Created
February 26, 2015 19:12
-
-
Save annawoodard/cc23e9bb05b350d3d56f to your computer and use it in GitHub Desktop.
LHE file merger
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/perl -w | |
################################################################################ | |
# merge.pl | |
# Richard Corke ([email protected]) | |
# | |
# Changed structure to not read entire files into memory | |
# | |
# Based on merge.pl v1.2 by Michel Herquet (UCL-CP3) | |
# DESCRIPTION : script to merge to LHE events files | |
# USAGE : | |
# ./merge.pl eventfile1.lhe.gz eventfile2.lhe.gz ... result.lhe.gz banner.txt | |
# Note that result can be eventfile1, eventfile2, ... | |
# OUTPUT : merged file, banner | |
################################################################################ | |
use Compress::Zlib; | |
use List::Util qw[min max]; | |
########################################################################### | |
# Initialisation | |
########################################################################### | |
# Set tag names for later | |
my $begin_header = '<header>'; | |
my $end_header = '</header>'; | |
my $begin_event = '<event>'; | |
my $end_event = '</event>'; | |
my $begin_init = '<init>'; | |
my $end_init = '</init>'; | |
# Parse the command line arguments | |
if ( $#ARGV < 2 ) { | |
die "This script must be called with at least three filenames as arguments!\n"; | |
} | |
my $bannerfile = pop(@ARGV); | |
my $outfile = pop(@ARGV); | |
########################################################################### | |
# First pass - extract number of events and cross-sections | |
########################################################################### | |
foreach $infile (@ARGV) { | |
$gzin = gzopen($infile, "r") || die ("Couldn't open file $infile\n"); | |
# No. events and cross-section from current file | |
$noevents = $xsec = -1; | |
# Storage for current file's init block | |
@gzinit = (); | |
# Keep track if we are in the init block or not | |
$initblock = 0; | |
while (1) { | |
$gzbytes = $gzin->gzreadline($gzline); | |
if ($gzbytes == -1) { die("Error reading from file $infile\n"); } | |
if ($gzbytes == 0) { last; } | |
# Extract <MGGenerationInfo> information | |
if ($initblock == 0) { | |
if ($gzline =~ s/# Number of Events\s*:\s*(.*)\n/$1/) { $noevents = $gzline; } | |
if ($gzline =~ s/# Integrated weight \(pb\)\s*:\s*(.*)\n/$1/) { $xsec = $gzline; } | |
# Check if we enter <init> block | |
if ($gzline =~ m/$begin_init/) { $initblock++; next; } | |
# Extract <init> information | |
} else { | |
# Check for end of <init> block | |
if ($gzline =~ m/$end_init/) { last; } | |
# Remove leading whitespace and split | |
$gzline =~ s/^\s+//; | |
@gzparam = split(/\s+/, $gzline); | |
# Check <init> block line has right no. of entries | |
if ($initblock == 1) { | |
if ($#gzparam != 9) { die "Not right number of params in init"; } | |
} else { | |
if ($#gzparam != 3) { die "Not right number of params in init"; } | |
} | |
push(@gzinit, [ @gzparam ]); | |
$initblock++; | |
} | |
} | |
$gzin->gzclose(); | |
# Check the file contained all the information we need | |
if ($noevents == -1 || $xsec == -1) { | |
die("Couldn't extract No. Events / Cross-section from $infile") | |
} | |
# Store information for later | |
push(@infiles, [ $infile, $noevents, $xsec, [ @gzinit ] ]); | |
} | |
########################################################################### | |
# Check/compute cross-sections, totals and unit weight | |
########################################################################### | |
$oldxsec = $infiles[0][2]; | |
@oldinit = @{$infiles[0][3]}; | |
$totevents = 0; $totxsec = 0.0; | |
foreach $infile (@infiles) { | |
print "Input file: $infile->[0]\n"; | |
print " No. Events = $infile->[1], Cross-section = $infile->[2]\n"; | |
# Check that cross sections do not differ too much | |
$newxsec = $infile->[2]; | |
if (abs(($newxsec - $oldxsec) / $newxsec) > 0.05 ) { | |
print " WARNING Cross sections do not agree with a 5\% precision!\n"; | |
} | |
$oldxsec = $newxsec; | |
@currinit = @{$infile->[3]}; | |
# Same number of lines in <init> block? | |
if ($#oldinit != $#currinit) { die("Init blocks do not match"); } | |
# Same number of entries on first line of <init> block? | |
if ($#{$oldinit[0]} != $#{$currinit[0]}) { die("Init blocks do not match"); } | |
# All entries the same on first line of <init> block? | |
for ($i = 0; $i <= $#{$oldinit[0]}; $i++) { | |
if ($oldinit[0][$i] != $currinit[0][$i]) | |
{ die("Init blocks do not match"); } | |
} | |
# Create new init block (overwrite first file's init block data) | |
for ($i = 1; $i <= $#oldinit; $i++) { | |
if ($oldinit[$i][3] != $currinit[$i][3]) { die("Init blocks do not match"); } | |
print " xsecup = $currinit[$i][0], xerrup = $currinit[$i][1]\n"; | |
print " xmaxup = $currinit[$i][2], lprup = $currinit[$i][3]\n"; | |
# XSECUP = sum(xsecup * no.events) / tot.events | |
# XERRUP = sqrt( sum(sigma^2 * no.events^2) ) / tot.events | |
# Here we temporarily store: | |
# sum(xsecup * no.events) | |
# sum(sigma^2 * no.events^2) | |
if (\$oldinit == \$currinit) { | |
$oldinit[$i][0] *= $infile->[1]; | |
$oldinit[$i][1] *= $oldinit[$i][1] * $infile->[1]**2; | |
} else { | |
$oldinit[$i][0] += ($currinit[$i][0] * $infile->[1]); | |
$oldinit[$i][1] += $currinit[$i][1]**2 * $infile->[1]**2; | |
} | |
# XMAXUP = max(xmaxup) | |
$oldinit[$i][2] = max($oldinit[$i][2], $currinit[$i][2]); | |
} | |
# Total number of events and total cross-section | |
$totevents += $infile->[1]; | |
$totxsec += ($infile->[1] * $infile->[2]); | |
print "\n"; | |
} | |
print "\n"; | |
# Finish calculation of XSECUP and XERRUP | |
for ($i = 1; $i <= $#oldinit; $i++) { | |
$oldinit[$i][0] /= $totevents; | |
$oldinit[$i][0] = sprintf('%0.5E', $oldinit[$i][0]); | |
$oldinit[$i][1] = sqrt($oldinit[$i][1]); | |
$oldinit[$i][1] /= $totevents; | |
$oldinit[$i][1] = sprintf('%0.5E', $oldinit[$i][1]); | |
} | |
# Finish calculation of total xsec and new unit weight | |
$totxsec /= $totevents; | |
$dispxsec = sprintf('%0.5E', $totxsec); | |
$uwgt = sprintf('%0.5E', $totxsec / $totevents); | |
# Display new information | |
print "Banner file: $bannerfile\n"; | |
print "Output file: $outfile\n"; | |
print " No. Events = $totevents, Cross-section = $dispxsec\n"; | |
for ($i = 1; $i <= $#oldinit; $i++) { | |
print " xsecup = $oldinit[$i][0], xerrup = $oldinit[$i][1]\n"; | |
print " xmaxup = $oldinit[$i][2], lprup = $oldinit[$i][3]\n"; | |
} | |
print "\n"; | |
print " Unit weight = $uwgt\n"; | |
########################################################################### | |
# Second pass - output file with new information | |
########################################################################### | |
# The first file is written out with changes to the header, init block | |
# and unit weight of each event. | |
# All events from other files are then written out with their unit weight | |
# changed. | |
########################################################################### | |
$gzout = gzopen($outfile, "w") || die ("Couldn't open file $outfile\n"); | |
open(BANNER, ">$bannerfile") || die ("Couldn't open file $outfile\n"); | |
$stage = 0; | |
$eventcount = 0; | |
foreach $infile (@infiles) { | |
$gzin = gzopen($infile->[0], "r") || die ("Couldn't open file $infile\n"); | |
while (1) { | |
$gzbytes = $gzin->gzreadline($gzline); | |
if ($gzbytes == -1) { die("Error reading from file $infile\n"); } | |
if ($gzbytes == 0) { last; } | |
# Pre-header | |
if ($stage == 0) { | |
if ($gzline =~ m/$begin_header/) { $stage++; } | |
# Header (and output to banner) | |
} elsif ($stage == 1) { | |
$gzline =~ s/# Integrated weight \(pb\)\s*:(.*)\n/# Integrated weight (pb) : $dispxsec\n/; | |
$gzline =~ s/# Number of Events\s*:(.*)\n/# Number of Events : $totevents\n/; | |
$gzline =~ s/# Unit wgt\s*:(.*)\n/# Unit wgt : $uwgt\n/; | |
if ($gzline =~ m/$end_header/) { $stage++; } | |
else { print BANNER $gzline; } | |
# Init block | |
} elsif ($stage == 2) { | |
if ($gzline =~ m/$end_init/) { | |
$gzline = "<init>\n"; | |
for ($i = 0; $i <= $#oldinit; $i++) { | |
$gzline .= " "; | |
for ($j = 0; $j <= $#{$oldinit[$i]}; $j++) { | |
$gzline .= "$oldinit[$i][$j] "; | |
} | |
$gzline .= "\n"; | |
} | |
$gzline .= "</init>\n"; | |
$stage++; | |
} else { next; } | |
# Pre-event | |
} elsif ($stage == 3) { | |
if ($gzline =~ m/$begin_event/) { $stage++; } else { next; } | |
# Event information | |
} elsif ($stage == 4) { | |
$gzline =~ s/^\s+//; | |
@gzparam = split(/\s+/, $gzline); | |
# Keep weight sign from original LHE file | |
$signed_uwgt = $uwgt; | |
if ($gzparam[2] < 0) { | |
$signed_uwgt = -1 * $signed_uwgt; | |
} | |
$gzline = " $gzparam[0] $gzparam[1] $signed_uwgt $gzparam[3] $gzparam[4] $gzparam[5]\n"; | |
if ($#gzparam != 5) { die "Not right number of param in first line of event"; } | |
$gzline = " $gzparam[0] $gzparam[1] $uwgt $gzparam[3] $gzparam[4] $gzparam[5]\n"; | |
$stage++; | |
# Event particles | |
} elsif ($stage == 5) { | |
if ($gzline =~ m/$end_event/) { $stage = 3; $eventcount++ } | |
} | |
# Write out the line | |
$gzout->gzwrite($gzline); | |
} | |
} | |
# Write out closing tag and close | |
$gzout->gzwrite("</LesHouchesEvents>\n"); | |
$gzout->gzclose(); | |
print "Wrote $eventcount events\n"; | |
exit(0); |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment