def _check_lumi_filter(self, block, idx_runs, idx_lumi):
lumi_filter = self._lumi_filter.lookup(block[DataProvider.Nickname], is_selector=False)
if not lumi_filter:
return
if (self._lumi_strict == LumiMode.strict) and ((idx_runs is None) or (idx_lumi is None)):
raise DatasetError('Strict lumi filter active but ' +
'dataset %s does not provide lumi information!' % DataProvider.get_block_id(block))
elif (self._lumi_strict == LumiMode.weak) and (idx_runs is None):
raise DatasetError('Weak lumi filter active but ' +
'dataset %s does not provide run information!' % DataProvider.get_block_id(block))
def _check_lumi_filter(self, block, idx_runs, idx_lumi):
lumi_filter = self._lumi_filter.lookup(block[DataProvider.Nickname], is_selector=False)
if not lumi_filter:
return
if (self._lumi_strict == LumiMode.strict) and ((idx_runs is None) or (idx_lumi is None)):
raise DatasetError('Strict lumi filter active but ' +
'dataset %s does not provide lumi information!' % DataProvider.get_block_id(block))
elif (self._lumi_strict == LumiMode.weak) and (idx_runs is None):
raise DatasetError('Weak lumi filter active but ' +
'dataset %s does not provide run information!' % DataProvider.get_block_id(block))
def create_dbs3_json_blocks(opts, dataset_blocks):
dbs3_proto_block_iter = create_dbs3_proto_blocks(opts, dataset_blocks)
for (block, block_dump, block_size, dataset_type) in dbs3_proto_block_iter:
dataset = block[DataProvider.Dataset]
try:
primary_dataset, processed_dataset, data_tier = dataset[1:].split(
'/')
except Exception:
raise DatasetError('Dataset name %s is not a valid DBS name!' %
dataset)
# add primary dataset information
block_dump['primds'] = {
'primary_ds_type': dataset_type,
'primary_ds_name': primary_dataset
}
# add dataset information
block_dump['dataset'] = {
'dataset': dataset,
'processed_ds_name': processed_dataset,
'data_tier_name': data_tier,
'physics_group_name': None,
'dataset_access_type': 'VALID',
'xtcrosssection':
None, # TODO: Add to metadata from FrameWorkJobReport, if possible!
}
# add block information
site_db = CRIC()
try:
origin_site_name = site_db.se_to_cms_name(
block[DataProvider.Locations][0])[0]
except IndexError:
clear_current_exception()
origin_site_name = 'UNKNOWN'
block_dump['block'] = {
'block_name': DataProvider.get_block_id(block),
'block_size': block_size,
'file_count': len(block[DataProvider.FileList]),
'origin_site_name': origin_site_name
}
if opts.do_close_blocks:
block_dump['block']['open_for_writing'] = 0
else:
block_dump['block']['open_for_writing'] = 1
# add acquisition_era, CRAB is important because of checks within DBS 3
block_dump['acquisition_era'] = {
'acquisition_era_name': 'CRAB',
'start_date': 0
}
# add processing_era
block_dump['processing_era'] = {
'processing_version': 1,
'description': 'grid-control'
}
yield validate_dbs3_json('blockBulk', block_dump)
def create_dbs3_proto_blocks(opts, dataset_blocks):
for dataset in dataset_blocks:
missing_info_blocks = []
dataset_types = set()
for block in dataset_blocks[dataset]:
block_dump = {'dataset_conf_list': [], 'files': [], 'file_conf_list': [], 'file_parent_list': []}
(block_size, block_dataset_types) = create_dbs3_json_files(opts, block, block_dump)
if len(block_dataset_types) > 1:
raise Exception('Data and MC files are mixed in block %s' % DataProvider.get_block_id(block))
elif len(block_dataset_types) == 1:
yield (block, block_dump, block_size, block_dataset_types.pop())
else:
missing_info_blocks.append((block, block_dump, block_size))
# collect dataset types in this dataset for blocks with missing type information
dataset_types.update(block_dataset_types)
if missing_info_blocks:
if len(dataset_types) > 1:
raise Exception('Data and MC files are mixed in dataset %s! ' +
'Unable to determine dataset type for blocks without type info')
elif len(dataset_types) == 0:
if not opts.datatype:
raise Exception('Please supply dataset type via --datatype!')
dataset_type = opts.datatype
else:
dataset_type = dataset_types.pop()
for (block, block_dump, block_size) in missing_info_blocks:
yield (block, block_dump, block_size, dataset_type)
def create_dbs3_json_blocks(opts, dataset_blocks):
dbs3_proto_block_iter = create_dbs3_proto_blocks(opts, dataset_blocks)
for (block, block_dump, block_size, dataset_type) in dbs3_proto_block_iter:
dataset = block[DataProvider.Dataset]
try:
primary_dataset, processed_dataset, data_tier = dataset[1:].split('/')
except Exception:
raise DatasetError('Dataset name %s is not a valid DBS name!' % dataset)
# add primary dataset information
block_dump['primds'] = {'primary_ds_type': dataset_type, 'primary_ds_name': primary_dataset}
# add dataset information
block_dump['dataset'] = {
'dataset': dataset, 'processed_ds_name': processed_dataset, 'data_tier_name': data_tier,
'physics_group_name': None, 'dataset_access_type': 'VALID',
'xtcrosssection': None, # TODO: Add to metadata from FrameWorkJobReport, if possible!
}
# add block information
site_db = SiteDB()
try:
origin_site_name = site_db.se_to_cms_name(block[DataProvider.Locations][0])[0]
except IndexError:
clear_current_exception()
origin_site_name = 'UNKNOWN'
block_dump['block'] = {'block_name': DataProvider.get_block_id(block), 'block_size': block_size,
'file_count': len(block[DataProvider.FileList]), 'origin_site_name': origin_site_name}
if opts.do_close_blocks:
block_dump['block']['open_for_writing'] = 0
else:
block_dump['block']['open_for_writing'] = 1
# add acquisition_era, CRAB is important because of checks within DBS 3
block_dump['acquisition_era'] = {'acquisition_era_name': 'CRAB', 'start_date': 0}
# add processing_era
block_dump['processing_era'] = {'processing_version': 1, 'description': 'grid-control'}
yield validate_dbs3_json('blockBulk', block_dump)
def _get_fi_class(self, fi, block):
run_range = self._run_range.lookup(DataProvider.get_block_id(block))
metadata_idx = block[DataProvider.Metadata].index('Runs')
return tuple(
imap(lambda r: int(r / run_range),
fi[DataProvider.Metadata][metadata_idx]))
def _get_fi_class(self, fi, block):
run_range = self._run_range.lookup(DataProvider.get_block_id(block))
metadata_idx = block[DataProvider.Metadata].index('Runs')
return tuple(imap(lambda r: int(r / run_range), fi[DataProvider.Metadata][metadata_idx]))
