name,

bin_width=5.0,

max_bin=2000,

):

gROOT.cd() # deal with TH1D/TFile/python scope issues!!

n_bins = int(1.0*max_bin/bin_width)

hist = TH1D('%s' % name, '', n_bins, 0, max_bin)

hist.Sumw2()

#print '----> making hist: %s' % hist.GetName()

weight_to_hist_dict = {}

weight_to_n_events_dict = {}

total_hist = get_hist('total')

n_all_events = 0

total_track_weight_hist = TH1D('track_weight_hist', '', 110, -100, 10)

total_track_weight_hist.SetLineColor(TColor.kBlue+1)

total_track_weight_hist.SetFillColor(TColor.kBlue+1)

n_IS_entries = 0

n_noIS_entries = 0

n_files = 0

for root_file_name in root_file_names:

n_files += 1

#if n_files > 4: break # debugging

basename = os.path.basename(root_file_name)

print '--> processing %s' % basename

root_file = TFile(root_file_name)

tree = root_file.Get('fTree')

#n_entries = tree.GetEntries()

n_entries = tree.Draw('fTotalEnergy', 'fTotalEnergy>0', 'goff')

if n_entries <= 0:

continue

# get some info from the first tree entry

tree.GetEntry(0)

mc_run = tree.fMCRun

n_events = mc_run.GetNEvents()

is_used = mc_run.GetUseImportanceSampling()

n_all_events += n_events

if is_used:

track_weight_hist = TH1D('track_weight_hist', '', 110, -100, 10)

tree.Draw(

'TMath::Log2(fSteps.fTrackWeight) >> track_weight_hist',

'fEdep>0',

'goff'

)

total_track_weight_hist.Add(track_weight_hist)

min_weight = 1.0

max_weight = -400

for i_bin in range(track_weight_hist.GetNbinsX()):

low_edge = track_weight_hist.GetBinLowEdge(i_bin)

counts = track_weight_hist.GetBinContent(i_bin)

if counts > 0:

#print i_bin, low_edge, counts

if low_edge > max_weight:

max_weight = low_edge

if low_edge < min_weight:

min_weight = low_edge

#print min_weight, max_weight

# these were all log2 values:

#min_weight = pow(2.0, min_weight)

#max_weight = pow(2.0, max_weight)

weight = max_weight # test!

else:

weight = 0.0

print '\t %i events | %i entries | weight: %s | eff: %.1e +/- %.1e' % (

n_events,

n_entries,

weight,

weight*n_entries/n_events,

weight*math.sqrt(n_entries)/n_events,

)

if is_used:

print '\t\t weight range: %s, %s' % (min_weight, max_weight)

n_IS_entries += n_entries

else:

n_noIS_entries += n_entries

try:

hist = weight_to_hist_dict[weight]

n_total_events = weight_to_n_events_dict[weight]

#print '--> found hist: %s' % hist.GetName()

except KeyError:

hist = get_hist(name = 'hist_weight%.3e' % weight)

weight_to_hist_dict[weight] = hist

hist = weight_to_hist_dict[weight]

n_total_events = 0

n_total_events += n_events

weight_to_n_events_dict[weight] = n_total_events

#print hist.GetEntries()

hist.GetDirectory().cd()

# this draws edep from each step

#tree.Draw(

# 'fEdep*1e3 >> +%s' % hist.GetName(),

# 'fTrackWeight*(fEdep>0)',

# 'goff'

#)

# this draws total edep -- assuming fTrackWeight taken from the last step

# in the first event applied to all events!

selection = '(fTotalEnergy>0)'

if is_used:

selection = 'fTrackWeight[1]*(fTotalEnergy>0)'

#print selection

n_drawn = tree.Draw(

'fTotalEnergy*1e3 >> +%s' % hist.GetName(),

selection,

'goff'

)

tree.Draw(

'fTotalEnergy*1e3 >> +%s' % total_hist.GetName(),

selection,

'goff'

)

#print hist.GetEntries(), n_drawn, total_hist.GetEntries()

print '\t IS used:', is_used, mc_run.GetBiasedParticleID(), mc_run.GetUseTimeWindow(), mc_run.GetUseImportanceProcessWindow()

#print hist.GetEntries()

# end loop over input files

weight_to_n_events_dict[4.0] = n_all_events

weight_to_hist_dict[4.0] = total_hist

weights = weight_to_hist_dict.keys()

weights.sort()

canvas = TCanvas('canvas', '')

canvas.SetLogy(1)

legend = TLegend(0.1, 0.91, 0.9, 0.99)

legend.SetNColumns(4)

# find the max

max_y_value = 0

min_max_y_value = weight_to_hist_dict.values()[0].GetMaximum()

min_y_value = weight_to_hist_dict.values()[0].GetMinimum()

hists = weight_to_hist_dict.values()

for hist in hists:

n_total_events = weight_to_n_events_dict[weight]

hist.Scale(1.0/n_total_events)

hist_max = hist.GetMaximum()

hist_min = hist.GetMinimum()

if hist_max > max_y_value: max_y_value = hist_max

if hist_min < min_y_value: min_y_value = hist_min

if hist_max < min_max_y_value: min_max_y_value = hist_max

for i_weight in range(len(weights)):

weight = weights[i_weight]

hist = weight_to_hist_dict[weight]

hist.SetLineWidth(2)

color = i_weight+2

hist.SetFillColor(color)

hist.SetLineColor(color)

#hist.SetMarkerColor(color)

n_entries = hist.GetEntries()

n_hits_per_decay = hist.Integral(0, hist.GetNbinsX())

try:

n_hits_err = n_hits_per_decay/math.sqrt(n_entries)

except:

n_hits_err = 0.0

print 'hist %i | weight: 2^(%i) = %.2e | hits: %i | eff: %.2e +/- %.2e' % (

i_weight,

weight,

pow(2.0, weight),

n_entries,

n_hits_per_decay,

n_hits_err,

)

draw_opt = 'e2'

if i_weight is 0:

hist.SetMaximum(max_y_value*2.0)

#hist.SetMinimum(min_y_value/2.0)

hist.SetMinimum(min_max_y_value/10.0)

hist.SetXTitle('Energy [keV]')

hist.SetYTitle('Counts / Decay / %.1f keV' % hist.GetBinWidth(1))

hist.GetYaxis().SetTitleOffset(1.2)

else:

draw_opt += ' same'

hist.Draw(draw_opt)

entry_label = '%s, %.1e entries' % (weight, n_entries)

legend.AddEntry(hist, entry_label, 'lf')

legend.Draw()

canvas.Update()

prefix = os.path.commonprefix(root_file_names)

prefix = os.path.basename(prefix)

#canvas.Print('%s_IS_Spectra.pdf' % prefix)

print '--- %i entries with IS, %i entries w/o IS' % (

n_IS_entries,

n_noIS_entries,

)

response = raw_input('--> enter to continue, i to see indiv hists, q to quit: ')

if response == 'q':

return

if response == 'i':

for i_weight in range(len(weights)):

weight = weights[i_weight]

hist = weight_to_hist_dict[weight]

print weight

hist.Draw('e2')

canvas.Update()

raw_input('--> return to continue ')

legend = TLegend(0.1, 0.91, 0.9, 0.99)

legend.SetNColumns(4)

print '--> drawing track weight distribution:'

total_track_weight_hist.Draw()

canvas.Update()

raw_input('--> return to continue ')

# now look at the output of each run:

hists = []

hist_min = 1e5

hist_max = 0.0

for i_file in range(len(root_file_names)):

root_file_name = root_file_names[i_file]

root_file = TFile(root_file_name)

tree = root_file.Get('fTree')

# get some info from the first tree entry

tree.GetEntry(0)

mc_run = tree.fMCRun

n_events = mc_run.GetNEvents()

is_used = mc_run.GetUseImportanceSampling()

run_id = mc_run.GetRunID()

i_color = i_file +2

hist = get_hist(name=run_id, bin_width=75.0)

hist.SetLineColor(i_color)

hist.SetFillColor(i_color)

hist.SetLineWidth(3)

entry_label = '%s' % run_id

legend.AddEntry(hist, entry_label, 'lf')

selection = '(fTotalEnergy>0)'

if is_used:

selection = 'fTrackWeight[1]*(fTotalEnergy>0)'

#print selection

n_drawn = tree.Draw(

'fTotalEnergy*1e3 >> +%s' % hist.GetName(),

selection,

'goff'

)

hist.Scale(1.0/n_events)

hists.append(hist)

print run_id, i_color

max = hist.GetMaximum()

min = hist.GetMinimum()

if min < hist_min: hist_min = min

if max > hist_max: hist_max = max

#if i_file > 5: break # debugging

# end loop over files

hists[0].SetMaximum(hist_max*2.0)

hists[0].SetMinimum(hist_max/1e7)

hists[0].Draw('e2')

for hist in hists:

hist.Draw('e2 same')

legend.Draw()

canvas.Update()