reconstruction.py

#!/usr/bin/env python3
 

 
# Limit the number of threads spawned by external libraries (e.g. XGBoost)
import os
os.environ['OMP_THREAD_LIMIT'] = "1"  # noqa
 
import basf2
 
from geometry import check_components
 
from svd import add_svd_reconstruction
from pxd import add_pxd_reconstruction
 
from rawdata import add_unpackers
 
from softwaretrigger.constants import ALWAYS_SAVE_OBJECTS, RAWDATA_OBJECTS, DEFAULT_HLT_COMPONENTS
 
from tracking import (
    add_mc_tracking_reconstruction,
    add_prefilter_tracking_reconstruction,
    add_postfilter_tracking_reconstruction,
    add_cr_tracking_reconstruction,
    add_prune_tracks,
)
 
from softwaretrigger.path_utils import (
    add_filter_software_trigger,
    add_skim_software_trigger
)
 
 
CDST_TRACKING_OBJECTS = (
    'EventLevelTrackingInfo',
    'RecoTracks',
    'Tracks',
    'V0s',
    'TrackFitResults',
    'EventT0',
    'CDCDedxHits',
    'CDCDedxTracks',
    'SVDShaperDigitsFromTracks',
    'PXDClustersFromTracks',
    'VXDDedxTracks',
    'CDCDedxLikelihoods',
    'VXDDedxLikelihoods'
)
 
 
DIGITS_OBJECTS = (
    'ARICHDigits',
    'CDCHits',
    'ECLDigits',
    'ECLDsps',
    'KLMDigits',
    'PXDDigits',
    'SVDEventInfoSim',
    'SVDShaperDigits',
    'TOPDigits'
)
 
 
def default_event_abort(module, condition, error_flag):
    """Default event abort outside of HLT: Ignore the error flag and just stop
    processing by giving an empty path"""
    p = basf2.Path()
    module.if_value(condition, p, basf2.AfterConditionPath.END)
 
 
def add_reconstruction(path, components=None, pruneTracks=True, add_trigger_calculation=True, skipGeometryAdding=False,
                       trackFitHypotheses=None, addClusterExpertModules=True,
                       with_cdc_cellular_automaton=False,
                       use_second_cdc_hits=False, add_muid_hits=False, reconstruct_cdst=None,
                       event_abort=default_event_abort, use_random_numbers_for_hlt_prescale=True,
                       pxd_filtering_offline=False,
                       create_intercepts_for_pxd_ckf=False,
                       append_full_grid_cdc_eventt0=True,
                       legacy_ecl_charged_pid=False, emulate_HLT=False,
                       skip_full_grid_cdc_eventt0_if_svd_time_present=True,
                       switch_off_slow_modules_for_online=False):
    """
    This function adds the standard reconstruction modules to a path.
    Consists of clustering, tracking and the PID modules essentially in this structure:
 
    | :func:`add_reconstruction()`
    | ├── :func:`add_prefilter_reconstruction()`
    | │ ├── :func:`add_prefilter_pretracking_reconstruction()`   : Clustering
    | │ ├── ``add_prefilter_tracking_reconstruction()``          : Tracking essential for HLT filter calculation
    | │ └── :func:`add_prefilter_posttracking_reconstruction()`  : PID and clustering essential for HLT
    | └── :func:`add_postfilter_reconstruction()`
    |   ├── ``add_postfilter_tracking_reconstruction()``         : Rest of the tracking
    |   └── :func:`add_postfilter_posttracking_reconstruction()` : Rest of PID and clustering
 
    plus the modules to calculate the software trigger cuts.
 
    :param path: Add the modules to this path.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param pruneTracks: Delete all hits except the first and last of the tracks after the V0Finder modules.
    :param skipGeometryAdding: Advances flag: The tracking modules need the geometry module and will add it,
        if it is not already present in the path. In a setup with multiple (conditional) paths however, it can not
        determine, if the geometry is already loaded. This flag can be used to just turn off the geometry adding at
        all (but you will have to add it on your own then).
    :param trackFitHypotheses: Change the additional fitted track fit hypotheses. If no argument is given,
        the fitted hypotheses are pion, kaon and proton, i.e. [211, 321, 2212].
    :param addClusterExpertModules: Add the cluster expert modules in the KLM and ECL. Turn this off to reduce
        execution time.
    :param with_cdc_cellular_automaton: If true, in the CDC track finding the cellular automaton algorithm will be used too,
        after the global algorithm (Legendre).
    :param use_second_cdc_hits: If true, the second hit information will be used in the CDC track finding.
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
    :param add_trigger_calculation: add the software trigger modules for monitoring (do not make any cut)
    :param reconstruct_cdst: None for mdst, 'rawFormat' to reconstruct cdsts in rawFormat, 'fullFormat' for the
        full (old) format. This parameter is needed when reconstructing cdsts, otherwise the
        required PXD objects won't be added.
    :param event_abort: A function to abort event processing at the given point. Should take three arguments: a module,
        the condition and the error_flag to be set if these events are kept. If run on HLT this will not abort the event
        but just remove all data except for the event information.
    :param use_random_numbers_for_hlt_prescale: If True, the HLT filter prescales are applied using randomly
        generated numbers, otherwise are applied using an internal counter.
    :param pxd_filtering_offline: If True, PXD data reduction (ROI filtering) is applied during the track reconstruction.
        The reconstructed SVD/CDC tracks are used to define the ROIs and reject all PXD clusters outside of these.
    :param create_intercepts_for_pxd_ckf: If True, the PXDROIFinder is added to the path to create PXDIntercepts to be used
        for hit filtering when creating the CKF relations. This independent of the offline PXD digit filtering which is
        steered by 'pxd_filtering_offline'. This can be applied for both data and MC.
    :param append_full_grid_cdc_eventt0: If True, the module FullGridChi2TrackTimeExtractor is added to the path
                                      and provides the CDC temporary EventT0.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
        MVA based charged particle ID (false).
    :param emulate_HLT: if True, it runs the reconstruction as it is run on HLT (e.g. without PXD).
        If you want to use this flag on raw data, you should also exclude the following branches from RootInput: ROIs, ROIpayload
    :param skip_full_grid_cdc_eventt0_if_svd_time_present: if true, and if also append_full_grid_cdc_eventt0 is true, the
        FullGridChi2TrackTimeExtractor is only executed in the events where no SVD-based EventT0 is found. If false, but
        append_full_grid_cdc_eventt0 is true, FullGridChi2TrackTimeExtractor will be executed in each event regardless of
        SVD EventT0 being present. Has no effect if append_full_grid_cdc_eventt0 is false. Default: true
    :param switch_off_slow_modules_for_online: if true, it switches off some modules in the reconstruction chain by overriding
        other flags (e.g.: this flag overrides ``append_full_grid_cdc_eventt0``. On HLT and ExpressReco, this flag is set
        to true in order to speed up the reconstruction.
    """
 
    # Set the run for beam data
    basf2.declare_beam()
 
    # By default, the FullGrid module is not used in the reconstruction chain.
    # It is needed for detectors that perform post-tracking calibration with respect to CDC EventT0 using cDST
    if reconstruct_cdst == 'rawFormat':
        append_full_grid_cdc_eventt0 = True
 
    if emulate_HLT:
        components = DEFAULT_HLT_COMPONENTS
 
    # If switch_off_slow_modules_for_online is True, we override some flags to make sure some slow modules are not executed
    if switch_off_slow_modules_for_online:
        append_full_grid_cdc_eventt0 = False
        legacy_ecl_charged_pid = True
 
    # pre-filter reconstruction
    add_prefilter_reconstruction(path,
                                 components=components,
                                 add_modules_for_trigger_calculation=add_trigger_calculation,
                                 skipGeometryAdding=skipGeometryAdding,
                                 trackFitHypotheses=trackFitHypotheses,
                                 with_cdc_cellular_automaton=with_cdc_cellular_automaton,
                                 use_second_cdc_hits=use_second_cdc_hits,
                                 add_muid_hits=add_muid_hits,
                                 reconstruct_cdst=reconstruct_cdst,
                                 event_abort=event_abort,
                                 pxd_filtering_offline=pxd_filtering_offline,
                                 create_intercepts_for_pxd_ckf=create_intercepts_for_pxd_ckf,
                                 append_full_grid_cdc_eventt0=append_full_grid_cdc_eventt0,
                                 skip_full_grid_cdc_eventt0_if_svd_time_present=skip_full_grid_cdc_eventt0_if_svd_time_present,
                                 switch_off_slow_modules_for_online=switch_off_slow_modules_for_online)
 
    # Add the modules calculating the software trigger cuts (but not performing them)
    if add_trigger_calculation and (not components or ("CDC" in components and "ECL" in components and "KLM" in components)):
        add_filter_software_trigger(path,
                                    use_random_numbers_for_prescale=use_random_numbers_for_hlt_prescale)
 
    # post-filter reconstruction
    add_postfilter_reconstruction(path,
                                  components=components,
                                  pruneTracks=pruneTracks,
                                  addClusterExpertModules=addClusterExpertModules,
                                  reconstruct_cdst=reconstruct_cdst,
                                  legacy_ecl_charged_pid=legacy_ecl_charged_pid,
                                  switch_off_slow_modules_for_online=switch_off_slow_modules_for_online)
 
    # Add the modules calculating the software trigger skims
    if add_trigger_calculation and (not components or ("CDC" in components and "ECL" in components and "KLM" in components)):
        add_skim_software_trigger(path)
 
 
def add_prefilter_reconstruction(path,
                                 components=None,
                                 add_modules_for_trigger_calculation=True,
                                 skipGeometryAdding=False,
                                 trackFitHypotheses=None,
                                 with_cdc_cellular_automaton=False,
                                 use_second_cdc_hits=False,
                                 add_muid_hits=False,
                                 reconstruct_cdst=None,
                                 event_abort=default_event_abort,
                                 pxd_filtering_offline=False,
                                 create_intercepts_for_pxd_ckf=False,
                                 append_full_grid_cdc_eventt0=True,
                                 skip_full_grid_cdc_eventt0_if_svd_time_present=True,
                                 switch_off_slow_modules_for_online=False):
    """
    This function adds only the reconstruction modules required to calculate HLT filter decision to a path.
    Consists of essential tracking and the functionality provided by :func:`add_prefilter_posttracking_reconstruction()`.
 
    :param path: Add the modules to this path.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param add_modules_for_trigger_calculation: add the modules necessary for computing the software trigger decision
        during later stages (do not make any cut), relevant only when reconstruct_cdst is not None.
    :param skipGeometryAdding: Advances flag: The tracking modules need the geometry module and will add it,
        if it is not already present in the path. In a setup with multiple (conditional) paths however, it can not
        determine, if the geometry is already loaded. This flag can be used to just turn off the geometry adding at
        all (but you will have to add it on your own then).
    :param trackFitHypotheses: Change the additional fitted track fit hypotheses. If no argument is given,
        the fitted hypotheses are pion, kaon and proton, i.e. [211, 321, 2212].
    :param with_cdc_cellular_automaton: If true, in the CDC track finding the cellular automaton algorithm will be used too,
        after the global algorithm (Legendre).
    :param use_second_cdc_hits: If true, the second hit information will be used in the CDC track finding.
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
    :param reconstruct_cdst: None for mdst, 'rawFormat' to reconstruct cdsts in rawFormat, 'fullFormat' for the
        full (old) format. This parameter is needed when reconstructing cdsts, otherwise the
        required PXD objects won't be added.
    :param event_abort: A function to abort event processing at the given point. Should take three arguments: a module,
        the condition and the error_flag to be set if these events are kept. If run on HLT this will not abort the event
        but just remove all data except for the event information.
    :param pxd_filtering_offline: If True, PXD data reduction (ROI filtering) is applied during the track reconstruction.
        The reconstructed SVD/CDC tracks are used to define the ROIs and reject all PXD clusters outside of these.
    :param create_intercepts_for_pxd_ckf: If True, the PXDROIFinder is added to the path to create PXDIntercepts to be used
        for hit filtering when creating the CKF relations. This independent of the offline PXD digit filtering which is
        steered by 'pxd_filtering_offline'. This can be applied for both data and MC.
    :param append_full_grid_cdc_eventt0: If True, the module FullGridChi2TrackTimeExtractor is added to the path
        and provides the CDC temporary EventT0.
    :param skip_full_grid_cdc_eventt0_if_svd_time_present: if true, and if also append_full_grid_cdc_eventt0 is true, the
        FullGridChi2TrackTimeExtractor is only executed in the events where no SVD-based EventT0 is found. If false, but
        append_full_grid_cdc_eventt0 is true, FullGridChi2TrackTimeExtractor will be executed in each event regardless of
        SVD EventT0 being present. Has no effect if append_full_grid_cdc_eventt0 is false. Default: true
    :param switch_off_slow_modules_for_online: if true, it switches off some modules in the reconstruction chain by overriding
        other flags (e.g.: this flag overrides ``append_full_grid_cdc_eventt0``. On HLT and ExpressReco, this flag is set
        to true in order to speed up the reconstruction.
    """
 
    # If switch_off_slow_modules_for_online is True, we override some flags to make sure some slow modules are not executed
    if switch_off_slow_modules_for_online:
        append_full_grid_cdc_eventt0 = False
 
    # Always avoid the top-level 'import ROOT'.
    from ROOT import Belle2  # noqa
 
    # Check components.
    check_components(components)
 
    # Do not even attempt at reconstructing events w/ abnormally large occupancy.
    doom = path.add_module("EventsOfDoomBuster")
    event_abort(doom, ">=1", Belle2.EventMetaData.c_ReconstructionAbort)
    path.add_module('StatisticsSummary').set_name('Sum_EventsofDoomBuster')
 
    # Add modules that have to be run BEFORE track reconstruction
    add_prefilter_pretracking_reconstruction(path, components=components)
 
    # Add prefilter tracking reconstruction modules
    add_prefilter_tracking_reconstruction(
        path,
        components=components,
        mcTrackFinding=False,
        skipGeometryAdding=skipGeometryAdding,
        trackFitHypotheses=trackFitHypotheses,
        with_cdc_cellular_automaton=with_cdc_cellular_automaton,
        use_second_cdc_hits=use_second_cdc_hits,
        pxd_filtering_offline=pxd_filtering_offline,
        create_intercepts_for_pxd_ckf=create_intercepts_for_pxd_ckf,
        append_full_grid_cdc_eventt0=append_full_grid_cdc_eventt0,
        skip_full_grid_cdc_eventt0_if_svd_time_present=skip_full_grid_cdc_eventt0_if_svd_time_present)
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_Prefilter_Tracking')
 
    # In case of cdst reconstruction ...
    if reconstruct_cdst:
        add_special_vxd_modules(path, components=components)
    if reconstruct_cdst == 'rawFormat' and not add_modules_for_trigger_calculation:
        return
 
    # Add prefilter posttracking modules
    add_prefilter_posttracking_reconstruction(path,
                                              components=components,
                                              add_muid_hits=add_muid_hits)
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_Prefilter_PostTracking')
 
 
def add_postfilter_reconstruction(path,
                                  components=None,
                                  pruneTracks=False,
                                  addClusterExpertModules=True,
                                  reconstruct_cdst=None,
                                  legacy_ecl_charged_pid=False,
                                  switch_off_slow_modules_for_online=False):
    """
    This function adds the reconstruction modules not required to calculate HLT filter decision to a path.
 
    :param path: Add the modules to this path.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param pruneTracks: Delete all hits expect the first and the last from the found tracks.
    :param addClusterExpertModules: Add the cluster expert modules in the KLM and ECL. Turn this off to
        reduce execution time.
    :param reconstruct_cdst: None for mdst, 'rawFormat' to reconstruct cdsts in rawFormat, 'fullFormat' for the
        full (old) format. This parameter is needed when reconstructing cdsts, otherwise the
        required PXD objects won't be added.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
      MVA based charged particle ID (false).
    :param switch_off_slow_modules_for_online: if true, it switches off some modules in the reconstruction chain by overriding
        other flags (e.g.: this flag overrides ``append_full_grid_cdc_eventt0``. On HLT and ExpressReco, this flag is set
        to true in order to speed up the reconstruction.
    """
 
    # If switch_off_slow_modules_for_online is True, we override some flags to make sure some slow modules are not executed
    flip_recoTrack = True
    kink_finding = True
    run_klm_dnn = True
    if switch_off_slow_modules_for_online:
        legacy_ecl_charged_pid = True
        flip_recoTrack = False
        kink_finding = False
        run_klm_dnn = False
 
    # Add postfilter tracking reconstruction modules
    add_postfilter_tracking_reconstruction(
        path,
        components=components,
        pruneTracks=False,
        flip_recoTrack=flip_recoTrack,
        kink_finding=kink_finding
    )
 
    path.add_module('StatisticsSummary').set_name('Sum_Postfilter_Tracking')
 
    # Skip postfilter posttracking modules except dedx for raw format cdst reconstruction
    if reconstruct_cdst == 'rawFormat':
        add_dedx_modules(
            path,
            components=components,
            enableDebugOutput=True
        )
        if pruneTracks:
            add_prune_tracks(
                path,
                components
            )
        return
 
    # Add postfilter posttracking modules
    add_postfilter_posttracking_reconstruction(
        path,
        components=components,
        addClusterExpertModules=addClusterExpertModules,
        legacy_ecl_charged_pid=legacy_ecl_charged_pid,
        run_klm_dnn=run_klm_dnn
    )
 
    # Prune tracks
    if pruneTracks:
        add_prune_tracks(
            path,
            components
        )
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_Postfilter_PostTracking')
 
 
def add_cosmics_reconstruction(
        path,
        components=None,
        pruneTracks=True,
        skipGeometryAdding=False,
        eventTimingExtraction=True,
        addClusterExpertModules=True,
        merge_tracks=True,
        use_second_cdc_hits=False,
        add_muid_hits=False,
        reconstruct_cdst=False,
        posttracking=True,
        legacy_ecl_charged_pid=False,
        ):
    """
    This function adds the standard reconstruction modules for cosmic data to a path.
    Consists of tracking and the functionality provided by :func:`add_posttracking_reconstruction()`,
    plus the modules to calculate the software trigger cuts.
 
    :param path: Add the modules to this path.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param pruneTracks: Delete all hits except the first and last of the tracks after the dEdX modules.
    :param skipGeometryAdding: Advances flag: The tracking modules need the geometry module and will add it,
        if it is not already present in the path. In a setup with multiple (conditional) paths however, it can not
        determine, if the geometry is already loaded. This flag can be used to just turn off the geometry adding at
        all (but you will have to add it on your own then).
 
    :param eventTimingExtraction: extract the event time
    :param addClusterExpertModules: Add the cluster expert modules in the KLM and ECL. Turn this off to reduce
        execution time.
 
    :param merge_tracks: The upper and lower half of the tracks should be merged together in one track
    :param use_second_cdc_hits: If true, the second hit information will be used in the CDC track finding.
 
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
 
    :param reconstruct_cdst: run only the minimal reconstruction needed to produce the cdsts (raw+tracking+dE/dx)
    :param posttracking: run reconstruction for outer detectors.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
      MVA based charged particle ID (false).
    """
 
    # Set the run for cosmics data
    basf2.declare_cosmics()
 
    # Check components.
    check_components(components)
 
    # Add modules that have to be run before track reconstruction
    add_prefilter_pretracking_reconstruction(path,
                                             components=components)
 
    # Add cdc tracking reconstruction modules
    add_cr_tracking_reconstruction(path,
                                   components=components,
                                   prune_tracks=False,
                                   skip_geometry_adding=skipGeometryAdding,
                                   event_time_extraction=eventTimingExtraction,
                                   merge_tracks=merge_tracks,
                                   use_second_cdc_hits=use_second_cdc_hits)
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_Tracking')
 
    if posttracking:
        if reconstruct_cdst:
            add_special_vxd_modules(path, components=components)
            add_dedx_modules(path, components=components, enableDebugOutput=True)
            add_prune_tracks(path, components=components)
 
        else:
            # Add further reconstruction modules
            add_posttracking_reconstruction(path,
                                            components=components,
                                            pruneTracks=pruneTracks,
                                            addClusterExpertModules=addClusterExpertModules,
                                            add_muid_hits=add_muid_hits,
                                            cosmics=True,
                                            legacy_ecl_charged_pid=legacy_ecl_charged_pid)
 
 
def add_mc_reconstruction(path, components=None, pruneTracks=True, addClusterExpertModules=True,
                          use_second_cdc_hits=False, add_muid_hits=False, legacy_ecl_charged_pid=False):
    """
    This function adds the standard reconstruction modules with MC tracking
    to a path.
 
    @param components list of geometry components to include reconstruction for, or None for all components.
    @param use_second_cdc_hits: If true, the second hit information will be used in the CDC track finding.
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
      MVA based charged particle ID (false).
    """
 
    # Set the run for beam data
    basf2.declare_beam()
 
    # Add modules that have to be run before track reconstruction
    add_prefilter_pretracking_reconstruction(path,
                                             components=components)
 
    # tracking
    add_mc_tracking_reconstruction(path,
                                   components=components,
                                   pruneTracks=False,
                                   use_second_cdc_hits=use_second_cdc_hits)
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_MC_Tracking')
 
    # add further reconstruction modules
    add_posttracking_reconstruction(path,
                                    components=components,
                                    pruneTracks=pruneTracks,
                                    add_muid_hits=add_muid_hits,
                                    addClusterExpertModules=addClusterExpertModules,
                                    legacy_ecl_charged_pid=legacy_ecl_charged_pid)
 
 
def add_prefilter_pretracking_reconstruction(path, components=None):
    """
    This function adds the standard reconstruction modules BEFORE tracking
    to a path.
 
    :param path: The path to add the modules to.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    """
 
    add_ecl_modules(path, components)
 
    # Statistics summary
    path.add_module('StatisticsSummary').set_name('Sum_Clustering')
 
 
def add_prefilter_posttracking_reconstruction(path,
                                              components=None,
                                              add_muid_hits=False,
                                              for_cdst_analysis=False,
                                              add_eventt0_combiner_for_cdst=False):
    """
    This function adds to the path the standard reconstruction modules after prefilter tracking
    whoose outputs are also needed in the filter.
 
    :param path: The path to add the modules to.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
    :param for_cdst_analysis: if True, EventT0Combiner is not added to path.
           This is only needed by prepare_cdst_analysis().
    :param add_eventt0_combiner_for_cdst: if True, the EventT0Combiner module is added to the path even if
           for_cdst_analysis is False. This is useful for validation purposes for avoiding to run the full
           add_reconstruction(). Note that, with the default settings (for_cdst_analysis=False and
           add_eventt0_combiner_for_cdst=False), the EventT0Combiner module is added to the path.
    """
 
    add_ext_module(path, components)
 
    # Add EventT0Combiner, if this function is not called from prepare_cdst_analysis() or if requested also there.
    if not for_cdst_analysis or add_eventt0_combiner_for_cdst:
        path.add_module("EventT0Combiner")
    add_ecl_finalizer_module(path, components)
    add_ecl_mc_matcher_module(path, components)
    add_klm_modules(path, components)
    add_klm_mc_matcher_module(path, components)
    add_muid_module(path, add_hits_to_reco_track=add_muid_hits, components=components)
    add_ecl_track_cluster_modules(path, components)
    add_ecl_cluster_properties_modules(path, components)
 
 
def add_postfilter_posttracking_reconstruction(path,
                                               components=None,
                                               addClusterExpertModules=True,
                                               cosmics=False,
                                               for_cdst_analysis=False,
                                               legacy_ecl_charged_pid=False,
                                               run_klm_dnn=True):
    """
    This function adds to the path the standard reconstruction modules whoose outputs are not needed in the filter.
 
    :param path: The path to add the modules to.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param addClusterExpertModules: Add the cluster expert modules in the KLM and ECL. Turn this off to reduce
        execution time.
    :param cosmics: if True, steer TOP for cosmic reconstruction.
    :param for_cdst_analysis: if True, the OnlineEventT0Creator module is not added to the path.
        This is only needed by prepare_cdst_analysis().
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
        MVA based charged particle ID (false). This flag is automatically set to true on HLT and ExpressReco.
    :param run_klm_dnn: If True, add the ``KLMMuonIDDNNExpert`` module to the path. This flag is automatically set to
        false on HLT and ExpressReco.
    """
 
    add_dedx_modules(path, components, for_cdst_analysis=for_cdst_analysis)
    add_top_modules(path, components, cosmics=cosmics)
    add_arich_modules(path, components)
 
    # only add the OnlineEventT0Creator if not preparing cDST
    if not for_cdst_analysis:
        path.add_module("OnlineEventT0Creator")
 
    add_ecl_chargedpid_module(path, components, legacy_ecl_charged_pid)
    add_pid_module(path, components, run_klm_dnn)
 
    if addClusterExpertModules:
        # FIXME: Disabled for HLT until execution time bug is fixed
        add_cluster_expert_modules(path, components)
 
    add_ecl_track_brem_finder(path, components)
 
 
def add_posttracking_reconstruction(path,
                                    components=None,
                                    pruneTracks=True,
                                    addClusterExpertModules=True,
                                    add_muid_hits=False,
                                    cosmics=False,
                                    for_cdst_analysis=False,
                                    add_eventt0_combiner_for_cdst=False,
                                    legacy_ecl_charged_pid=False):
    """
    This function adds the standard reconstruction modules after tracking
    to a path.
 
    :param path: The path to add the modules to.
    :param components: list of geometry components to include reconstruction for, or None for all components.
    :param pruneTracks: Delete all hits except the first and last after the post-tracking modules.
    :param addClusterExpertModules: Add the cluster expert modules in the KLM and ECL. Turn this off to reduce
        execution time.
    :param add_muid_hits: Add the found KLM hits to the RecoTrack. Make sure to refit the track afterwards.
    :param cosmics: if True, steer TOP for cosmic reconstruction.
    :param for_cdst_analysis: if True, the dEdx and PruneTracks modules are not added to the path, as well
           as all EventT0 related modules.
           This is only needed by prepare_cdst_analysis().
    :param add_eventt0_combiner_for_cdst: if True, the EventT0Combiner module is added to the path even if
           for_cdst_analysis is True. This is useful for validation purposes for avoiding to run the full
           add_reconstruction(). Note that, with the default settings (for_cdst_analysis=False and
           add_eventt0_combiner_for_cdst=False), the EventT0Combiner module is added to the path.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
      MVA based charged particle ID (false).
    """
 
    add_prefilter_posttracking_reconstruction(path,
                                              components=components,
                                              add_muid_hits=add_muid_hits,
                                              for_cdst_analysis=for_cdst_analysis,
                                              add_eventt0_combiner_for_cdst=add_eventt0_combiner_for_cdst)
 
    add_postfilter_posttracking_reconstruction(path,
                                               components=components,
                                               addClusterExpertModules=addClusterExpertModules,
                                               cosmics=cosmics,
                                               for_cdst_analysis=for_cdst_analysis,
                                               legacy_ecl_charged_pid=legacy_ecl_charged_pid)
 
    # Prune tracks as soon as the post-tracking steps are complete
    # Not add prune tracks modules in prepare_cdst_analysis()
    if not for_cdst_analysis:
        if pruneTracks:
            add_prune_tracks(path, components)
 
    path.add_module('StatisticsSummary').set_name('Sum_Posttracking_Reconstruction')
 
 
def add_cdst_output(path,
                    mc=True,
                    filename='cdst.root',
                    additionalBranches=None,
                    dataDescription=None,
                    ignoreInputModulesCheck=False):
    """
    This function adds the `RootOutput` module to a path with the settings needed to produce a cDST output.
    The actual cDST output content depends on the value of the parameter `mc`:
    * if `mc` is `False` (default setting), the cDST content is raw + tracking dataobjects;
    * if `mc` is `True`, the cDST content is digits + MCParticles + tracking dataobjects.
 
    @param path Path to add modules to.
    @param mc Define the type of cDST output content: `False` for raw + tracking dataobjects, `True` for digits +
           MCParticles + tracking dataobjects.
    @param filename Output file name.
    @param additionalBranches Additional objects/arrays of event durability to save
    @param dataDescription Additional key->value pairs to be added as data description
           fields to the output FileMetaData.
    @param ignoreInputModulesCheck If True, do not enforce check on missing PXD modules in the input path.
           Needed when a conditional path is passed as input.
    """
 
    branches = list(CDST_TRACKING_OBJECTS)
    persistentBranches = ['FileMetaData']
 
    if not mc:
        branches += ALWAYS_SAVE_OBJECTS + RAWDATA_OBJECTS
    else:
        branches += list(DIGITS_OBJECTS) + [
            'MCParticles',
            'EventLevelTriggerTimeInfo',
            'SoftwareTriggerResult',
            'TRGSummary']
        persistentBranches += ['BackgroundInfo']
 
    if not ignoreInputModulesCheck and "PXDClustersFromTracks" not in [module.name() for module in path.modules()]:
        basf2.B2ERROR("PXDClustersFromTracks is required in CDST output but its module is not found in the input path!")
 
    if dataDescription is None:
        dataDescription = {}
        dataDescription.setdefault("dataLevel", "cdst")
 
    if additionalBranches is not None:
        branches += additionalBranches
 
    return path.add_module("RootOutput", outputFileName=filename, branchNames=branches,
                           branchNamesPersistent=persistentBranches, additionalDataDescription=dataDescription)
 
 
def add_arich_modules(path, components=None):
    """
    Add the ARICH reconstruction to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'ARICH' in components:
        path.add_module('ARICHFillHits')
        path.add_module('ARICHReconstructor',
                        storePhotons=1)  # enabled for ARICH DQM plots
 
 
def add_top_modules(path, components=None, cosmics=False):
    """
    Add the TOP reconstruction to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param cosmics: if True, steer TOP for cosmic reconstruction
    """
    if components is None or 'TOP' in components:
        path.add_module('TOPChannelMasker')
        if cosmics:
            path.add_module('TOPCosmicT0Finder')
        else:
            path.add_module('TOPBunchFinder')
        path.add_module('TOPReconstructor')
 
 
def add_cluster_expert_modules(path, components=None):
    """
    Add the KLMExpert and ClusterMatcher modules to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or ('KLM' in components and 'ECL' in components):
        path.add_module('KLMExpert')
        path.add_module('ClusterMatcher')
 
 
def add_pid_module(path, components=None, run_klm_dnn=True):
    """
    Add the PID modules to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param run_klm_dnn: If True, add the ``KLMMuonIDDNNExpert`` module to the path.
        This flag is automatically set to false on HLT and ExpressReco.
    """
    if components is None or 'SVD' in components or 'CDC' in components:
        path.add_module('MdstPID')
    if (components is None or 'KLM' in components) and run_klm_dnn:
        path.add_module('KLMMuonIDDNNExpert')
 
 
def add_klm_modules(path, components=None):
    """
    Add the KLM reconstruction modules to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'KLM' in components:
        path.add_module('KLMReconstructor')
        path.add_module('KLMClustersReconstructor')
        path.add_module('KLMClusterAna')
 
 
def add_klm_mc_matcher_module(path, components=None):
    """
    Add the KLM mc matcher module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'KLM' in components:
        path.add_module('MCMatcherKLMClusters')
 
 
def add_muid_module(path, add_hits_to_reco_track=False, components=None):
    """
    Add the MuID module to the path.
 
    :param path: The path to add the modules to.
    :param add_hits_to_reco_track: Add the found KLM hits also to the RecoTrack. Make sure to refit the track afterwards.
    :param components: The components to use or None to use all standard components.
    """
    # Muid is needed for muonID computation AND ECLCluster-Track matching.
    if components is None or ('CDC' in components and 'ECL' in components and 'KLM' in components):
        path.add_module('Muid',
                        addHitsToRecoTrack=add_hits_to_reco_track)
    if components is not None and 'CDC' in components:
        if ('ECL' not in components and 'KLM' in components):
            basf2.B2WARNING('You added KLM to the components list but not ECL: the module Muid, that is necessary '
                            'for correct muonID computation, will not be added to your reconstruction path. '
                            'Make sure that this is fine for your purposes, otherwise please include also ECL.')
        if ('ECL' in components and 'KLM' not in components):
            basf2.B2WARNING('You added ECL to the components list but not KLM: the module Muid, that is necessary '
                            'for correct ECLCluster-Track matching, will not be added to your reconstruction path. '
                            ' Make sure that this is fine for your purposes, otherwise please include also KLM.')
 
 
def add_ecl_modules(path, components=None):
    """
    Add the ECL reconstruction modules to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'ECL' in components:
        path.add_module('ECLWaveformFit')
        path.add_module('ECLDigitCalibrator')
        path.add_module('ECLEventT0')
        path.add_module('ECLCRFinder')  # connected region finder
        path.add_module('ECLLocalMaximumFinder')
        path.add_module('ECLSplitterN1')
        path.add_module('ECLSplitterN2')
        path.add_module('ECLShowerCorrector')
        path.add_module('ECLShowerShape')
        path.add_module('ECLClusterPSD')
        path.add_module('ECLCovarianceMatrix')
        # The module ECLFinalizer must run after EventT0Combiner
 
 
def add_ecl_finalizer_module(path, components=None):
    """
        Add the ECL finalizer module to the path.
 
        :param path: The path to add the modules to.
        :param components: The components to use or None to use all standard components.
        """
 
    if components is None or 'ECL' in components:
        path.add_module('ECLFinalizer')
 
 
def add_ecl_track_cluster_modules(path, components=None):
    """
    Add the ECL track cluster matching module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or ('ECL' in components and ('PXD' in components or 'SVD' in components or 'CDC' in components)):
        path.add_module('ECLTrackClusterMatching')
 
 
def add_ecl_cluster_properties_modules(path, components=None):
    """
    Add the ECL cluster properties module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or ('ECL' in components and ('PXD' in components or 'SVD' in components or 'CDC' in components)):
        path.add_module('ECLClusterProperties')
 
 
def add_ecl_track_brem_finder(path, components=None):
    """
    Add the bremsstrahlung finding module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or ('ECL' in components and ('PXD' in components or 'SVD' in components)):
        path.add_module('ECLTrackBremFinder')
 
 
def add_ecl_chargedpid_module(path, components=None, legacyMode=False):
    """
    Add the ECL charged PID module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param legacyMode: Uses the simple E/p based charged PID instead of the MVA based charged PID.
        This flag is automatically set to true on HLT and ExpressReco.
    """
    if components is None or 'ECL' in components:
        # charged PID
        if legacyMode:
            path.add_module('ECLChargedPID')
        else:
            path.add_module('ECLFillCellIdMapping')
            path.add_module('ECLChargedPIDMVA')
 
 
def add_ecl_mc_matcher_module(path, components=None):
    """
    Add the ECL MC matcher module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'ECL' in components:
        path.add_module('MCMatcherECLClusters')
 
 
def add_ext_module(path, components=None):
    """
    Add the extrapolation module to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
    if components is None or 'CDC' in components:
        path.add_module('Ext')
 
 
def add_dedx_modules(path, components=None, for_cdst_analysis=False, enableDebugOutput=False):
    """
    Add the dE/dX reconstruction modules to the path.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param for_cdst_analysis: if True, add only DedxPIDCreator module, otherwise add both
    :param enableDebugOutput: enable/disable writing out debugging information to CDCDedxTracks
    """
    # CDC dE/dx PID
    if components is None or 'CDC' in components:
        if for_cdst_analysis:
            path.add_module('CDCDedxPIDCreator', enableDebugOutput=True)
        else:
            path.add_module('CDCDedxHitSaver')
            path.add_module('CDCDedxPIDCreator', enableDebugOutput=enableDebugOutput)
    # VXD dE/dx PID
    # only run this if the SVD is enabled - PXD is disabled by default
    if components is None or 'SVD' in components:
        if for_cdst_analysis:
            path.add_module('VXDDedxPIDRemaker')
        else:
            path.add_module('VXDDedxPID')
 
 
def add_special_vxd_modules(path, components=None):
    """
    Add two modules that are not part of the standard reconstruction.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    """
 
    if not components or ('PXD' in components):
        path.add_module("PXDClustersFromTracks")
    if not components or ('SVD' in components):
        path.add_module("SVDShaperDigitsFromTracks")
 
 
def prepare_cdst_analysis(path, components=None, mc=False, add_eventt0_combiner=False, legacy_ecl_charged_pid=False):
    """
    Adds to a (analysis) path all the modules needed to analyse a cDST file in the raw+tracking format
    for collisions/cosmics data or in the digits+tracking format for MC data.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param mc: Are we running over MC data or not? If so, do not run the unpackers.
    :param add_eventt0_combiner: If True, it adds the EventT0Combiner module when the post-tracking
      reconstruction is run. This must NOT be used during the calibration, but it may be necessary
      for validation purposes or for the user analyses.
    :param legacy_ecl_charged_pid: Bool denoting whether to use the legacy EoP based charged particleID in the ECL (true) or
      MVA based charged particle ID (false).
    """
    # Add the unpackers only if not running on MC, otherwise check the components and simply add
    # the Gearbox and the Geometry modules
    if not mc:
        add_unpackers(path,
                      components=components)
    else:
        check_components(components)
        path.add_module('Gearbox')
        path.add_module('Geometry')
 
    # This currently just calls add_ecl_modules
    add_prefilter_pretracking_reconstruction(path,
                                             components=components)
 
    # Needed to retrieve the PXD and SVD clusters out of the raw data
    if components is None or 'SVD' in components:
        add_svd_reconstruction(path)
    if components is None or 'PXD' in components:
        add_pxd_reconstruction(path)
 
    # check, this one may not be needed...
    path.add_module('SetupGenfitExtrapolation',
                    energyLossBrems=False,
                    noiseBrems=False)
 
    # Add the posttracking modules needed for the cDST analysis
    add_posttracking_reconstruction(path,
                                    components=components,
                                    for_cdst_analysis=True,
                                    add_eventt0_combiner_for_cdst=add_eventt0_combiner,
                                    legacy_ecl_charged_pid=legacy_ecl_charged_pid)
 
 
def prepare_user_cdst_analysis(path, components=None, mc=False):
    """
    Adds to a (analysis) path all the modules needed to analyse a cDST file in the raw+tracking format
    for collisions/cosmics data or in the digits+tracking format for MC data.
    Differently from prepare_cdst_analysis(), this function add the EventT0Combiner module to the path,
    which makes this function suitable for all the users and not only for the calibration expertes.
    Note that the EventT0Combiner module is necessary for applying the proper EventT0 correction to
    our data.
 
    :param path: The path to add the modules to.
    :param components: The components to use or None to use all standard components.
    :param mc: Are we running over MC data or not? If so, do not run the unpackers.
    """
    prepare_cdst_analysis(path=path, components=components, mc=mc, add_eventt0_combiner=True)