IOPool/Input/scripts/edmDumpEventContent - master - githubmirror/cmssw

IOPool/Input/scripts/edmDumpEventContent
241 строка9 KB
Matti Kortelainen
Move edmDumpEventContent to IOPool/Input
22 май 2026, 23:43
22 май 2026, 23:431ef7651
#! /usr/bin/env python3

from builtins import object
import os
import sys
import copy
from argparse import ArgumentParser
import re
import copy
import subprocess

# define regex
wrapperRE      = re.compile (r'edm::Wrapper<(.+)\s*>$')
trailingDotRE  = re.compile (r'\.$')
underscoreRE   = re.compile (r'_')
leadingSlashRE = re.compile (r'^/')
webRE          = re.compile (r'^\w+://')

def getOutput (command):
    '''Given an input command, returns STDOUT and STDERR'''
    proc = subprocess.Popen(command,
                            shell=True,
                            stdout=subprocess.PIPE,
                            stderr=subprocess.PIPE
                            )
    return proc.communicate()


def expandFilename (filename, options):
    '''Intelligently guesses as to whether or not it is given an PFN
    or LFN.  Returns PFN.'''
    # if this is a web address, don't modify it
    if webRE.search (filename):
        return filename
    # if it doesn't start with a slash, return it as is
    if not leadingSlashRE.search (filename):
        return filename
    # if the file exists locally AND 'lfn' option is not given, return
    # as is
    if not options.lfn and os.path.exists (filename):
        return filename
    # OK.  If we're here, then we probably have an LFN.  Convert it
    stdout, stderr = getOutput ('edmFileUtil -d %s' % filename)
    if stdout:
        return stdout.strip()
    raise RuntimeError("Could not successfully run 'edmFileUtil.' Error %s" \
          % stderr)


class Branch (object):
    # default options
    form = "%(type)5s  %(module)5s  %(label)5s  %(process)5s"
    mode = 'normal'
    forceColumns = False
    inQuotes = ['module', 'label', 'process']

    def __init__ (self, branchInfo, regexList = None):
        """Takes the needed information from Root's Branch Info"""
        match = wrapperRE.match( branchInfo.GetTypeName() )
        if match:
            self.type = match.group (1)
        else:
            # this isn't a type we're interested in
            raise ValueError("Not edm::Wrapper")
        name = trailingDotRE.sub ('', branchInfo.GetName())
        pieces = underscoreRE.split (name)
        if len (pieces) != 4:
            raise ValueError("%s not formatted as expected" % name)
        self.typeString = pieces[0] # not currently used
        self.module     = pieces[1]
        self.label      = pieces[2]
        self.process    = pieces[3]
        self.fullname   = "_".join (pieces)
        self.splitlevel = branchInfo.GetSplitLevel()
        # if we're past a regex list, then make sure we've got a match
        if not regexList:
            # nothing more to do
            return
        found = False
        for regex in regexList:
            # search the branch name
            if regex.search (self.fullname):
                found = True
                break
            if regex.search (self.type):
                found = True
                break
        if not found:
            raise ValueError("%s does not match regexs provided" \
                  % self.fullname)

        
    def __str__ (self):
        """String representation """
        copyDict = copy.copy (self.__dict__)
        for key in Branch.inQuotes:
            copyDict[key] = '"%s"' % copyDict[key]
        return Branch.form % copyDict
    
    
    @staticmethod
    def _setForm (branchList):
        '''Loop through lists and set widths and form appropriately'''
        # each item in this dictionary is a variable in Branch where
        # we want to keep track of the length.  The first number in
        # the list is the maximum length of all of the Branch items in
        # the list.  The second entry is the maximum length that it
        # should take IFF we are enforcing maximum length.
        lengthDict = { 'type'    : [7, 35],
                       'module'  : [7, 25],
                       'label'   : [7, 15],
                       'process' : [7, 15],
                       'fullname': [7, 50]}
        # Since we want this script to print out the columns in the
        # order that we want and not some random order, we specify
        # that order here.
        order = ['type', 'module', 'label', 'process']
        for obj in branchList:
            for key, twoList in lengthDict.items():
                attribute = getattr (obj, key)
                if len (attribute) + 2 > twoList[0]:
                    twoList[0] = len (attribute) + 2
        totalLength = 0
        for twoList in list(lengthDict.values()):
            totalLength += twoList[0]
        useMin = True
        if Branch.forceColumns or totalLength < 70:
            useMin = False
        Branch.form = ""
        if Branch.mode == 'name':
            # name
            Branch.form += '%(fullname)s'
        else:
            for piece in order:
                Branch.form += '%('
                Branch.form +=  piece
                if useMin:
                    Branch.form += ')-%ds   ' % min ( lengthDict[piece] )
                else:
                    Branch.form += ')-%ds   ' % lengthDict[piece][0]
        if Branch.mode == 'all':
            # name
            Branch.form += '     %('
            if useMin:
                Branch.form += 'fullname)-%ds' % min ( lengthDict['fullname'] )
            else:
                Branch.form += 'fullname)-%ds' % lengthDict['fullname'][0]
            # split level
            Branch.form += '     %(splitlevel)s'


    @staticmethod
    def printList (branchList):
        '''Print out nicely formatted list of all branches'''
        Branch._setForm (branchList)
        title = Branch.form % {'type'     : 'Type',
                               'module'   : 'Module',
                               'label'    : 'Label',
                               'process'  : 'Process',
                               'fullname' : 'Full Name',
                               'splitlevel' : 'Split level'}
        print(title)
        print('-' * len(title))
        for branch in branchList:
            print(branch)

#branches = ''
if __name__ == "__main__":

    parser = ArgumentParser(description="Prints out info on edm file.")
    nameAllGroup = parser.add_mutually_exclusive_group()
    nameAllGroup.add_argument('--name', dest='name', action='store_true',
                              help='print out only branch names')
    nameAllGroup.add_argument('--all',  dest='all',  action='store_true',
                              help='Print out everything: type, module, label, '\
                              'process, branch name, and branch split level')
    parser.add_argument('--lfn', dest='lfn', action='store_true',
                        help="Force LFN2PFN translation (usually not necessary)")
    lumiRunGroup = parser.add_mutually_exclusive_group()
    lumiRunGroup.add_argument('--lumi', dest='lumi', action='store_true',
                              help="Look at 'lumi' tree")
    lumiRunGroup.add_argument('--run', dest='run', action='store_true',
                              help="Look at 'run' tree")
    parser.add_argument("--regex", dest='regex', action="append",
                        type=str, default=[],
                        help="Filter results based on regex")
    parser.add_argument('--skipping', dest='skipping', action='store_true',
                        help="Print out branches being skipped")
    parser.add_argument('--forceColumns', dest='forceColumns',
                        action='store_true',
                        help="Forces printouts to be in nice columns")
    parser.add_argument("templates", metavar="templates.root", type=str)
    options = parser.parse_args()
    filename = expandFilename (options.templates, options)
    ###################
    # process options #
    ###################
    if options.name:
        Branch.mode = 'name'
    elif options.all:
        Branch.mode = 'all'
    Branch.forceColumns = options.forceColumns

    treeName = 'Events'
    if options.lumi:
        treeName = 'LuminosityBlocks'
    elif options.run:
        treeName = 'Runs'
    regexList = []
    # are we asked to filter this list?
    for regexString in options.regex:
        regexList.append( re.compile( regexString, re.IGNORECASE ) )
    # Because PyRoot is, well, greedy, we want to make sure we have
    # setup and parsed the command line options before importing ROOT
    # otherwise ROOT will try and do this for us.
    import ROOT
    ROOT.gROOT.SetBatch() # setting batch mode
    # Here we turn off stderr so that we don't get all of the errors
    # saying we don't have a dictionary for all of the objects in the
    # root file.  When we've loaded the file, we turn it back on
    oldStderr = os.dup( sys.stderr.fileno() )
    newStderr = open ( '/dev/null', 'w')
    os.dup2( newStderr.fileno(), sys.stderr.fileno() )
    rootfile = ROOT.TFile.Open( filename )
    os.dup2( oldStderr, sys.stderr.fileno() )
    if not rootfile:
        raise RuntimeError("Could not open file '%s'." % filename)
    Tree = rootfile.Get (treeName)
    branches = Tree.GetListOfBranches()
    branchList = []
    for branchInfo in branches:
        try:
            branch = Branch (branchInfo, regexList)
        except Exception as detail:
            if options.skipping:
                print("Skipping ", detail)
            continue
        branchList.append (branch)
    Branch.printList (branchList)
    rootfile.Close()
    del rootfile