum = float(dbs_cfg.get('DBSChangeCksumFail'))
dbs_size = float(dbs_cfg.get('DBSChangeSizeFail'))
# if value is None, the cast will fail, which means no failures are used
except TypeError:
failure_rates = None
else:
failure_rates = dict(PhedexSkipFileFail=phedex_file)
failure_rates.update(PhedexChangeCksumFail=phedex_cksum)
failure_rates.update(PhedexChangeSizeFail=phedex_size)
failure_rates.update(DBSSkipFileFail=dbs_file)
failure_rates.update(DBSChangeCksumFail=dbs_cksum)
failure_rates.update(DBSChangeSizeFail=dbs_size)
print failure_rates
phedex_provider = PhedexProvider(dbs_name=phedex_dbs_name,
failure_rates=failure_rates)
dbs_provider = DBSProvider(failure_rates=failure_rates)
for _ in xrange(number_of_datasets):
#clone initial payload
payload = deepcopy(initial_payload)
phedex_provider.generate_dataset()
phedex_provider.add_blocks(number_of_blocks)
phedex_provider.add_files(number_of_files)
payload['workflow']['Phedex'] = [phedex_provider.dataset()]
payload['workflow']['DBS'] = dbs_provider.block_dump(
number_of_runs, number_of_lumis)
phedex_provider.reset()
new_payload.append(payload)
with open(fout, 'w') as output:
json.dump(new_payload, output)
if verbose:
print "\n### output workflow"
print pprint.pformat(new_payload)
