Python JobSpec.transformation方法代码示例

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
    def defineEvgen16Job(self, i):
        """Define an Evgen16 job based on predefined values and randomly generated names
        """

        job = JobSpec()
        job.computingSite = self.__site
        job.cloud = self.__cloud

        job.jobDefinitionID = int(time.time()) % 10000
        job.jobName = "%s_%d" % (uuid.uuid1(), i)
        job.AtlasRelease = 'Atlas-16.6.2'
        job.homepackage = 'AtlasProduction/16.6.2.1'
        job.transformation = 'Evgen_trf.py'
        job.destinationDBlock = self.__datasetName
        job.destinationSE = self.__destName
        job.currentPriority = 10000
        job.prodSourceLabel = 'test'
        job.cmtConfig = 'i686-slc5-gcc43-opt'

        #Output file
        fileO = FileSpec()
        fileO.lfn = "%s.evgen.pool.root" % job.jobName
        fileO.destinationDBlock = job.destinationDBlock
        fileO.destinationSE = job.destinationSE
        fileO.dataset = job.destinationDBlock
        fileO.destinationDBlockToken = 'ATLASDATADISK'
        fileO.type = 'output'
        job.addFile(fileO)

        #Log file
        fileL = FileSpec()
        fileL.lfn = "%s.job.log.tgz" % job.jobName
        fileL.destinationDBlock = job.destinationDBlock
        fileL.destinationSE = job.destinationSE
        fileL.dataset = job.destinationDBlock
        fileL.destinationDBlockToken = 'ATLASDATADISK'
        fileL.type = 'log'
        job.addFile(fileL)

        job.jobParameters = "2760 105048 19901 101 200 MC10.105048.PythiaB_ccmu3mu1X.py %s NONE NONE NONE MC10JobOpts-latest-test.tar.gz" % fileO.lfn
        return job

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
#site = sys.argv[1]
site = 'ANALY_BNL-LSST'  #orig
#site = 'BNL-LSST'
#site = 'SWT2_CPB-LSST'
#site = 'UTA_SWT2-LSST'
#site = 'ANALY_SWT2_CPB-LSST'

datasetName = 'panda.user.jschovan.lsst.%s' % commands.getoutput('uuidgen')
destName    = None

job = JobSpec()
job.jobDefinitionID   = int(time.time()) % 10000
job.jobName           = "%s" % commands.getoutput('uuidgen')
### job.transformation    = 'http://www.usatlas.bnl.gov/~wenaus/lsst-trf/lsst-trf.sh'
#job.transformation    = 'http://pandawms.org/pandawms-jobcache/lsst-trf.sh'
job.transformation = 'http://pandawms.org/pandawms-jobcache/lsst-trf-phosim332.sh'
job.destinationDBlock = datasetName
#job.destinationSE     = destName
job.destinationSE     = 'local' 
job.currentPriority   = 1000
#job.prodSourceLabel   = 'ptest'
#job.prodSourceLabel = 'panda'
#job.prodSourceLabel = 'ptest'
#job.prodSourceLabel = 'test'
#job.prodSourceLabel = 'ptest'
### 2014-01-27
#job.prodSourceLabel = 'user'
job.prodSourceLabel = 'panda'
job.computingSite     = site
job.jobParameters = ""
job.VO = "lsst"

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
    def createJobSpec(self, task, outdataset, job, jobset, jobdef, site, jobname, lfnhanger, allsites, jobid):
        """Create a spec for one job

        :arg TaskWorker.DataObject.Task task: the task to work on
        :arg str outdataset: the output dataset name where all the produced files will be placed
        :arg WMCore.DataStructs.Job job: the abstract job
        :arg int jobset: the PanDA jobset corresponding to the current task
        :arg int jobdef: the PanDA jobdef where to append the current jobs --- not used
        :arg str site: the borkered site where to run the jobs
        :arg str jobname: the job name
        :arg str lfnhanger: the random string to be added in the output file name
        :arg list str allsites: all possible sites where the job can potentially run
        :arg int jobid: incremental job number
        :return: the sepc object."""

        pandajob = JobSpec()
        ## always setting a job definition ID
        pandajob.jobDefinitionID = jobdef if jobdef else -1
        ## always setting a job set ID
        pandajob.jobsetID = jobset if jobset else -1
        pandajob.jobName = jobname
        pandajob.prodUserID = task['tm_user_dn']
        pandajob.destinationDBlock = outdataset
        pandajob.prodDBlock = task['tm_input_dataset']
        pandajob.prodSourceLabel = 'user'
        pandajob.computingSite = site
        pandajob.cloud = getSite(pandajob.computingSite)
        pandajob.destinationSE = 'local'
        pandajob.transformation = task['tm_transformation']
        ## need to initialize this
        pandajob.metadata = ''

        def outFileSpec(of=None, log=False):
            """Local routine to create an FileSpec for the an job output/log file

               :arg str of: output file base name
               :return: FileSpec object for the output file."""
            outfile = FileSpec()
            if log:
                outfile.lfn = "job.log_%d_%s.tgz" % (jobid, lfnhanger)
                outfile.type = 'log'
            else:
                outfile.lfn = '%s_%d_%s%s' %(os.path.splitext(of)[0], jobid, lfnhanger, os.path.splitext(of)[1])
                outfile.type = 'output'
            outfile.destinationDBlock = pandajob.destinationDBlock
            outfile.destinationSE = task['tm_asyncdest']
            outfile.dataset = pandajob.destinationDBlock
            return outfile

        alloutfiles = []
        outjobpar = {}
        outfilestring = ''
        for outputfile in task['tm_outfiles']:
            outfilestring += '%s,' % outputfile
            filespec = outFileSpec(outputfile)
            alloutfiles.append(filespec)
            #pandajob.addFile(filespec)
            outjobpar[outputfile] = filespec.lfn
        for outputfile in task['tm_tfile_outfiles']:
            outfilestring += '%s,' % outputfile
            filespec = outFileSpec(outputfile)
            alloutfiles.append(filespec)
            #pandajob.addFile(filespec)
            outjobpar[outputfile] = filespec.lfn
        for outputfile in task['tm_edm_outfiles']:
            outfilestring += '%s,' % outputfile
            filespec = outFileSpec(outputfile)
            alloutfiles.append(filespec)
            #pandajob.addFile(filespec)
            outjobpar[outputfile] = filespec.lfn
        outfilestring = outfilestring[:-1]

        infiles = []
        for inputfile in job['input_files']:
            infiles.append( inputfile['lfn'] )

        pandajob.jobParameters = '-a %s ' % task['tm_user_sandbox']
        pandajob.jobParameters += '--sourceURL %s ' % task['tm_cache_url']
        pandajob.jobParameters += '--jobNumber=%s ' % jobid
        pandajob.jobParameters += '--cmsswVersion=%s ' % task['tm_job_sw']
        pandajob.jobParameters += '--scramArch=%s ' % task['tm_job_arch']
        pandajob.jobParameters += '--inputFile=\'%s\' ' % json.dumps(infiles)

        self.jobParametersSetting(pandajob, job, self.jobtypeMapper[task['tm_job_type']])

        pandajob.jobParameters += '-o "%s" ' % str(outjobpar)
        pandajob.jobParameters += '--dbs_url=%s ' % task['tm_dbs_url']
        pandajob.jobParameters += '--publish_dbs_url=%s ' % task['tm_publish_dbs_url']
        pandajob.jobParameters += '--publishFiles=%s ' % ('True' if task['tm_publication'] == 'T' else 'False')
        pandajob.jobParameters += '--saveLogs=%s ' % ('True' if task['tm_save_logs'] == 'T' else 'False')
        pandajob.jobParameters += '--availableSites=\'%s\' ' %json.dumps(allsites)

        pandajob.jobParameters += '--group=%s ' % (task['tm_user_group'] if task['tm_user_group'] else '')
        pandajob.jobParameters += '--role=%s ' % (task['tm_user_role'] if task['tm_user_role'] else '')

        self.logger.info(type(task['tm_user_infiles']))
        self.logger.info(task['tm_user_infiles'])

        if task['tm_user_infiles']:
            addinfilestring = ''
#.........这里部分代码省略.........

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]


jobList = []

for i in range(1):
  for lfn in files.keys():
    job = JobSpec()
    job.jobDefinitionID   = int(time.time()) % 10000
    job.jobName           = commands.getoutput('uuidgen') 
    job.AtlasRelease      = 'Atlas-12.0.7'
    job.homepackage       = 'AtlasProduction/12.0.7.1'
# Need different args too
#    job.AtlasRelease      = 'Atlas-13.0.30'
#    job.homepackage       = 'AtlasProduction/13.0.30.2'
    job.transformation    = 'csc_simul_trf.py'
    job.destinationDBlock = datasetName
    job.cloud = cloud
    job.computingSite     = site
#    job.prodDBlock        = 'mc12.005001.pythia_minbias.evgen.EVNT.v12000701_tid012303'
    job.prodDBlock        = files[lfn]
    job.prodSourceLabel   = 'test'
#    job.prodSourceLabel   = 'cloudtest'
    job.currentPriority   = 1001

    fileI = FileSpec()
    fileI.dataset    = job.prodDBlock
    fileI.prodDBlock = job.prodDBlock
    fileI.lfn = lfn
    fileI.type = 'input'
    job.addFile(fileI)

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
for idx,argv in enumerate(sys.argv):
    if argv == '-s':
        aSrvID = sys.argv[idx+1]
        sys.argv = sys.argv[:idx]
        break

site = sys.argv[1]

datasetName = 'panda.destDB.%s' % commands.getoutput('uuidgen')
destName    = 'local'

job = JobSpec()
job.jobDefinitionID   = int(time.time()) % 10000
job.jobName           = "%s" % commands.getoutput('uuidgen')
# MPI transform on Titan that will run actual job                               
job.transformation    = '/lustre/atlas/proj-shared/csc108/panitkin/alicetest1/m\
pi_wrapper_alice_ppbench.py'

job.destinationDBlock = datasetName
job.destinationSE     = destName
job.currentPriority   = 1000
job.prodSourceLabel   = 'panda'
job.computingSite     = site
job.jobParameters     = " "
job.VO                = 'alice'

fileOL = FileSpec()
fileOL.lfn = "%s.job.log.tgz" % job.jobName
fileOL.destinationDBlock = job.destinationDBlock
fileOL.destinationSE     = job.destinationSE
fileOL.dataset           = job.destinationDBlock

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]

#.........这里部分代码省略.........
        elif job.splitter._name not in ['DQ2JobSplitter', 'ArgSplitter', 'ArgSplitterTask']:
            raise ApplicationConfigurationError('Panda splitter must be DQ2JobSplitter or ArgSplitter')
        
        if job.backend.site == 'AUTO':
            raise ApplicationConfigurationError('site is still AUTO after brokerage!')

#       output dataset
        if job.outputdata:
            if job.outputdata._name != 'DQ2OutputDataset':
                raise ApplicationConfigurationError('Panda backend supports only DQ2OutputDataset')
        else:
            logger.info('Adding missing DQ2OutputDataset')
            job.outputdata = DQ2OutputDataset()

        job.outputdata.datasetname,outlfn = dq2outputdatasetname(job.outputdata.datasetname, job.id, job.outputdata.isGroupDS, job.outputdata.groupname)

        self.outDsLocation = Client.PandaSites[job.backend.site]['ddm']

        try:
            Client.addDataset(job.outputdata.datasetname,False,location=self.outDsLocation)
            logger.info('Output dataset %s registered at %s'%(job.outputdata.datasetname,self.outDsLocation))
            dq2_set_dataset_lifetime(job.outputdata.datasetname, location=self.outDsLocation)
        except exceptions.SystemExit:
            raise BackendError('Panda','Exception in Client.addDataset %s: %s %s'%(job.outputdata.datasetname,sys.exc_info()[0],sys.exc_info()[1]))

        # handle the libds
        if job.backend.libds:
            self.libDataset = job.backend.libds
            self.fileBO = getLibFileSpecFromLibDS(self.libDataset)
            self.library = self.fileBO.lfn
        elif job.backend.bexec:
            self.libDataset = job.outputdata.datasetname+'.lib'
            self.library = '%s.tgz' % self.libDataset
            try:
                Client.addDataset(self.libDataset,False,location=self.outDsLocation)
                dq2_set_dataset_lifetime(self.libDataset, location=self.outDsLocation)
                logger.info('Lib dataset %s registered at %s'%(self.libDataset,self.outDsLocation))
            except exceptions.SystemExit:
                raise BackendError('Panda','Exception in Client.addDataset %s: %s %s'%(self.libDataset,sys.exc_info()[0],sys.exc_info()[1]))

        # collect extOutFiles
        self.extOutFile = []
        for tmpName in job.outputdata.outputdata:
            if tmpName != '':
                self.extOutFile.append(tmpName)

        for tmpName in job.outputsandbox:
            if tmpName != '':
                self.extOutFile.append(tmpName)

        for tmpName in job.backend.extOutFile:
            if tmpName != '':
                self.extOutFile.append(tmpName)

        # create build job
        if job.backend.bexec != '':
            jspec = JobSpec()
            jspec.jobDefinitionID   = job.id
            jspec.jobName           = commands.getoutput('uuidgen 2> /dev/null')
            jspec.transformation    = '%s/buildGen-00-00-01' % Client.baseURLSUB
            if Client.isDQ2free(job.backend.site):
                jspec.destinationDBlock = '%s/%s' % (job.outputdata.datasetname,self.libDataset)
                jspec.destinationSE     = 'local'
            else:
                jspec.destinationDBlock = self.libDataset
                jspec.destinationSE     = job.backend.site
            jspec.prodSourceLabel   = configPanda['prodSourceLabelBuild']
            jspec.processingType    = configPanda['processingType']
            jspec.assignedPriority  = configPanda['assignedPriorityBuild']
            jspec.computingSite     = job.backend.site
            jspec.cloud             = job.backend.requirements.cloud
            jspec.jobParameters     = '-o %s' % (self.library)
            if self.inputsandbox:
                jspec.jobParameters     += ' -i %s' % (self.inputsandbox)
            else:
                raise ApplicationConfigurationError('Executable on Panda with build job defined, but inputsandbox is emtpy !')
            matchURL = re.search('(http.*://[^/]+)/',Client.baseURLCSRVSSL)
            if matchURL:
                jspec.jobParameters += ' --sourceURL %s ' % matchURL.group(1)
            if job.backend.bexec != '':
                jspec.jobParameters += ' --bexec "%s" ' % urllib.quote(job.backend.bexec)
                jspec.jobParameters += ' -r %s ' % '.'
                

            fout = FileSpec()
            fout.lfn  = self.library
            fout.type = 'output'
            fout.dataset = self.libDataset
            fout.destinationDBlock = self.libDataset
            jspec.addFile(fout)

            flog = FileSpec()
            flog.lfn = '%s.log.tgz' % self.libDataset
            flog.type = 'log'
            flog.dataset = self.libDataset
            flog.destinationDBlock = self.libDataset
            jspec.addFile(flog)
            return jspec
        else:
            return None

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
    def prepare(self, app, appsubconfig, appmasterconfig, jobmasterconfig):
        """Prepare the specific aspec of each subjob.
           Returns: subjobconfig list of objects understood by backends."""

        from pandatools import Client
        from pandatools import AthenaUtils
        from taskbuffer.JobSpec import JobSpec
        from taskbuffer.FileSpec import FileSpec
        from GangaAtlas.Lib.ATLASDataset.DQ2Dataset import dq2_set_dataset_lifetime
        from GangaPanda.Lib.Panda.Panda import refreshPandaSpecs
        
        # make sure we have the correct siteType
        refreshPandaSpecs()

        job = app._getParent()
        masterjob = job._getRoot()

        logger.debug('ProdTransPandaRTHandler prepare called for %s',
                     job.getFQID('.'))

        job.backend.actualCE = job.backend.site
        job.backend.requirements.cloud = Client.PandaSites[job.backend.site]['cloud']

        # check that the site is in a submit-able status
        if not job.splitter or job.splitter._name != 'DQ2JobSplitter':
            allowed_sites = job.backend.list_ddm_sites()

        try:
            outDsLocation = Client.PandaSites[job.backend.site]['ddm']
            tmpDsExist = False
            if (configPanda['processingType'].startswith('gangarobot') or configPanda['processingType'].startswith('hammercloud')):
                #if Client.getDatasets(job.outputdata.datasetname):
                if getDatasets(job.outputdata.datasetname):
                    tmpDsExist = True
                    logger.info('Re-using output dataset %s'%job.outputdata.datasetname)
            if not configPanda['specialHandling']=='ddm:rucio' and not  configPanda['processingType'].startswith('gangarobot') and not configPanda['processingType'].startswith('hammercloud') and not configPanda['processingType'].startswith('rucio_test'):
                Client.addDataset(job.outputdata.datasetname,False,location=outDsLocation,allowProdDisk=True,dsExist=tmpDsExist)
            logger.info('Output dataset %s registered at %s'%(job.outputdata.datasetname,outDsLocation))
            dq2_set_dataset_lifetime(job.outputdata.datasetname, outDsLocation)
        except exceptions.SystemExit:
            raise BackendError('Panda','Exception in adding dataset %s: %s %s'%(job.outputdata.datasetname,sys.exc_info()[0],sys.exc_info()[1]))
        
        # JobSpec.
        jspec = JobSpec()
        jspec.currentPriority = app.priority
        jspec.jobDefinitionID = masterjob.id
        jspec.jobName = commands.getoutput('uuidgen 2> /dev/null')
        jspec.coreCount = app.core_count
        jspec.AtlasRelease = 'Atlas-%s' % app.atlas_release
        jspec.homepackage = app.home_package
        jspec.transformation = app.transformation
        jspec.destinationDBlock = job.outputdata.datasetname
        if job.outputdata.location:
            jspec.destinationSE = job.outputdata.location
        else:
            jspec.destinationSE = job.backend.site
        if job.inputdata:
            jspec.prodDBlock = job.inputdata.dataset[0]
        else:
            jspec.prodDBlock = 'NULL'
        if app.prod_source_label:
            jspec.prodSourceLabel = app.prod_source_label
        else:
            jspec.prodSourceLabel = configPanda['prodSourceLabelRun']
        jspec.processingType = configPanda['processingType']
        jspec.specialHandling = configPanda['specialHandling']
        jspec.computingSite = job.backend.site
        jspec.cloud = job.backend.requirements.cloud
        jspec.cmtConfig = app.atlas_cmtconfig
        if app.dbrelease == 'LATEST':
            try:
                latest_dbrelease = getLatestDBReleaseCaching()
            except:
                from pandatools import Client
                latest_dbrelease = Client.getLatestDBRelease()
            m = re.search('(.*):DBRelease-(.*)\.tar\.gz', latest_dbrelease)
            if m:
                self.dbrelease_dataset = m.group(1)
                self.dbrelease = m.group(2)
            else:
                raise ApplicationConfigurationError(None, "Error retrieving LATEST DBRelease. Try setting application.dbrelease manually.")
        else:
            self.dbrelease_dataset = app.dbrelease_dataset
            self.dbrelease = app.dbrelease
        jspec.jobParameters = app.job_parameters

        if self.dbrelease:
            if self.dbrelease == 'current':
                jspec.jobParameters += ' --DBRelease=current' 
            else:
                if jspec.transformation.endswith("_tf.py") or jspec.transformation.endswith("_tf"):
                    jspec.jobParameters += ' --DBRelease=DBRelease-%s.tar.gz' % (self.dbrelease,)
                else:
                    jspec.jobParameters += ' DBRelease=DBRelease-%s.tar.gz' % (self.dbrelease,)
                dbspec = FileSpec()
                dbspec.lfn = 'DBRelease-%s.tar.gz' % self.dbrelease
                dbspec.dataset = self.dbrelease_dataset
                dbspec.prodDBlock = jspec.prodDBlock
                dbspec.type = 'input'
                jspec.addFile(dbspec)
#.........这里部分代码省略.........

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
    def prepare(self,app,appconfig,appmasterconfig,jobmasterconfig):
        '''prepare the subjob specific configuration'''
 
        # PandaTools
        from pandatools import Client
        from pandatools import AthenaUtils
        from taskbuffer.JobSpec import JobSpec
        from taskbuffer.FileSpec import FileSpec

        job = app._getParent()
        logger.debug('AthenaMCPandaRTHandler prepare called for %s', job.getFQID('.'))
        
        try:
            assert self.outsite
        except:
            logger.error("outsite not set. Aborting")
            raise Exception()
        
        job.backend.site = self.outsite
        job.backend.actualCE = self.outsite
        cloud = job._getRoot().backend.requirements.cloud
        job.backend.requirements.cloud = cloud
        

        # now just filling the job from AthenaMC data
        
        jspec = JobSpec()
        jspec.jobDefinitionID   = job._getRoot().id
        jspec.jobName           = commands.getoutput('uuidgen 2> /dev/null')  
        jspec.AtlasRelease      = 'Atlas-%s' % app.atlas_rel
        
        if app.transform_archive:
            jspec.homepackage       = 'AnalysisTransforms'+app.transform_archive
        elif app.prod_release:
            jspec.homepackage       = 'AnalysisTransforms-AtlasProduction_'+str(app.prod_release)
        jspec.transformation    = '%s/runAthena-00-00-11' % Client.baseURLSUB
            
        #---->????  prodDBlock and destinationDBlock when facing several input / output datasets?

        jspec.prodDBlock    = 'NULL'
        if job.inputdata and len(app.inputfiles)>0 and app.inputfiles[0] in app.dsetmap:
            jspec.prodDBlock    = app.dsetmap[app.inputfiles[0]]

        # How to specify jspec.destinationDBlock  when more than one type of output is available? Panda prod jobs seem to specify only the last output dataset
        outdset=""
        for type in ["EVNT","RDO","HITS","AOD","ESD","NTUP"]:
            if type in app.outputpaths.keys():
                outdset=string.replace(app.outputpaths[type],"/",".")
                outdset=outdset[1:-1]
                break
        if not outdset:
            try:
                assert len(app.outputpaths.keys())>0
            except:
                logger.error("app.outputpaths is empty: check your output datasets")
                raise
            type=app.outputpaths.keys()[0]
            outdset=string.replace(app.outputpaths[type],"/",".")
            outdset=outdset[1:-1]
            
        jspec.destinationDBlock = outdset
        jspec.destinationSE = self.outsite
        jspec.prodSourceLabel   = 'user'
        jspec.assignedPriority  = 1000
        jspec.cloud             = cloud
        # memory
        if job.backend.requirements.memory != -1:
            jspec.minRamCount = job.backend.requirements.memory
        jspec.computingSite     = self.outsite
        jspec.cmtConfig         = AthenaUtils.getCmtConfig(athenaVer=app.atlas_rel)
#       library (source files)
        flib = FileSpec()
        flib.lfn            = self.library
#        flib.GUID           = 
        flib.type           = 'input'
#        flib.status         = 
        flib.dataset        = self.libDataset
        flib.dispatchDBlock = self.libDataset
        jspec.addFile(flib)

        #       input files FIXME: many more input types
        for lfn in app.inputfiles:
            useguid=app.turls[lfn].replace("guid:","")
            finp = FileSpec()
            finp.lfn            = lfn
            finp.GUID           = useguid
            finp.dataset        = app.dsetmap[lfn]
            finp.prodDBlock     = app.dsetmap[lfn]
            finp.prodDBlockToken = 'local'
            finp.dispatchDBlock = app.dsetmap[lfn]
            finp.type           = 'input'
            finp.status         = 'ready'
            jspec.addFile(finp)
        # add dbfiles if any:
        for lfn in app.dbfiles:
            useguid=app.dbturls[lfn].replace("guid:","")
            finp = FileSpec()
            finp.lfn            = lfn
            finp.GUID           = useguid
            finp.dataset        = app.dsetmap[lfn]
#.........这里部分代码省略.........

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]

#.........这里部分代码省略.........
                                         if unmergedDs.status in ['completed','cleanup','tobeclosed']:
                                             _logger.debug('%s skip %s due to status=%s' % (self.pandaID,unmergedDsName,unmergedDs.status))
                                         else:
                                             # set status
                                             unmergedDs.status = finalStatus
                                             # append to avoid repetition
                                             topUserDsList.append(unmergedDsName)
                                             # update DB
                                             retTopT = self.taskBuffer.updateDatasets([unmergedDs],withLock=True,withCriteria="status<>:crStatus",
                                                                                      criteriaMap={':crStatus':unmergedDs.status})
                                             if len(retTopT) > 0 and retTopT[0]==1:
                                                 _logger.debug('%s set %s to parent dataset : %s' % (self.pandaID,unmergedDs.status,unmergedDsName))
                                             else:
                                                 _logger.debug('%s failed to update parent dataset : %s' % (self.pandaID,unmergedDsName))
                     if self.pandaDDM and self.job.prodSourceLabel=='managed':
                         # instantiate SiteMapper
                         if self.siteMapper == None:
                             self.siteMapper = SiteMapper(self.taskBuffer)
                         # get file list for PandaDDM
                         retList = self.taskBuffer.queryFilesWithMap({'destinationDBlock':destinationDBlock})
                         lfnsStr = ''
                         guidStr = ''
                         for tmpFile in retList:
                             if tmpFile.type in ['log','output']:
                                 lfnsStr += '%s,' % tmpFile.lfn
                                 guidStr += '%s,' % tmpFile.GUID
                         if lfnsStr != '':
                             guidStr = guidStr[:-1]
                             lfnsStr = lfnsStr[:-1]
                             # create a DDM job
                             ddmjob = JobSpec()
                             ddmjob.jobDefinitionID   = int(time.time()) % 10000
                             ddmjob.jobName           = "%s" % commands.getoutput('uuidgen')
                             ddmjob.transformation    = 'http://pandaserver.cern.ch:25080/trf/mover/run_dq2_cr'
                             ddmjob.destinationDBlock = 'testpanda.%s' % ddmjob.jobName
                             ddmjob.computingSite     = "BNL_ATLAS_DDM"
                             ddmjob.destinationSE     = ddmjob.computingSite
                             ddmjob.currentPriority   = 200000
                             ddmjob.prodSourceLabel   = 'ddm'
                             ddmjob.transferType      = 'sub'
                             # append log file
                             fileOL = FileSpec()
                             fileOL.lfn = "%s.job.log.tgz" % ddmjob.jobName
                             fileOL.destinationDBlock = ddmjob.destinationDBlock
                             fileOL.destinationSE     = ddmjob.destinationSE
                             fileOL.dataset           = ddmjob.destinationDBlock
                             fileOL.type = 'log'
                             ddmjob.addFile(fileOL)
                             # make arguments
                             dstDQ2ID = 'BNLPANDA'
                             srcDQ2ID = self.siteMapper.getSite(self.job.computingSite).ddm
                             callBackURL = 'https://%s:%s/server/panda/datasetCompleted?vuid=%s&site=%s' % \
                                           (panda_config.pserverhost,panda_config.pserverport,
                                            dataset.vuid,dstDQ2ID)
                             _logger.debug(callBackURL)
                             # set src/dest
                             ddmjob.sourceSite      = srcDQ2ID
                             ddmjob.destinationSite = dstDQ2ID
                             # if src==dst, send callback without ddm job
                             if dstDQ2ID == srcDQ2ID:
                                 comout = commands.getoutput('curl -k %s' % callBackURL)
                                 _logger.debug(comout)
                             else:
                                 # run dq2_cr
                                 callBackURL = urllib.quote(callBackURL)
                                 # get destination dir

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
if len(sys.argv)>1:
    site = sys.argv[1]
else:
    site = None

datasetName = 'panda.destDB.%s' % commands.getoutput('uuidgen')

index = 0

job = JobSpec()
job.jobDefinitionID   = int(time.time()) % 10000
job.jobName           = "%s_%d" % (commands.getoutput('uuidgen'),index)
job.AtlasRelease      = 'Atlas-14.1.0\nAtlas-14.1.0'
job.homepackage       = 'AtlasProduction/14.1.0.3\nAtlasProduction/14.1.0.3'
job.transformation    = 'csc_digi_trf.py\ncsc_reco_trf.py'
job.destinationDBlock = datasetName

job.computingSite     = site

job.prodDBlock        = 'valid1.005200.T1_McAtNlo_Jimmy.simul.HITS.e322_s429_tid022081'
    
job.prodSourceLabel   = 'test'    
job.currentPriority   = 10000
job.cloud             = 'US'

for lfn in ['HITS.022081._00001.pool.root','HITS.022081._00002.pool.root']:
    fileI = FileSpec()
    fileI.dataset    = job.prodDBlock
    fileI.prodDBlock = job.prodDBlock
    fileI.lfn = lfn

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
files = {
    inputFile:None,
    }

jobList = []

index = 0
for lfn in files.keys():
    index += 1
    job = JobSpec()
    job.jobDefinitionID   = (time.time()) % 10000
    job.jobName           = "%s_%d" % (commands.getoutput('uuidgen'),index)
    job.AtlasRelease      = 'Atlas-15.3.1'
    job.homepackage       = 'AtlasProduction/15.3.1.5'
    job.transformation    = 'csc_atlasG4_trf.py'
    job.destinationDBlock = datasetName
    job.computingSite     = site
    job.prodDBlock        = prodDBlock
    
    job.prodSourceLabel   = 'test'
    job.processingType    = 'test'
    job.currentPriority   = 10000
    job.cloud             = cloud
    job.cmtConfig         = 'i686-slc4-gcc34-opt'

    fileI = FileSpec()
    fileI.dataset    = job.prodDBlock
    fileI.prodDBlock = job.prodDBlock
    fileI.lfn = lfn
    fileI.type = 'input'

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
    # AtlasRelease
    if len(re.findall('\.',trfVer)) > 2:
        match = re.search('^(\d+\.\d+\.\d+)',trfVer)
        job.AtlasRelease  = 'Atlas-%s' % match.group(1)
    else:
        job.AtlasRelease  = 'Atlas-%s' % trfVer
    # homepackage
    vers = trfVer.split('.')
    if int(vers[0]) <= 11:
        job.homepackage = 'JobTransforms'
        for ver in vers:
            job.homepackage += "-%02d" % int(ver)
    else:
        job.homepackage = 'AtlasProduction/%s' % trfVer
    # trf
    job.transformation    = trf
    job.destinationDBlock = oDatasets[0]
    # prod DBlock
    job.prodDBlock        = iDataset
    # souce lavel
    job.prodSeriesLabel   = 'pandatest'
    job.prodSourceLabel   = 'managed'
    # priority
    job.assignedPriority  = priority
    job.currentPriority   = priority
    # CPU, memory,disk ### FIXME

    # attempt number   ### FIXME

    # input files
    if iDataset != 'NULL':

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
def send_job(jobid, siteid):
    _logger.debug('Jobid: ' + str(jobid))

    site = sites_.get(siteid)

    job = jobs_.get(int(jobid))
    cont = job.container
    files_catalog = cont.files

    fscope = getScope(job.owner.username)
    datasetName = '{}:{}'.format(fscope, cont.guid)

    distributive = job.distr.name
    release = job.distr.release

    # Prepare runScript
    parameters = job.distr.command
    parameters = parameters.replace("$COMMAND$", job.params)
    parameters = parameters.replace("$USERNAME$", job.owner.username)
    parameters = parameters.replace("$WORKINGGROUP$", job.owner.working_group)

    # Prepare metadata
    metadata = dict(user=job.owner.username)

    # Prepare PanDA Object
    pandajob = JobSpec()
    pandajob.jobDefinitionID = int(time.time()) % 10000
    pandajob.jobName = cont.guid
    pandajob.transformation = client_config.DEFAULT_TRF
    pandajob.destinationDBlock = datasetName
    pandajob.destinationSE = site.se
    pandajob.currentPriority = 1000
    pandajob.prodSourceLabel = 'user'
    pandajob.computingSite = site.ce
    pandajob.cloud = 'RU'
    pandajob.VO = 'atlas'
    pandajob.prodDBlock = "%s:%s" % (fscope, pandajob.jobName)
    pandajob.coreCount = job.corecount
    pandajob.metadata = json.dumps(metadata)
    #pandajob.workingGroup = job.owner.working_group

    if site.encode_commands:
        # It requires script wrapper on cluster side
        pandajob.jobParameters = '%s %s %s "%s"' % (cont.guid, release, distributive, parameters)
    else:
        pandajob.jobParameters = parameters


    has_input = False
    for fcc in files_catalog:
        if fcc.type == 'input':
            f = fcc.file
            guid = f.guid
            fileIT = FileSpec()
            fileIT.lfn = f.lfn
            fileIT.dataset = pandajob.prodDBlock
            fileIT.prodDBlock = pandajob.prodDBlock
            fileIT.type = 'input'
            fileIT.scope = fscope
            fileIT.status = 'ready'
            fileIT.GUID = guid
            pandajob.addFile(fileIT)

            has_input = True
        if fcc.type == 'output':
            f = fcc.file
            fileOT = FileSpec()
            fileOT.lfn = f.lfn
            fileOT.destinationDBlock = pandajob.prodDBlock
            fileOT.destinationSE = pandajob.destinationSE
            fileOT.dataset = pandajob.prodDBlock
            fileOT.type = 'output'
            fileOT.scope = fscope
            fileOT.GUID = f.guid
            pandajob.addFile(fileOT)

            # Save replica meta
            fc.new_replica(f, site)

    if not has_input:
        # Add fake input
        fileIT = FileSpec()
        fileIT.lfn = "fake.input"
        fileIT.dataset = pandajob.prodDBlock
        fileIT.prodDBlock = pandajob.prodDBlock
        fileIT.type = 'input'
        fileIT.scope = fscope
        fileIT.status = 'ready'
        fileIT.GUID = "fake.guid"
        pandajob.addFile(fileIT)

    # Prepare lof file
    fileOL = FileSpec()
    fileOL.lfn = "%s.log.tgz" % pandajob.jobName
    fileOL.destinationDBlock = pandajob.destinationDBlock
    fileOL.destinationSE = pandajob.destinationSE
    fileOL.dataset = '{}:logs'.format(fscope)
    fileOL.type = 'log'
    fileOL.scope = 'panda'
    pandajob.addFile(fileOL)
#.........这里部分代码省略.........

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
files={'daq.m5_combined.0028997.Default.L1TT-b00000110.LB0000.SFO-1._0001.data':'M5.0028997.Default.L1TT-b00000110.RAW.v010803',}

if cloud=='IT':
  files={'daq.m5_combined.0029118.Default.L1TT-b00000010.LB0000.SFO-1._0001.data':'M5.0029118.Default.L1TT-b00000010.RAW.v010803'}


jobList = []

for i in range(1):
  for lfn in files.keys():
    job = JobSpec()
    job.jobDefinitionID   = int(time.time()) % 10000
    job.jobName           = commands.getoutput('uuidgen') 
    job.AtlasRelease      = 'Atlas-13.0.35'
    job.homepackage       = 'AtlasPoint1/13.0.35.1'
    job.transformation    = 'csc_cosmics_trf.py'
    job.destinationDBlock = datasetName
    job.cloud = cloud
    job.computingSite     = site
    job.prodDBlock        = files[lfn]
    job.prodSourceLabel   = 'test'
    job.currentPriority   = 1001

    fileI = FileSpec()
    fileI.dataset    = job.prodDBlock
    fileI.prodDBlock = job.prodDBlock
    fileI.lfn = lfn
    fileI.type = 'input'
    job.addFile(fileI)

    fileD = FileSpec()

# 需要导入模块: from taskbuffer.JobSpec import JobSpec [as 别名]
# 或者: from taskbuffer.JobSpec.JobSpec import transformation [as 别名]
for idx,argv in enumerate(sys.argv):
    if argv == '-s':
        aSrvID = sys.argv[idx+1]
        sys.argv = sys.argv[:idx]
        break

site = sys.argv[1]

datasetName = 'panda.destDB.%s' % commands.getoutput('uuidgen')
destName    = 'local'

job = JobSpec()
job.jobDefinitionID   = int(time.time()) % 10000
job.jobName           = "%s" % commands.getoutput('uuidgen')
# MPI transform on Titan that will run actual job                               
job.transformation    = '/lustre/atlas/proj-shared/csc108/transforms/mpi_wrapper_alice_A01alicegeo.py'

job.destinationDBlock = datasetName
job.destinationSE     = destName
job.currentPriority   = 1000
job.prodSourceLabel   = 'panda'
job.computingSite     = site
job.jobParameters     = " "
job.VO                = 'alice'

fileOL = FileSpec()
fileOL.lfn = "%s.job.log.tgz" % job.jobName
fileOL.destinationDBlock = job.destinationDBlock
fileOL.destinationSE     = job.destinationSE
fileOL.dataset           = job.destinationDBlock
fileOL.type = 'log'