RatesAnalysisOnlineProcessing Namespace Reference

Classes
class	RateEntry

Functions
def	readLbStartFromCool (runNumber, lb)
def	readLbEndFromCool (runNumber, lb)
def	readLbFromCool (runNumber, lb)
def	readChainsGroups (smkey="", dbAlias="")
def	readHLTPrescales (pskey="", dbAlias="")
def	readL1Prescales (pskey="", dbAlias="")
def	findKeysForRange (psk, lbStart, lbEnd)
def	toCSV (fileName, dirName, data)
def	saveMetadata (dirName, data)
def	readAvgLumi (runNumber, lbStart, lbEnd)
def	main ()

Variables
	log

Function Documentation

findKeysForRange()

def RatesAnalysisOnlineProcessing.findKeysForRange	(		psk,
			lbStart,
			lbEnd
	)

Retrieve all prescale keys combinations within given lumiblocks

Definition at line 135 of file RatesAnalysisOnlineProcessing.py.

def findKeysForRange(psk, lbStart, lbEnd):
    ''' 
    Retrieve all prescale keys combinations within given lumiblocks
    '''
    # Convert str to list of ntups
    it = iter("".join(c for c in psk if c not in "()[] ").split(","))
    pskList = [(int(x), int(y), int(z)) for x, y, z in zip(it, it, it)]
 
    log.debug("Reading the key ranges for {0}".format(psk))
 
    # COOL format of prescale entry is (key, lbStart, lbEnd)
    foundKeys = []
    for prescaleEntry in pskList:
        if prescaleEntry[1] <= lbEnd and prescaleEntry[2] >= lbStart:
            log.debug("Found prescale key in the range: {0}".format(prescaleEntry))
            foundKeys.append(prescaleEntry)
 
    if not foundKeys:
        log.warning("Cannot find one prescale for lumiblocks {0} to {1}. Available values: {2}".format(lbStart, lbEnd, pskList))
        return []
 
    return foundKeys
 
 

main()

def RatesAnalysisOnlineProcessing.main

(

)

Definition at line 217 of file RatesAnalysisOnlineProcessing.py.

def main():
    from argparse import ArgumentParser
    parser = ArgumentParser()
    parser.add_argument('--runNumber', required=True, type=int, 
                        help='Number of run to process')
    parser.add_argument('--lbStart', required=True, type=int, 
                        help='First lumiblock to record Enhanced Bias data')
    parser.add_argument('--lbEnd', required=True, type=int, 
                        help='Last lumiblock to record Enhanced Bias data')
    parser.add_argument('-s','--server', default='https://atlasop.cern.ch',
                        help="Pbeast server url. For GPN: https://atlasop.cern.ch, for P1: http://pc-tdq-bst-05.cern.ch:8080")
    parser.add_argument('--loglevel', type=int, default=3, 
                        help='Verbosity level: 1 - VERBOSE, 2 - DEBUG, 3 - INFO')
    args = parser.parse_args()
    log.setLevel(args.loglevel)
 
    startOfRange = readLbStartFromCool(args.runNumber, args.lbStart)
    endOfRange = readLbEndFromCool(args.runNumber, args.lbEnd)
 
    from time import ctime
    log.info("Rates will be retrieved for lumiblocks {0}-{1}: {2} - {3}".format(args.lbStart, args.lbEnd, ctime(startOfRange/1E6), ctime(endOfRange/1E6)))
    log.debug("Start and end timestamps: {0} {1}".format(startOfRange, endOfRange))
 
    # Read prescales and groups to save in the csv summary
    # The format of the configuration is defined in https://gitlab.cern.ch/atlas/athena/-/blob/main/Trigger/TrigCost/TrigCostAnalysis/python/CostMetadataUtil.py#L238
    from TrigCostAnalysis.CostMetadataUtil import readHLTConfigKeysFromCOOL
    configMetadata = readHLTConfigKeysFromCOOL(args.runNumber)
 
    # The start and stop lumiblock should not include the lumiblocks where we did the keys change
    hltAvailableKeys = findKeysForRange(configMetadata[3]["HLTPSK"], args.lbStart, args.lbEnd)
    l1AvailableKeys = findKeysForRange(configMetadata[4]["LVL1PSK"], args.lbStart, args.lbEnd)
 
    # Create list of key ranges - it will store tuples (l1Psk, hltPsk, rangeStart, rangeEnd)
    # Four cases are handled: L1 range starts before HLT range and ends later, HLT range starts before L1 range and ends later, 
    #       L1 range is withing HLT range, and HLT range is within L1 range
    keyRangesList = []
    for hltEntry in hltAvailableKeys:
        hltLbStart = hltEntry[1]
        hltLbEnd = hltEntry[2]
 
        for l1Entry in l1AvailableKeys:
            l1LbStart = l1Entry[1]
            l1LbEnd = l1Entry[2]
 
            if (hltLbStart >= l1LbStart) and (hltLbEnd >= l1LbEnd) and (hltLbStart <= l1LbEnd):
                keyRangesList.append((l1Entry[0], hltEntry[0], hltLbStart, l1LbEnd))
            elif (hltLbStart <= l1LbStart) and (hltLbEnd <= l1LbEnd) and (l1LbStart <= hltLbEnd):
                keyRangesList.append((l1Entry[0], hltEntry[0], l1LbStart, hltLbEnd))
            elif (hltLbStart <= l1LbStart) and (hltLbEnd >= l1LbEnd):
                keyRangesList.append((l1Entry[0], hltEntry[0], l1LbStart, l1LbEnd))
            elif (hltLbStart >= l1LbStart) and (hltLbEnd <= l1LbEnd):
                keyRangesList.append((l1Entry[0], hltEntry[0], hltLbStart, hltLbEnd))
 
    log.debug("Available key ranges are {0}".format(keyRangesList))
    chainGroups = readChainsGroups(configMetadata[2]["SMK"], configMetadata[0]["DB"])
 
    pbeast = None
    try:
        import libpbeastpy
        pbeast = libpbeastpy.ServerProxy(args.server)
    except ImportError as e:
        log.error("Exeption when reading the pbeast information. Remember to setup the tdaq release!\n{0}".format(e))
        return
 
    hltChains = {}
    l1Items = {}
    groups = {}
    for keysRange in keyRangesList:
        # Save the range start and end and make sure we are in given limits
        lbStart = keysRange[2] if keysRange[2] > args.lbStart else args.lbStart
        lbEnd = keysRange[3] if keysRange[3] < args.lbEnd else args.lbEnd
        startOfKeysRange = readLbStartFromCool(args.runNumber, lbStart)
        endOfKeysRange = readLbEndFromCool(args.runNumber, lbEnd)
        log.debug("Current range is {0}-{1}. Timestamps are {2}-{3}".format(lbStart, lbEnd, ctime(startOfKeysRange/1E6), ctime(endOfKeysRange/1E6)))
 
        hltPrescales = readHLTPrescales(keysRange[1], configMetadata[0]["DB"])
        l1Prescales = readL1Prescales(keysRange[0], configMetadata[0]["DB"])
 
        # Read prescaled IS rates from pbeast and calculate unprescaled rates
        # Queries are based on TRP grafana dashboard
        try:
            hltRates =  pbeast.get_data('ATLAS', 'HLT_Rate', 'Output', 'ISS_TRP.HLT_.*', True, startOfKeysRange, endOfKeysRange)[0].data
            l1Rates =  pbeast.get_data('ATLAS', 'L1_Rate', 'TAV', 'ISS_TRP.L1_.*', True, startOfKeysRange, endOfKeysRange)[0].data
            streamRates = pbeast.get_data('ATLAS', 'HLT_Rate', 'Output', 'ISS_TRP.str_.*', True, startOfKeysRange, endOfKeysRange)[0].data
            groupRates = pbeast.get_data('ATLAS', 'HLT_Rate', 'Output', 'ISS_TRP.grp_.*', True, startOfKeysRange, endOfKeysRange)[0].data
            recordingRates = pbeast.get_data('ATLAS', 'SFOngCounters', 'WritingEventRate', 'DF.TopMIG-IS:HLT.Counters.*', True, startOfKeysRange, endOfKeysRange)[0].data
        except RuntimeError as e:
            log.error("Exception when reading the pbeast information. Remember to export the pbeast server sso!\n{0}".format(e))
            return
 
        for chain in hltRates:
            chainName = chain.replace("ISS_TRP.", "")
            if chainName not in chainGroups:
                log.warning("Chain {0} is missing from the current menu".format(chainName))
                continue
 
            l1Item = "L1_" + chainName.split("L1")[1]
            l1psk = l1Prescales[l1Item] if l1Item in l1Prescales else "-"
            keysStr = "L1:{0} HLT:{1}".format(l1psk, hltPrescales[chainName])
            if chainName not in hltChains:
                hltChains[chainName] = RateEntry(chainName, chainGroups[chainName], keysStr)
            else:
                hltChains[chainName].appendKeys(keysStr)
            
            for dataPoint in hltRates[chain]:
                if dataPoint.value == 0:
                    # Skip avg rate = 0
                    continue
                elif dataPoint.ts < startOfKeysRange or dataPoint.ts > endOfKeysRange:
                    # Skip data point outside analyzed range
                    continue
 
                rate = dataPoint.value
                if rate > 0 and hltPrescales[chainName] <=0:
                    log.warning("Rate for disabled chain {0} is higher than 0! {1} timestamp {2}".format(chainName, rate, ctime(dataPoint.ts/1E6)))
 
                rateUn = rate * hltPrescales[chainName] * (l1Prescales[l1Item] if l1Item in l1Prescales else 1)
                hltChains[chainName].appendRate(rate, rateUn)
 
        for item in l1Rates:
            itemName = item.replace("ISS_TRP.", "")
            if "--enabled" not in itemName:
                continue
            else:
                itemName = itemName.replace("--enabled", "")
 
            if itemName not in l1Items:
                l1Items[itemName] = RateEntry(itemName, "-", "L1:{0}".format(l1Prescales[itemName]))
            else:
                l1Items[itemName].appendKeys("L1:{0}".format(l1Prescales[itemName]))
 
            for dataPoint in l1Rates[item]:
                if dataPoint.value == 0:
                    continue
                elif dataPoint.ts < startOfKeysRange or dataPoint.ts > endOfKeysRange:
                    continue
                rate = dataPoint.value
                if rate > 0 and l1Prescales[itemName] <=0:
                    log.warning("Rate for disabled chain {0} is higher than 0! {1} timestamp {2}".format(chainName, rate, ctime(dataPoint.ts/1E6)))
                rateUn = rate * l1Prescales[itemName]
 
                l1Items[itemName].appendRate(rate, rateUn)
 
        groupRates = groupRates | streamRates | recordingRates
        for group in groupRates:
            groupName = group.replace("ISS_TRP.", "") \
                            .replace("grp", "RATE") \
                            .replace("str", "STREAM") \
                            .replace("DF.TopMIG-IS:HLT.Counters.", "HLT_recording_")     
 
            if groupName not in groups:    
                groups[groupName] = RateEntry(groupName, "-", "Multiple")
 
            for dataPoint in groupRates[group]:
                if dataPoint.value == 0:
                    continue
                elif dataPoint.ts < startOfKeysRange or dataPoint.ts > endOfKeysRange:
                    continue
                groups[groupName].appendRate(dataPoint.value, 0)
 
    # Save the results
    from RatesAnalysis.Util import getTableName
    L1Table = getTableName("L1")
    HLTTable = getTableName("HLT")
    GroupTable = getTableName("Group")
 
    metadata = [
        {'PredictionLumi' : readAvgLumi(args.runNumber, args.lbStart, args.lbEnd)},
        {'RunNumber' : args.runNumber},
        {"First lumiblock" : args.lbStart},
        {"Last lumiblock" : args.lbEnd},
        {'SMK' :  configMetadata[2]["SMK"]},
        {'DB' : configMetadata[0]["DB"]},
        {'LVL1PSK' :  l1AvailableKeys},
        {'HLTPSK' : hltAvailableKeys}
    ]
 
    prescaledDirName = "costMonitoring_OnlineTRPRates-onlinePS-LB{0}-{1}_{2}/".format(args.lbStart, args.lbEnd, args.runNumber)
    unprescaledDirName = "costMonitoring_OnlineTRPRates-noPS-LB{0}-{1}_{2}/".format(args.lbStart, args.lbEnd, args.runNumber)
    log.info("Exporting " + HLTTable)
    hltChainsList = [hltChains[chain].getCsvEntry() for chain in hltChains]
    hltChainsUnpsList = [hltChains[chain].getUnprescaledCsvEntry() for chain in hltChains]
    toCSV(HLTTable, prescaledDirName + "csv/", hltChainsList)
    toCSV(HLTTable, unprescaledDirName + "csv/", hltChainsUnpsList)
 
    log.info("Exporting " + L1Table)
    l1ItemsList = [l1Items[item].getCsvEntry() for item in l1Items]
    l1ItemsUnpsList = [l1Items[item].getUnprescaledCsvEntry() for item in l1Items]
    toCSV(L1Table, prescaledDirName + "csv/", l1ItemsList)
    toCSV(L1Table, unprescaledDirName + "csv/", l1ItemsUnpsList)
 
    log.info("Exporting " + GroupTable)
    groupsList = [groups[group].getCsvEntry() for group in groups]
    toCSV(GroupTable, prescaledDirName + "csv/", groupsList)
 
    prescaledMd = [*metadata, {"Details" : "Averaged rates with online prescales from online monitoring"}]
    unprescaledMd = [*metadata, {"Details" : "Averaged rates with prescales removed from online monitoring"}]
    saveMetadata(prescaledDirName, prescaledMd)
    saveMetadata(unprescaledDirName, unprescaledMd)
 
 

readAvgLumi()

def RatesAnalysisOnlineProcessing.readAvgLumi	(		runNumber,
			lbStart,
			lbEnd
	)

Read Inst Lumi and calculate the average for given run number and lumiblock ranges

Definition at line 191 of file RatesAnalysisOnlineProcessing.py.

def readAvgLumi(runNumber, lbStart, lbEnd):
    ''' 
    Read Inst Lumi and calculate the average for given run number and lumiblock ranges
    '''
    from PyCool import cool
    from DQUtils.sugar import RunLumi
    from TrigConfStorage.TriggerCoolUtil import TriggerCoolUtil
    db = TriggerCoolUtil.GetConnection('CONDBR2')
    folder = db.getFolder("/TRIGGER/LUMI/OnlPrefLumi")
    folderIterator = folder.browseObjects(RunLumi(runNumber, lbStart), RunLumi(runNumber, lbEnd), cool.ChannelSelection())    
 
    avg = 0
    counter = 0
    while folderIterator.goToNext():
        payload=folderIterator.currentRef().payload()
        avg += payload["LBAvInstLumi"]
        counter +=1
 
    if counter == 0:
        log.error("No entries in COOL database, in /TRIGGER/LUMI/OnlPrefLumi folder were found for run {0}, lumiblocks {1}-{2}".format(runNumber, lbStart, lbEnd))
        return 0
 
    log.debug("Avg inst lumi {0} for {1} events".format((avg/counter), counter))
 
    return avg/counter * 1e30
 

readChainsGroups()

def RatesAnalysisOnlineProcessing.readChainsGroups	(		smkey = "",
			dbAlias = ""
	)

Returns dictionary of chain and it's groups based on HLT Menu

Definition at line 90 of file RatesAnalysisOnlineProcessing.py.

def readChainsGroups(smkey="", dbAlias=""):
    ''' 
    Returns dictionary of chain and it's groups based on HLT Menu
    '''
    log.debug("Reading HLT menu from {0} {1}".format(smkey, dbAlias))
    from TrigConfIO.HLTTriggerConfigAccess import HLTMenuAccess
    hltChains = HLTMenuAccess(dbalias = dbAlias, smkey = smkey).chains()
 
    groups = {}
    for chainName in hltChains:
        groups[chainName] = " ".join(hltChains[chainName]["groups"])
 
    return groups
 
 

readHLTPrescales()

def RatesAnalysisOnlineProcessing.readHLTPrescales	(		pskey = "",
			dbAlias = ""
	)

Returns dictionary of chain and it's prescale

Definition at line 105 of file RatesAnalysisOnlineProcessing.py.

def readHLTPrescales(pskey="", dbAlias=""):
    ''' 
    Returns dictionary of chain and it's prescale
    '''
    log.debug("Reading HLT prescales set from {0} {1}".format(pskey, dbAlias))
    from TrigConfIO.HLTTriggerConfigAccess import HLTPrescalesSetAccess
    prescalesSet = HLTPrescalesSetAccess(dbalias = dbAlias, hltpskey = pskey).prescales()
 
    prescales = {}
    for chainName in prescalesSet:
        prescales[chainName] = prescalesSet[chainName]["prescale"]
 
    return prescales
 
 

readL1Prescales()

def RatesAnalysisOnlineProcessing.readL1Prescales	(		pskey = "",
			dbAlias = ""
	)

Returns dictionary of item and it's prescale

Definition at line 120 of file RatesAnalysisOnlineProcessing.py.

def readL1Prescales(pskey="", dbAlias=""):
    ''' 
    Returns dictionary of item and it's prescale
    '''
    log.debug("Reading L1 prescales set from {0} {1}".format(pskey, dbAlias))
    from TrigConfIO.L1TriggerConfigAccess import L1PrescalesSetAccess
    prescalesAccess = L1PrescalesSetAccess(dbalias = dbAlias, l1pskey = pskey)
 
    prescales = {}
    for itemName in prescalesAccess.itemNames():
        prescales[itemName] = prescalesAccess.prescale(itemName)
 
    return prescales
 
 

readLbEndFromCool()

def RatesAnalysisOnlineProcessing.readLbEndFromCool	(		runNumber,
			lb
	)

Returns end of the lumiblock from given run in microseconds

Definition at line 62 of file RatesAnalysisOnlineProcessing.py.

def readLbEndFromCool(runNumber, lb):
    ''' 
    Returns end of the lumiblock from given run in microseconds
    '''
    timestamp = readLbFromCool(runNumber, lb)["EndTime"]
 
    # Returns value UTC nanoseconds since 1970  - convert to microseconds (used by IS)
    return int(timestamp/1000) if timestamp else -1
 
 

readLbFromCool()

def RatesAnalysisOnlineProcessing.readLbFromCool	(		runNumber,
			lb
	)

Returns payload for a lumiblock from given run read from COOL database
Based on TriggerCoolUtil.getRunStartTime

Definition at line 72 of file RatesAnalysisOnlineProcessing.py.

def readLbFromCool(runNumber, lb):
    ''' 
    Returns payload for a lumiblock from given run read from COOL database
    Based on TriggerCoolUtil.getRunStartTime
    '''
    from TrigConfStorage.TriggerCoolUtil import TriggerCoolUtil
    dbconn = TriggerCoolUtil.GetConnection("CONDBR2")
    f = dbconn.getFolder( "/TRIGGER/LUMI/LBLB" )
 
    limmin = RunLumi(runNumber, lb)
    limmax = RunLumi(runNumber, lb+1)
 
    from PyCool import cool
    objs = f.browseObjects(limmin, limmax, cool.ChannelSelection(0))
    objs.goToNext()
    return objs.currentRef().payload()
 
 

readLbStartFromCool()

def RatesAnalysisOnlineProcessing.readLbStartFromCool	(		runNumber,
			lb
	)

Returns start of the lumiblock from given run in microseconds

Definition at line 52 of file RatesAnalysisOnlineProcessing.py.

def readLbStartFromCool(runNumber, lb):
    ''' 
    Returns start of the lumiblock from given run in microseconds
    '''
    timestamp = readLbFromCool(runNumber, lb)["StartTime"]
 
    # Returns value UTC nanoseconds since 1970  - convert to microseconds (used by IS)
    return int(timestamp/1000) if timestamp else -1
 
 

saveMetadata()

def RatesAnalysisOnlineProcessing.saveMetadata	(		dirName,
			data
	)

Save metadata dictionary

Definition at line 176 of file RatesAnalysisOnlineProcessing.py.

def saveMetadata(dirName, data):
    ''' 
    Save metadata dictionary
    '''
 
    mdDict = {}
    mdDict['text'] = 'metadata'
    mdDict['children'] = data
 
    outFilename = dirName + '/' + "metadata.json"
    with open(outFilename, mode='w') as outputFile:
        import json
        json.dump(obj=mdDict, fp=outputFile, indent=2, sort_keys=True)
 
 

toCSV()

def RatesAnalysisOnlineProcessing.toCSV	(		fileName,
			dirName,
			data
	)

Save rates to csv file

Definition at line 159 of file RatesAnalysisOnlineProcessing.py.

def toCSV(fileName, dirName, data):
    ''' 
    Save rates to csv file
    '''
    import csv, os
 
    os.makedirs(dirName, exist_ok=True)
    with open(dirName + '/' + fileName, mode='w') as outputFile:
        ratesWriter = csv.writer(outputFile, delimiter=',', quotechar='"', quoting=csv.QUOTE_MINIMAL)
 
        ratesWriter.writerow(['Name','Group','Rate [Hz]','Rate Err [Hz]', 'Prescale'])
        ratesWriter.writerow(['Trigger name','The group this chain belongs to','Online rate','Error on rate','The prescale of this chain. Only displayed for simple combinations'])
 
        for trig in data:
            ratesWriter.writerow(trig)
 
 

Variable Documentation

log

RatesAnalysisOnlineProcessing.log

Definition at line 15 of file RatesAnalysisOnlineProcessing.py.