| 42 |
|
def __init__(self, owner, dataset, dataTiers, cfg_params): |
| 43 |
|
|
| 44 |
|
# Attributes |
| 45 |
< |
self.dbsdataset='/'+dataset+'/datatier/'+owner |
| 45 |
> |
self.owner = owner |
| 46 |
> |
self.dataset = dataset |
| 47 |
|
self.dataTiers = dataTiers |
| 48 |
|
self.cfg_params = cfg_params |
| 49 |
|
|
| 57 |
|
""" |
| 58 |
|
Contact DBS |
| 59 |
|
""" |
| 59 |
– |
parents = [] |
| 60 |
– |
parentsblocksinfo = {} |
| 60 |
|
|
| 61 |
|
## add the PU among the required data tiers if the Digi are requested |
| 62 |
|
if (self.dataTiers.count('Digi')>0) & (self.dataTiers.count('PU')<=0) : |
| 63 |
|
self.dataTiers.append('PU') |
| 64 |
|
|
| 65 |
|
## get info about the requested dataset |
| 66 |
< |
dbs=DBSInfo(self.dbsdataset,self.dataTiers) |
| 66 |
> |
dbs=DBSInfo() |
| 67 |
> |
self.datasets = dbs.getMatchingDatasets(self.owner, self.dataset) |
| 68 |
> |
if len(self.datasets) == 0: |
| 69 |
> |
raise DataDiscoveryError("Owner=%s, Dataset=%s unknown to DBS" % (self.owner, self.dataset)) |
| 70 |
> |
if len(self.datasets) > 1: |
| 71 |
> |
raise DataDiscoveryError("Owner=%s, Dataset=%s is ambiguous" % (self.owner, self.dataset)) |
| 72 |
|
try: |
| 73 |
< |
self.blocksinfo=dbs.getDatasetContents() |
| 73 |
> |
self.dbsdataset = self.datasets[0].getDatasetPath() |
| 74 |
> |
self.blocksinfo = dbs.getDatasetContents(self.dbsdataset) |
| 75 |
> |
self.allblocks.append (self.blocksinfo.keys ()) # add also the current fileblocksinfo |
| 76 |
> |
self.dbspaths.append(self.dbsdataset) |
| 77 |
|
except DBSError, ex: |
| 78 |
|
raise DataDiscoveryError(ex.getErrorMessage()) |
| 79 |
|
|
| 80 |
|
if len(self.blocksinfo)<=0: |
| 81 |
< |
msg="\nERROR Data %s do not exist in DBS! \n Check the dataset/owner variables in crab.cfg !"%self.dbsdataset |
| 81 |
> |
msg="\nERROR Data for %s do not exist in DBS! \n Check the dataset/owner variables in crab.cfg !"%self.dbsdataset |
| 82 |
|
raise NotExistingDatasetError(msg) |
| 83 |
|
|
| 77 |
– |
currentdatatier=string.split(self.blocksinfo.keys()[0],'/')[2] |
| 78 |
– |
fakedatatier=string.split(self.dbsdataset,'/')[2] |
| 79 |
– |
currentdbsdataset=string.replace(self.dbsdataset, fakedatatier, currentdatatier) |
| 80 |
– |
|
| 81 |
– |
self.dbspaths.append(currentdbsdataset) # add the requested dbspath |
| 84 |
|
|
| 85 |
|
## get info about the parents |
| 86 |
|
try: |
| 87 |
< |
parents=dbs.getDatasetProvenance() |
| 87 |
> |
parents=dbs.getDatasetProvenance(self.dbsdataset, self.dataTiers) |
| 88 |
|
except DBSInvalidDataTierError, ex: |
| 89 |
|
msg=ex.getErrorMessage()+' \n Check the data_tier variable in crab.cfg !\n' |
| 90 |
|
raise DataDiscoveryError(msg) |
| 92 |
|
raise DataDiscoveryError(ex.getErrorMessage()) |
| 93 |
|
|
| 94 |
|
## check that the user asks for parent Data Tier really existing in the DBS provenance |
| 95 |
< |
self.checkParentDataTier(parents, self.dataTiers, currentdbsdataset) |
| 95 |
> |
self.checkParentDataTier(parents, self.dataTiers) |
| 96 |
|
|
| 97 |
|
## for each parent get the corresponding fileblocks |
| 98 |
< |
for aparent in parents: |
| 99 |
< |
## fill a list of dbspaths |
| 100 |
< |
parentdbsdataset=aparent.getDatasetPath() |
| 101 |
< |
self.dbspaths.append(parentdbsdataset) |
| 102 |
< |
pdbs=DBSInfo(parentdbsdataset,[]) |
| 103 |
< |
try: |
| 104 |
< |
parentsblocksinfo=pdbs.getDatasetContents() |
| 105 |
< |
except DBSError, ex: |
| 98 |
> |
try: |
| 99 |
> |
for p in parents: |
| 100 |
> |
## fill a list of dbspaths |
| 101 |
> |
parentPath = p.getDatasetPath() |
| 102 |
> |
self.dbspaths.append (parentPath) |
| 103 |
> |
parentBlocks = dbs.getDatasetContents (parentPath) |
| 104 |
> |
self.allblocks.append (parentBlocks.keys ()) # add parent fileblocksinfo |
| 105 |
> |
except DBSError, ex: |
| 106 |
|
raise DataDiscoveryError(ex.getErrorMessage()) |
| 107 |
|
|
| 106 |
– |
self.allblocks.append(parentsblocksinfo.keys()) # add parent fileblocksinfo |
| 107 |
– |
|
| 108 |
– |
## all the required blocks |
| 109 |
– |
self.allblocks.append(self.blocksinfo.keys()) # add also the current fileblocksinfo |
| 110 |
– |
|
| 111 |
– |
|
| 108 |
|
# ################################################# |
| 109 |
< |
def checkParentDataTier(self, parents, user_datatiers, currentdbsdataset ): |
| 109 |
> |
def checkParentDataTier(self, parents, dataTiers): |
| 110 |
|
""" |
| 111 |
|
check that the data tiers requested by the user really exists in the provenance of the given dataset |
| 112 |
|
""" |
| 113 |
< |
|
| 114 |
< |
current_datatier=string.split(currentdbsdataset,'/')[2] |
| 115 |
< |
|
| 116 |
< |
parent_datatypes=[] |
| 117 |
< |
for aparent in parents: |
| 118 |
< |
parent_datatypes.append(aparent.getDataType()) |
| 123 |
< |
|
| 124 |
< |
for datatier in user_datatiers: |
| 125 |
< |
if parent_datatypes.count(datatier)<=0: |
| 126 |
< |
# the current datatier is not supposed to be in the provenance |
| 127 |
< |
if not (datatier == current_datatier): |
| 128 |
< |
msg="\nERROR Data %s not published in DBS with asked data tiers : the data tier not found is %s !\n Check the data_tier variable in crab.cfg !"%(currentdbsdataset,datatier) |
| 129 |
< |
raise NoDataTierinProvenanceError(msg) |
| 113 |
> |
startType = string.split(self.dbsdataset,'/')[2] |
| 114 |
> |
parentTypes = map(lambda p: p.getDataType(), parents) |
| 115 |
> |
for tier in dataTiers: |
| 116 |
> |
if parentTypes.count(tier) <= 0 and tier != startType: |
| 117 |
> |
msg="\nERROR Data %s not published in DBS with asked data tiers : the data tier not found is %s !\n Check the data_tier variable in crab.cfg !"%(self.dbsdataset,tier) |
| 118 |
> |
raise NoDataTierinProvenanceError(msg) |
| 119 |
|
|
| 120 |
|
|
| 121 |
|
# ################################################# |
