DictFromChainName.py

Go to the documentation of this file.

#! /bin/env python
# Copyright (C) 2002-2025 CERN for the benefit of the ATLAS collaboration
 
"""
Class to obtain the chain configuration dictionary from the short or long name
 
Authors of original code in TriggerMenu: Joerg Stelzer, Moritz Backes, Catrin Bernius
 
"""
__author__  = 'Catrin Bernius'
__version__=""
__doc__="Decoding of chain name into a dictionary"
 
from TriggerMenuMT.HLT.Menu.EventBuildingInfo import isRoIBasedPEB
from TrigConfHLTUtils.HLTUtils import string2hash
from AthenaCommon.Logging import logging
import re
from copy import deepcopy
 
import collections.abc
 
log = logging.getLogger( __name__ )
 
def getOverallL1item(flags, chainName):
    """
    Extracts the L1 seed from the chain name, checks for issue in naming, if found, fails
 
    In simple case the L1 items is just the last part of the name, after the _L1...
    There are though more complicated names like ...._L1...._L1..._L1... in such cases the L1 seed item is the last one
    """
    assert '_L1' in chainName, 'ERROR IN CHAIN {}, missing L1 seed at the end i.e. _L1...' .format(chainName)
 
    from TriggerMenuMT.HLT.Menu.L1Seeds import valid_multiseeds, getSpecificL1Seeds
    from TrigConfigSvc.TriggerConfigAccess import getL1MenuAccess
 
    # this assumes that the last string of a chain name is the overall L1 item
    cNameParts = chainName.rsplit("_L1",1)
    l1seed = 'L1_' + cNameParts[-1]
    # For reference, remapping of L1seeds lived originally in LVL1MenuConfig/LVL1Menu/L1Seeds.py
    # L1 collections can be read out from there in case they are needed again
    if l1seed == 'L1_All':
        return ''
    if l1seed == 'L1_test': #Multiseeded chains are build like this
        return 'L1_EM24VHI,L1_MU20'
    if l1seed in valid_multiseeds:
        # For these item seed specifications we need to derive the precise list of item names from the L1Menu.
        lvl1access = getL1MenuAccess(flags)
        itemsDict = lvl1access.items(includeKeys = ['name','ctpid','triggerType'])
        l1seedlist = getSpecificL1Seeds(l1seed, itemsDict, flags.Trigger.triggerMenuSetup)
        return l1seedlist
 
    return l1seed
 
def getL1item(flags, chainName):
    mainL1 = getOverallL1item(flags, chainName)
    return mainL1
 
def getAllThresholdsFromItem(item):
    """
    Breaks the item name into parts separated by the _ and from each element extract the threshold name
 
    Fails when used on topo items or ill formatted items
    Examples: L1_MU4 returns [MU4]
    L1_2MU4 returns [MU4, MU4]
    L1_MU4_MU6 returns [MU4, MU6]
    """
    assert item.startswith("L1_"), "The L1 item  {} name does not start with L1".format( item )
    assert "-" not in item, "The {} is Topo item and can not be parsed".format( item )
 
    multThressholds = item.split("_")[1:] # skips the L1_
    thresholds = []
    for mt in multThressholds:
        assert mt[0] != '1', 'Issue while parsing {}, multiplicity of 1, should not be configured, if multiplicity greater than 9 is needed then support for it needs to be added'.format( mt )
        if mt[0] in '23456789': # first letter is a number == multiplicity, multiplicity 1 is not specified
            thresholds.extend( int(mt[0])*[mt[1:]] )
        else:
            thresholds.append( mt ) # no muliplicity prefix
    return thresholds
 
def getUniqueThresholdsFromItem(item):
    """
    As above but eliminates repeated thresholds
    """
    allThreshold = getAllThresholdsFromItem(item)
    s = set()
    def _(t):
        s.add(t)
        return t
    return  [ _(t) for t in allThreshold if t not in s ]
 
def getEBPartFromParts( chainName, chainParts ):
    """
    If EB identifier is in the name, return it.
 
    Checks if there is only one identifier
    """
    # ---- event building identifier ----
    from TriggerMenuMT.HLT.Menu.EventBuildingInfo import getAllEventBuildingIdentifiers
    eventBuildTypes = set( getAllEventBuildingIdentifiers() ).intersection( chainParts )
    assert len(eventBuildTypes) <= 1, 'Chain {} has more than one Event Building identifier: {}, that is not supported'.format( chainName, eventBuildTypes)
    if eventBuildTypes:
        return eventBuildTypes.pop()
    return ''
 
 
def getChainThresholdFromName(chainParts, signature):
    """
    Decode threshold value from the chain name
    """
 
    from TriggerMenuMT.HLT.Menu.SignatureDicts import getBasePattern
    pattern = getBasePattern()
    trigType = []
    thresholdToPass = 0
 
    allThresh = []
    for cpart in chainParts:
        m = pattern.match( cpart )
        if m:
            log.debug("In getChainThresholdFromName: Pattern found in this string: %s", cpart)
            groupdict = m.groupdict()
            allThresh.append(groupdict['threshold'])
            trigType.append(groupdict['trigType'])
            if signature == groupdict['trigType']:
                thresholdToPass = groupdict['threshold']
                break
    return thresholdToPass
 
def getChainMultFromDict(chainDict):
    """
    Look for all multiplicities stored in chains
    """
    allMultis = []
   
    for cpart in chainDict['chainParts']:
        if cpart['multiplicity'] != '':
            allMultis.append( int(cpart['multiplicity']))
    return allMultis
 
 
# For certain chains, we need to include the L1 seed threshold
# in the leg name for disambiguation.
# This checks that the explicit threshold matches the threshold list
# In this case, the seed thresholds cannot be inferred from the L1 name
matchL1 = re.compile('L1([a-zA-Z0-9]+)(_.*)?')
def verifyExplicitL1Thresholds(chainname,chainparts,L1thresholds):
    assert len(L1thresholds)>0, f'L1 thresholds cannot be inferred from L1 name for chain {chainname} with explicit L1 seed thresholds'
 
    log.debug('Verifying explicit L1 thresholds for %s, received L1thresholds %s',chainname,L1thresholds)
 
    counter = 0
    for part, threshold in zip(chainparts,L1thresholds):
        if '_L1' in part:
            matches = matchL1.findall(part)
            assert len(matches)==1, f"Inappropriate number of L1 thresholds {matches} in chain part '{part}'"
            thisL1, remainder = matches[0] # list of tuples
            assert remainder == '', f"Explicit L1 threshold should be last element in chain part {part}"
            log.verbose('  ChainPart %s has explicit L1 threshold %s',part,thisL1)
            assert 'PROBE' not in thisL1, f"Omit 'PROBE' from explicit L1 threshold '{thisL1}'"
            assert thisL1 == threshold.replace('PROBE',''), f"Explicit L1 threshold for chainpart '{thisL1}' does not match threshold '{threshold}' from list"
            counter += 1
    assert counter>0, f"Did not find explicit L1 seeds in chain parts for {chainname}!"
 
def unifyJetRecoParts(chainParts):
    from TriggerMenuMT.HLT.Menu.SignatureDicts import getSignatureInformation, JetRecoKeys
 
    """
    Postprocess the jet chainParts to set all the reco config consistently
    For formatting reasons, we extract these from the last jet chainPart
    then propagate to all jet chainParts.
    Any inconsistencies in non-default options are considered a malformed chain
    Technically we can't tell if the default was explicitly set, though.
    """
 
    jetDefaults, _ = getSignatureInformation('Jet')
    # Extract the jet chainParts, and get the reco specs from the last one
    _jetRecoKeys = JetRecoKeys
    jetChainParts = [cPart for cPart in chainParts if cPart['signature'] in ['Jet','Bjet']]
    jetRecoDict = {k:jetChainParts[-1][k] for k in _jetRecoKeys}
 
    # Check consistency of the preceding jet parts
    for p in jetChainParts[:-1]:
        for k in _jetRecoKeys:
            # Assume that keys are all defined
            if p[k] != jetDefaults[k]: # OK if the spec is default
                log.error('Inconsistent jet reco setting for %s in chainPart %s', k, p['chainPartName'])
                log.error('Jet chainParts: %s', jetChainParts)
                raise RuntimeError('Jet reco should be specified only in the last jet chainPart')
 
    # Substitute the reco settings consistently and return
    newChainParts = []
    for cPart in chainParts:
        newCPart = deepcopy(cPart)
        if newCPart['signature'] in ['Jet','Bjet']:
            for k in _jetRecoKeys:
                newCPart[k] = jetRecoDict[k]
        newChainParts.append(newCPart)
 
    return newChainParts
 
def analyseChainName(chainName, L1thresholds, L1item):
    """
    Function to obtain the chain configuration dictionary from the short name by parsing its
    components and finding the corresponding properties which are defined in SignatureDicts.
    The naming convention is defined in https://twiki.cern.ch/twiki/bin/view/Atlas/TriggerNamingRun2
    and https://twiki.cern.ch/twiki/bin/view/Atlas/TriggerNamingRun3
    """
 
    # ---- dictionary with all chain properties ----
    from TriggerMenuMT.HLT.Menu.SignatureDicts import ChainDictTemplate
    from TriggerMenuMT.HLT.Menu.SignatureDicts import getSignatureInformation
    from TriggerMenuMT.HLT.Jet.JetRecoCommon import etaRangeAbbrev
    genchainDict = deepcopy(ChainDictTemplate)
    genchainDict['chainName'] = chainName
 
    # ---- remove the L1 item from the name ----
    hltChainName = chainName.rsplit("_L1",1)[0]
 
    # ---- check for HLT_HLT in name ---
    if 'HLT_HLT' in hltChainName:
        log.error("%s is incorrect, please fix the HLT_HLT part", hltChainName)
        raise RuntimeError("[analyseChainName] chain name contains HLT_HLT, please fix in corresponding menu where this chain is included")
 
    # ---- specific chain part information ----
    allChainProperties=[]
    cparts = hltChainName.split("_")
    if 'HLT' in hltChainName:
        cparts.remove('HLT')
 
    # ---- handle EB directive ----
    genchainDict['eventBuildType'] = getEBPartFromParts( chainName, cparts )
    if genchainDict['eventBuildType']:
        cparts.remove( genchainDict['eventBuildType'] )
 
    hltChainNameShort = '_'.join(cparts)
 
    # ---- identify the topo algorithm and add to genchainDict -----
    from TriggerMenuMT.HLT.Menu.SignatureDicts import AllowedTopos, AllowedTopos_comb, AllowedTopos_Bphysics_topoVariant, AllowedTopos_Bphysics_topoExtra
    topo = ''
    topos=[]
    extraComboHypos = []
    bphys_topoVariant=[]
    bphys_topoExtra = []
    bphysTopos = False 
    toposIndexed={}
    topoindex = -5
    for cindex, cpart in enumerate(cparts):
        #should make this if...elif...?
        if cpart in AllowedTopos:
            log.debug('" %s" is in this part of the name %s -> topo alg', AllowedTopos, cpart)
            topo = cpart
            topoindex = cindex
            toposIndexed.update({topo : topoindex})
            hltChainNameShort=hltChainNameShort.replace('_'+cpart, '')
            topos.append(topo)
        elif cpart in AllowedTopos_Bphysics_topoVariant:
            log.debug('[analyseChainName] chain part %s is a BLS topo variant, adding to bphys_topoVariant', cpart)
            bphys_topoVariant.append(cpart)
            toposIndexed.update({cpart : cindex})
            bphysTopos = True
        elif cpart in AllowedTopos_Bphysics_topoExtra:
            log.debug('[analyseChainName] chain part %s is a BLS extra topo hypo, adding to bphys_topoExtra', cpart)            
            bphys_topoExtra.append(cpart)
            toposIndexed.update({cpart : cindex})
            bphysTopos = True
        else:
            log.debug('[analyseChainName] chain part %s is not a general topo, BLS extra or variant topo hypo, checking comb topos next', cpart)            
        if cpart in AllowedTopos_comb:
             log.debug('[analyseChainName] chain part %s is a combined topo hypo, adding to extraComboHypo', cpart)
             toposIndexed.update({cpart : cindex})
             extraComboHypos.append(cpart)
 
    genchainDict['topo'] = topos
    genchainDict['extraComboHypos'] = extraComboHypos
 
    if bphysTopos is True:
        genchainDict['topoVariant'] = bphys_topoVariant
        genchainDict['topoExtra'] = bphys_topoExtra
        
    # remove the parts that have been already identified
    for t, i in enumerate(toposIndexed):
        if (t in cparts):
            log.debug('topo %s with index %s is going to be deleted from chain parts', t, i)
            del cparts[i]
 
 
    # ---- Find the signature defining patterns ----
    # ---- and write them out in dictionary     ----
    # ---- expected format: <Multiplicity(int)><TriggerType(str)>
    #      <Threshold(int)><isolation,...(str|str+int)> ----
    # EXCEPT FOR CHAINS ...
    from TriggerMenuMT.HLT.Menu.SignatureDicts import getBasePattern
    pattern = getBasePattern()
    mdicts=[]
    multichainindex=[]
 
 
    # ---- obtain dictionary parts for signature defining patterns ----
    from TriggerMenuMT.HLT.Menu.SignatureDicts import getSignatureNameFromToken, AllowedCosmicChainIdentifiers, \
        AllowedCalibChainIdentifiers, AllowedMonitorChainIdentifiers, AllowedBeamspotChainIdentifiers
    
    from TriggerMenuMT.HLT.Config.Utility.MenuAlignmentTools import get_alignment_group_from_pattern as getAlignmentGroupFromPattern
    
    def buildDict(signature, sigToken ):
        groupdict = {'signature': signature, 'threshold': '', 'multiplicity': '',
                     'trigType': sigToken, 'extra': ''}
        mdicts.append( groupdict )
 
    for cpart in cparts:
 
        log.debug("Looping over chain part: %s", cpart)
        m = pattern.match(cpart)
 
        if m:
            log.debug("Pattern found in this string: %s", cpart)
            groupdict = m.groupdict()
 
            multiChainIndices = [i for i in range(len(hltChainNameShort)) if ( hltChainNameShort.startswith(cpart, i) ) ]
            log.debug("MultiChainIndices: %s", multiChainIndices)
            for theMultiChainIndex in multiChainIndices:
                # this check is necessary for the bjet chains, example: j45_bloose_3j45
                # j45 would be found in [0, 13], and 3j45 in [12]
                # so need to make sure the multiplicities are considered here!
                if (theMultiChainIndex != 0) & (hltChainNameShort[theMultiChainIndex-1] != '_'):
                    continue
 
                if theMultiChainIndex not in multichainindex:
                    multichainindex.append(theMultiChainIndex)
 
                    log.debug("HLTChainName: %s", hltChainName)
                    log.debug("HLTChainNameShort: %s", hltChainNameShort)
                    log.debug("cpart: %s", cpart)
                    log.debug("groupdict: %s", groupdict)
                    log.debug("multichainindex: %s", multichainindex)
 
                sName = getSignatureNameFromToken(cpart)
            
            groupdict['signature'] = sName
            log.debug("groupdict['signature']: %s",groupdict['signature'])
       
            mdicts.append(groupdict)
 
        elif cpart =='noalg':
            multichainindex.append(hltChainNameShort.index(cpart))
            buildDict( 'Streaming', 'streamer')
            break # stop loop here so that further parts like noalg_idmon are discarded, this allows have copies with different output streams and prescales. Caveat: all noalg chains go into the Streaming slice
        elif cpart =='acceptedevts':
            multichainindex.append(hltChainNameShort.index(cpart))
            buildDict( 'Calib', 'calib')
            break # stop loop here so that further parts like acceptedevts_larnoiseburst are discarded. They are only used to define different prescales
        else:
            for chainCategory in [(['mb'], 'MinBias', 'mb'),
                                  (['hi'], 'HeavyIon', 'mb'),
                                  (AllowedCosmicChainIdentifiers, 'Cosmic', 'cosmic'),
                                  (AllowedCalibChainIdentifiers, 'Calib', 'calib'),
                                  (AllowedMonitorChainIdentifiers, 'Monitor', 'calib'),
                                  (AllowedBeamspotChainIdentifiers, 'Beamspot', 'beamspot'),
                                  (['eb'], 'EnhancedBias', 'eb')]:
                if cpart in chainCategory[0]:
                    log.debug('Doing chain type %s', chainCategory[1])
                    multichainindex.append(hltChainNameShort.index(cpart))
                    buildDict(chainCategory[1], chainCategory[2])
 
 
    # If multiple parts exist, split the string and analyse each
    # part depending on the signature it belongs to
    multichainparts=[]
    multichainindex = sorted(multichainindex, key=int)
    cN = deepcopy(hltChainNameShort)
    for i in reversed(multichainindex):
        if i!=0:
            log.debug('Appending to multichainparts (i!=0): %s', hltChainNameShort[i:len(cN)])
 
            multichainparts.append(hltChainNameShort[i:len(cN)])
            cN = cN[0:i-1]
        else:
            log.debug('Appending to multichainparts: %s', hltChainNameShort[i:len(cN)])
            multichainparts.append(cN)
    log.debug("multichainparts: %s",multichainparts)
 
    # build the chainProperties dictionary for each part of the chain
    # add it to a allChainProperties
    multichainparts.reverse()
    log.debug('multichainparts after reverse: %s', multichainparts)
 
    # verify if the L1 setup is consistent with the chain
    # sizes of seeds matter
    if len(L1thresholds) != 0:
        assert len(L1thresholds) ==  len(multichainparts), 'ERROR IN CHAIN {} definition, l1 thresholds specified {} have different length than chain parts {}'\
            .format(chainName, str(L1thresholds), str(multichainparts) )
 
    # tmp removed this check since the L1seeds depend on sym or asym cases
   # if len(L1thresholds) == 0:
        # getAllThresholdsFromItem, getUniqueThresholdsFromItem
    #    assert len( getAllThresholdsFromItem( L1item )) == len(multichainparts), 'ERROR IN CHAIN {} definition, l1 thresholds extracted from the L1 item name {} have different length than chain parts {}, set L1 thresholds ChainProp'\
     #       .format( chainName, str( getAllThresholdsFromItem( L1item ) ), str(multichainparts))
 
    # check the case when _L1 appears more than once in the name
    if chainName.count("_L1") > 1:
        verifyExplicitL1Thresholds(chainName,multichainparts,L1thresholds)
 
    for chainindex, chainparts in enumerate(multichainparts):
        chainProperties = {} #will contain properties for one part of chain if multiple parts
        chainpartsNoL1 = chainparts.rsplit('_L1',1)[0] if 'L1' in chainparts else chainparts
 
        if len(L1thresholds) != 0:
            chainProperties['L1threshold'] = L1thresholds[chainindex]
        else:
            __th = getAllThresholdsFromItem ( L1item )
            assert chainindex < len(__th), "In defintion of the chain {chainName} there is not enough thresholds to be used, index: {chainindex} >= number of thresholds, thresholds are: {__th}"
            chainProperties['L1threshold'] = __th[chainindex]  #replced getUniqueThresholdsFromItem
 
        parts=chainpartsNoL1.split('_')
        log.debug("[analyseChainName] chain parts w/o L1 are %s", parts)
        if None in parts:
            log.error("[analyseChainName] chainpartsNoL1 -> parts: %s -> %s", chainpartsNoL1, parts)
            raise Exception("[analyseChainName] parts contains None, please identify how this is happening")
        parts=list(filter(None,parts))
        log.debug("[analyseChainName] chainpartsNoL1 %s, parts %s", chainpartsNoL1, parts)
 
        # --- check for duplicate strings in chain part name ---
        duplicateParts = ([item for item, count in collections.Counter(parts).items() if count > 1])
        log.debug("[analyseChainName] chainpartsNoL1 %s contains duplicate strings %s", chainpartsNoL1, duplicateParts)
        if duplicateParts:
            log.error("[analyseChainName] chain %s has duplicate strings %s, please fix!!", chainName, duplicateParts)
            raise RuntimeError("[analyseChainName] Check the chain name, there are duplicate configurations: %s", duplicateParts)
 
        chainProperties['trigType']=mdicts[chainindex]['trigType']
        if 'extra' in chainProperties:
            raise RuntimeError("[analyseChainName] Overwrote 'extra' value, only one supported!")
        else:
            chainProperties['extra']=mdicts[chainindex]['extra']
        multiplicity = mdicts[chainindex]['multiplicity'] if not mdicts[chainindex]['multiplicity'] == '' else '1'
        chainProperties['multiplicity'] = multiplicity
        chainProperties['threshold']=mdicts[chainindex]['threshold']
        chainProperties['signature']=mdicts[chainindex]['signature']
       
        # if we have a L1 topo in a multi-chain then we want to remove it from the chain name
        # but only if it's the same as the L1item_main; otherwise it belongs to chain part and we q
        # have to keep it in the name
        chainProperties['chainPartName'] = chainparts
        if ('-' in L1item) and (len(multichainparts) > 1):
            chainProperties['chainPartName'] = chainpartsNoL1
        
        if len(multichainparts) > 1 and L1item.count("_") > 1 :
            chainProperties['chainPartName'] = chainpartsNoL1
 
        log.debug('[analyseChainname] Chainparts: %s', chainparts)
        if (chainProperties['signature'] != 'Cosmic') \
                & (chainProperties['signature'] != 'Calib')\
                & (chainProperties['signature'] != 'Streaming') \
                & (chainProperties['signature'] != 'Beamspot') \
                & (chainProperties['signature'] != 'Monitor') :
            parts.pop(0)
 
 
        #---- Check if topo is a bphysics topo -> change signature ----
        from TriggerMenuMT.HLT.Menu.SignatureDicts import AllAllowedTopos_Bphysics
        for t in genchainDict['topo']:
            if (t in AllAllowedTopos_Bphysics):
                chainProperties['signature'] = 'Bphysics'
                if "tnpInfo" in chainProperties.keys() and chainProperties['tnpInfo'] != "":
                    chainProperties['alignmentGroup'] = getAlignmentGroupFromPattern('Bphysics',chainProperties['tnpInfo'])
                else:
                    chainProperties['alignmentGroup'] = getAlignmentGroupFromPattern('Bphysics',chainProperties['extra'])
 
        # ---- import the relevant dictionaries for each part of the chain ----
        SignatureDefaultValues, allowedSignaturePropertiesAndValues = getSignatureInformation(chainProperties['signature'])
        log.debug('SignatureDefaultValues: %s', SignatureDefaultValues)
        allDefaults = list(SignatureDefaultValues.values())
        log.debug('All default values in a list')
 
        # ---- update chain properties with default properties ----
        result = deepcopy(SignatureDefaultValues)
        result.update(chainProperties)
        chainProperties = result
 
        # ---- check that all parts to be matched are not specified as defaults already ----
        overlaps = [x for x in parts if x in allDefaults]
        log.debug("parts that are also in defaults: %s ", overlaps)
 
        # Suppress the warning until we fully decomission the old DeepSet tau triggers in a couple of months
        if overlaps == ['tracktwoMVA']: overlaps = []
 
        if overlaps:
            #log.error("[analyseChainName] The following string(s) is/are already defined as defaults, please remove: %s", overlaps)
            #raise RuntimeError("[analyseChainname] Default config appearing in chain name, please remove: %s", overlaps)
            log.warning("[analyseChainName] The following chainpart %s contains the string(s) %s which is/are already defined as defaults, please remove!", chainparts, overlaps)
 
        # ---- check remaining parts for complete matches in allowedPropertiesAndValues Dict ----
        # ---- unmatched = list of tokens that are not found in the allowed values as a whole ----
        matchedparts = []
        log.debug("[analyseChainname] parts to match are %s", parts)
        for pindex, part in enumerate(parts):
            for prop, allowedValues in allowedSignaturePropertiesAndValues.items():
                if part in allowedValues:
                    if type(chainProperties[prop]) is list:
                        chainProperties[prop] += [part]
                    else:
                        chainProperties[prop] = part
                    matchedparts.append(part)
 
        # Jet-specific operations
        if chainProperties['signature']=='Jet':
 
        # ----- at this point we can figure out if the chain is a bJet chain and update defaults accordingly
            if chainProperties['bTag'] != '':
                log.debug('Setting b-jet chain defaults')
                # b-jet chain, so we now use the bJet defaults if they have not already been overriden
                bJetDefaultValues, allowedbJetPropertiesAndValues = getSignatureInformation('Bjet')
                for prop, value in bJetDefaultValues.items():
                    propSet=False
                    for value in allowedbJetPropertiesAndValues[prop]:
                        if value in matchedparts:
                            propSet=True
                            break
 
                    # if the property was not set already, then set if according to the b-jet defaults
                    if propSet is False:
                        log.debug('Changing %s from %s to %s', prop, str(chainProperties[prop]), str(bJetDefaultValues[prop]))
                        chainProperties[prop] = bJetDefaultValues[prop]
            
            
            if chainProperties['ditauTag'] != '':
                log.debug('Setting Ditau chain defaults')
                # boosted ditau chain, so we now use the ditau defaults if they have not already been overriden
                ditauJetDefaultValues, allowedDitauJetPropertiesAndValues = getSignatureInformation('Ditau')
                for prop, value in ditauJetDefaultValues.items():
                    propSet=False
                    for value in allowedDitauJetPropertiesAndValues[prop]:
                        if value in matchedparts:
                            propSet=True
                            break
 
                    # if the property was not set already, then set if according to the ditau defaults
                    if propSet is False:
                        log.debug('Changing %s from %s to %s', prop, str(chainProperties[prop]), str(ditauJetDefaultValues[prop]))
                        chainProperties[prop] = ditauJetDefaultValues[prop]
 
 
            if chainProperties['signature'] == 'Jet' and chainProperties['beamspotChain'] != '':
                log.debug('Setting beamspot chain defaults')
                BeamspotDefaultValues, allowedBeamspotPropertiesAndValues = getSignatureInformation('Beamspot_Jet')
                
                for prop, value in BeamspotDefaultValues.items():
                    propSet=False
                    for value in allowedBeamspotPropertiesAndValues[prop]:
                        if value in matchedparts:
                            propSet=True
                            break
 
                    # if the property was not set already, then set if according to the b-jet defaults
                    if propSet is False:
                        log.debug('Changing %s from %s to %s', prop, str(chainProperties[prop]), str(BeamspotDefaultValues[prop]))
                        chainProperties[prop] = BeamspotDefaultValues[prop]
                        
 
            # Substitute after b-jet defaults have been set otherwise they will be overridden
            if chainProperties['extra']:
                # Atypical handling here: jets translate the 1 char suffix to an eta range
                try:
                    etarange = etaRangeAbbrev[chainProperties['extra']]
                    chainProperties['etaRange'] = etarange
                except KeyError:
                    raise RuntimeError(f"Invalid jet 'extra' value {chainProperties['extra']} in {chainProperties['chainPartName']}. Only [a,c,f] allowed")
 
        # ---- checking the parts that haven't been matched yet ----
        log.debug("matched parts %s", matchedparts)
        leftoverparts = set(parts)-set(matchedparts)
        log.debug('leftoverparts %s', leftoverparts)
        for pindex, part in enumerate(leftoverparts):
            for prop, allowedValues in allowedSignaturePropertiesAndValues.items():
                if prop in chainProperties.keys():
                    continue
                for aV in allowedValues:
                    if (aV in part):
                        if (chainProperties['signature'] in ['Egamma', 'Muon'] )& (prop in ['trkInfo','hypoInfo']):
                            chainProperties[prop] = part
                            part = part.replace(part,'')
                        else:
                            chainProperties[prop] = aV
                            part = part.replace(aV,'')
                            break # done with allowed values for that property
 
            assert len(part.split()) == 0, "These parts of the chain name {} are not understood: {}".format(chainpartsNoL1,part)
 
 
        # ---- remove properties that aren't allowed in the chain properties for a given siganture ----
        forbiddenProperties = set(chainProperties.keys()) - set(allowedSignaturePropertiesAndValues.keys())
        log.debug('chainProperties: %s', sorted(set(chainProperties.keys())))
        log.debug('allowedSignaturePropertiesAndValues: %s', sorted(set(allowedSignaturePropertiesAndValues.keys())))
        for fb in forbiddenProperties:
            forbiddenValue = chainProperties.pop(fb)
            assert forbiddenValue == '', "Property {} not allowed for signature '{}', but specified '{}'".format (fb, chainProperties['signature'], forbiddenValue)
 
        # ---- set the alignment group here, once we have fully worked out the properties ----
        # ---- this is for the benefit of the probe leg in T&P chains ----
   
        if any(x in chainName for x in ('larnoiseburst', 'idcalib', 'metcalo', 'mettrk')):
            chainProperties['alignmentGroup'] = 'JetMET'
        else:
            if 'tnpInfo' in chainProperties.keys() and chainProperties['tnpInfo'] != "":
                chainProperties['alignmentGroup'] = getAlignmentGroupFromPattern(mdicts[chainindex]['signature'],chainProperties['tnpInfo'])
            else:
                chainProperties['alignmentGroup'] = getAlignmentGroupFromPattern(mdicts[chainindex]['signature'],chainProperties['extra'])
                
        # ---- the info of the general and the specific chain parts dict ----
        allChainProperties.append(chainProperties)
 
    # ---- depending on if signatures are different in this chain, break up the chainProperties dictionary ----
    # ---- finally also taking care of the signature key ----
    genchainDict['chainParts'] = allChainProperties
    for cPart in allChainProperties:
        if cPart['signature'] == 'Jet' and cPart['bTag'] != '':
            cPart['signature'] = 'Bjet'
            if 'tnpInfo' in cPart.keys() and cPart['tnpInfo'] != "":
                cPart['alignmentGroup'] = getAlignmentGroupFromPattern('Bjet', cPart['tnpInfo'])
            else:
                cPart['alignmentGroup'] = getAlignmentGroupFromPattern('Bjet', cPart['extra'])
        if cPart['signature'] == 'Jet' and cPart['ditauTag'] != '':
            cPart['signature'] = 'Tau'
            if 'tnpInfo' in cPart.keys() and cPart['tnpInfo'] != "":
                cPart['alignmentGroup'] = getAlignmentGroupFromPattern('Tau', cPart['tnpInfo'])
            else:
                cPart['alignmentGroup'] = getAlignmentGroupFromPattern('Tau', cPart['extra'])
        genchainDict['signatures'] += [cPart['signature']]
        genchainDict['alignmentGroups'] += [cPart['alignmentGroup']]
 
    # Postprocess chains with multiple jet chainParts 
    # to set them consistently from only one specification
    if len(genchainDict['chainParts'])>1:
        if 'Jet' in genchainDict['signatures'] or 'Bjet' in genchainDict['signatures']:
            genchainDict['chainParts'] = unifyJetRecoParts(genchainDict['chainParts'])
        
    #genchainDict['signature'] = allChainProperties[0]['signature']
    return genchainDict
 
 
def flattenChainGroups(A):
    rt = []
    for i in A:
        if isinstance(i,list): rt.extend(flattenChainGroups(i))
        else: rt.append(i)
    return rt
 
def checkChainStream(myStream):
    if len(myStream) == 1:
        if myStream[0] == 'express':
            return False
    return True
 
def dictFromChainName(flags, chainInfo):
    """
    Transforms ChainProp into the ChainDict
 
    The logic:
    ---- Loop over all chains (keys) in dictionary ----
    ---- Then complete the dict with other info    ----
    Default input format will be a ChainProp object:
    ChainProp(name: str, groups: list[str], l1SeedThresholds: list[str] = [], stream: list[str] = ['Main'], 
                monGroups: list[str] = [], mergingStrategy: str = 'auto', mergingOrder: list[str] = [], mergingOffset: int = -1),
    but for nwo plain chain name is also supported
    """
 
    # these if/elif/else statements are due to temporary development
    if type(chainInfo) is str:
        chainName       = chainInfo
        l1Thresholds    = []
        stream          = ''
        groups          = []
        mergingStrategy = 'parallel'
        mergingOffset   = -1
        mergingOrder    = []
        monGroups       = []
 
    elif 'ChainProp' in str(type(chainInfo)):   
        #this is how we define chains in the menu - the normal behaviour of this function
        chainName       = chainInfo.name
        l1Thresholds    = chainInfo.l1SeedThresholds
        stream          = chainInfo.stream
        groups          = flattenChainGroups(chainInfo.groups)
        mergingStrategy = chainInfo.mergingStrategy
        mergingOffset   = chainInfo.mergingOffset
        mergingOrder    = chainInfo.mergingOrder
        monGroups       = chainInfo.monGroups
 
    else:
        raise RuntimeError("Format of chainInfo passed to genChainDict not known")
 
    #check here the content of the stream
    if not checkChainStream(stream):
        raise RuntimeError("Chain {}, format of chainInfo:stream {} is not valid".format(chainName, stream))
 
    L1item = getL1item(flags, chainName)
 
    log.debug("Analysing chain with name: %s", chainName)
    chainDict = analyseChainName(chainName,  l1Thresholds, L1item)
    log.debug('ChainProperties: %s', chainDict)
 
    for chainPart in chainDict['chainParts']:
        # fill the sigFolder and subSigs folder
        for sf in chainPart['sigFolder']:
            if sf in chainDict['sigDicts']:
                chainDict['sigDicts'][sf].extend(chainPart['subSigs'])                
            else:
                chainDict['sigDicts'].update({sf:chainPart['subSigs']})
            if sf == 'Bjet' or 'Ditau' in chainPart['subSigs']:
                if 'Jet' not in chainDict['sigDicts']:
                    chainDict['sigDicts'].update({'Jet':['Jet']})
                
 
        thisSignature = chainPart['signature']
        thisSubSigs = chainPart['subSigs']
        thisAlignGroup = chainPart['alignmentGroup']
        thisExtra = chainPart['extra']
        thisL1 = chainPart['L1threshold']
        thisChainPartName = chainPart['chainPartName']
 
        if thisSignature in ['Muon','Bphysics']:
            if 'MuonnoL1' in thisAlignGroup or 'lateMu' in thisExtra:
                if 'FSNOSEED' not in thisL1:
                    log.error("Muon noL1 and lateMu chain should be seeded from FSNOSEED. Check %s seeded from %s (defined L1: %s), signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
            else:
                if 'MU' not in thisL1:
                    log.error("Standard muon and Bphysics chain should be seeded from L1_MU. Check %s seeded from %s (defined L1: %s), signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
                    #incorrectL1=True
 
        if thisSignature in ['Electron','Photon'] and not ('UPC' in stream): # don't apply for HI UPC chains
            if 'EM' not in thisL1 and 'BKee' not in thisL1 and 'All' not in thisL1:
                log.error("Standard egamma chains should be seeded from L1_EM. Check %s seeded from %s (defined L1: %s),  signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
                #incorrectL1=True
 
        if thisSignature in ['Tau'] and 'Ditau' not in thisSubSigs:
            if 'TAU' not in thisL1:
                log.error("Standard tau chains should be seeded from L1_TAU. Check %s seeded from %s (defined L1: %s), signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
                #incorrectL1=True
 
        if thisSignature in ['Jet','Bjet','MET','UnconventionalTracking'] or (thisSignature =='Tau' and 'Ditau' in thisSubSigs):
            if 'FSNOSEED' not in thisL1:
                log.error("Jet, b-jet, MET and Tau:Ditau chains should be seeded from FSNOSEED. Check %s seeded from %s (defined L1: %s),  signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
                #incorrectL1=True
 
        if thisChainPartName in ['noalg']:
            # All streamers should be unseeded except RoI-based PEB streamers which need a real RoI for PEB
            if 'FSNOSEED' not in thisL1 and not isRoIBasedPEB(chainDict['eventBuildType']):
                log.error("noalg chains should be seeded from FSNOSEED. Check %s seeded from %s (defined L1: %s),  signature %s",chainDict['chainName'],thisL1,l1Thresholds,thisSignature)
                #incorrectL1=True
 
#        if incorrectL1 is True:
#            raise Exception("You are using incorrect L1 seed, please check for ERROR messages...")
 
 
        log.debug('Parts: signature %s, align %s, extra %s, L1 %s', thisSignature, thisAlignGroup, thisExtra, thisL1)
 
 
 
    # setting the L1 item
    chainDict['L1item']          = L1item
    chainDict['stream']          = stream
    chainDict['groups']          = groups
    chainDict['mergingStrategy'] = mergingStrategy
    chainDict['mergingOffset']   = mergingOffset
    chainDict['mergingOrder']    = mergingOrder
    chainDict['monGroups']       = monGroups
    chainDict['chainNameHash']   = string2hash(chainDict['chainName'])
 
 
    allChainMultiplicities = getChainMultFromDict(chainDict)
    chainDict['chainMultiplicities'] = allChainMultiplicities
    log.debug('Setting chain multiplicities: %s ', allChainMultiplicities)
    
    
 
    return chainDict
 
def main():
 
    from AthenaConfiguration.AllConfigFlags import initConfigFlags
    flags = initConfigFlags()
 
    parser = flags.getArgumentParser()
    parser.add_argument(
        'chain',
        type=str,
        help='Chain name to be parsed into chain dictionary')
    parser.add_argument(
        '-t', '--thresholds',
        nargs='+',
        help='L1 thresholds, needed for multileg or fullscan chains')
    args = flags.fillFromArgs(parser=parser)
    flags.lock()
 
    if args.thresholds:
        from TriggerMenuMT.HLT.Config.Utility.ChainDefInMenu import ChainProp
        cd = dictFromChainName(flags, ChainProp(args.chain,l1SeedThresholds=args.thresholds,groups=[]))
    else:
        cd = dictFromChainName(flags, args.chain)
    
    from pprint import pprint
    pprint(cd)
 
    import json
    with open(f'{args.chain}.dict.json','w') as f:
        json.dump(cd, f, indent=2)
 
    return 0
 
if __name__=='__main__':
    main()