def save(self, filename=None, verbose=False, **inkwargs):
''' Save the current canvas contents to [filename] '''
kwargs = {
'png': True,
'pdf': True,
'dotc': False,
'dotroot': False,
'json': False
}
kwargs.update(self.defaults['save'] if 'save' in self.defaults else {})
kwargs.update(inkwargs)
#
if filename is None:
filename = self.canvas.name
#self.pad.Draw()
self.canvas.Update()
if not os.path.exists(self.outputdir):
os.makedirs(self.outputdir)
if verbose:
print 'saving ' + os.path.join(self.outputdir, filename) + '.png'
if kwargs['png']:
self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.png')
if kwargs['pdf']:
self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.pdf')
if kwargs['dotc']:
self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.C')
if kwargs['json']:
jdict = {}
for obj in self.keep:
if isinstance(obj, ROOT.TH1):
jdict[obj.GetTitle()] = [
obj.GetBinContent(1),
obj.GetBinError(1)
]
if isinstance(obj, ROOT.THStack):
jdict['hist_stack'] = {}
for i in obj.GetHists():
jdict['hist_stack'][i.GetTitle()] = [
i.GetBinContent(1),
i.GetBinError(1)
]
with open(os.path.join(self.outputdir, filename) + '.json',
'w') as jout:
jout.write(prettyjson.dumps(jdict))
if kwargs['dotroot']:
logging.error('This functionality still has to be implemented '
'properly, due to the awsome ROOT "features"')
rfile = os.path.join(self.outputdir, filename) + '.root'
with io.root_open(rfile, 'recreate') as tfile:
self.canvas.Write()
for obj in self.keep:
if isinstance(obj, plotting.HistStack):
for hist in obj.hists:
hist.Write()
obj.Write()
self.reset()
def save(self, filename=None, verbose=False, **inkwargs):
''' Save the current canvas contents to [filename] '''
kwargs = {
'png' : True, 'pdf' : True, 'dotc' : False,
'dotroot' : False, 'json' : False
}
kwargs.update(
self.defaults['save'] if 'save' in self.defaults else {}
)
kwargs.update(
inkwargs
)
#
if filename is None:
filename = self.canvas.name
#self.pad.Draw()
self.canvas.Update()
if not os.path.exists(self.outputdir):
os.makedirs(self.outputdir)
if verbose:
print 'saving '+os.path.join(self.outputdir, filename) + '.png'
if kwargs['png']: self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.png')
if kwargs['pdf']: self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.pdf')
if kwargs['dotc']:
self.canvas.SaveAs(os.path.join(self.outputdir, filename) + '.C')
if kwargs['json']:
jdict = {}
for obj in self.keep:
if isinstance(obj, ROOT.TH1):
jdict[obj.GetTitle()] = [obj.GetBinContent(1), obj.GetBinError(1)]
if isinstance(obj, ROOT.THStack):
jdict['hist_stack'] = {}
for i in obj.GetHists():
jdict['hist_stack'][i.GetTitle()] = [i.GetBinContent(1), i.GetBinError(1)]
with open(os.path.join(self.outputdir, filename) + '.json', 'w') as jout:
jout.write(prettyjson.dumps(jdict))
if kwargs['dotroot']:
logging.error(
'This functionality still has to be implemented '
'properly, due to the awsome ROOT "features"')
rfile = os.path.join(self.outputdir, filename) + '.root'
with io.root_open(rfile, 'recreate') as tfile:
self.canvas.Write()
for obj in self.keep:
if isinstance(obj, plotting.HistStack):
for hist in obj.hists:
hist.Write()
obj.Write()
self.reset()
parser.add_argument(
'--category',
default='*',
help='category to be used for drawing correlation matrix (POSIX regex)')
parser.add_argument('--pickEvery',
type=int,
default=10,
help='pick one event every to draw the correlation matrix')
parser.add_argument('--batch', action='store_true', help='batch mode')
args = parser.parse_args()
args_dict = deepcopy(args.__dict__)
current_file = os.path.abspath(inspect.getfile(inspect.currentframe()))
watermark = TObjString(prettyjson.dumps(args_dict))
codeset = open(current_file).read(
) #zlib.compress(open(current_file).read()) compressing does not work well with root...
codemark = TObjString(codeset)
#
# CORRELATION MATRIX
#
scripts_dir = os.path.join(os.environ['CTRAIN'], 'scripts')
fname_regex = re.compile(
'[a-zA-Z_0-9]+_(?P<category>[a-zA-Z]+)_(?P<flavor>[A-Z]+)\.root')
qcd_txt_path = os.path.join(scripts_dir, 'data/flat_trees/qcd_flat.list')
input_files = [i.strip() for i in open(qcd_txt_path)]
if args.category != '*':
input_files = [
#splot, All_tp0_category_classifications[i,valid_disc0_indices], All_tp1_category_classifications[i,valid_disc1_indices] = plot_data(lda, vars_dist, class_type, class_type_pred, cat_type_list[i][0][3], cat_type_list[i][0][4], cat_type_list[i][1][4], fig_index=i+1)
#set_trace()
plot_lda_cov(lda, splot)
plt.axis('tight')
#plt.axis([0, 1000, -10, 30])
plt.ylabel(ytitle)
#plt.xlabel('$t_{h}$ $p_{T}$ [GeV]')
plt.xlabel('$t_{h}$ $p_{T}$ [GeV]')
#plt.show()
#category_classifications(tp0_category_classifications, tp1_category_classifications)
#category_classifications(All_tp0_category_classifications, All_tp1_category_classifications)
#print 'ttJetsM1000_THadPt_vs_%s_%s_%s_Comparison.png' % (discname, cat_type_list[0][0][3], cat_type_list[1][0][3])
#set_trace()
outname = 'ttJetsM1000_THadPt_vs_%s_%s_%s_Comparison_split' % (
discname, cat_type_list[0][0][3], cat_type_list[1][0][3])
out_dir = '/home/jdulemba/ROOT/ttbar_reco_3J/plots/Split'
fig.savefig('%s/%s.png' % (out_dir, outname))
#pt.print_table(rows, filename='%s.txt' % outname)
with open('%s/%s.json' % (out_dir, outname), 'w') as f:
f.write(prettyjson.dumps(json))
print ' %s/%s.png and\n %s/%s.json created.' % (out_dir, outname, out_dir,
outname)
#set_trace()
log.setLevel(rootpy.log.DEBUG)
input_files = [i.strip() for i in open('../data_trees/inputs/%s.list' % args.sample)]
fname_regex = re.compile('[a-zA-Z_0-9\/]*\/?[a-zA-Z_0-9]+_(?P<category>[a-zA-Z]+)_(?P<flavor>[A-Z]+)\.root')
yields = {}
for flavor in flavors:
yields[flavor] = {}
for category in sv_categories:
yields[flavor][category] = {}
for fname in input_files:
log.info('processing file %s' % fname)
with io.root_open(fname) as tfile:
match = fname_regex.match(fname)
if not match:
raise ValueError("Could not match the regex to the file %s" % fname)
flavor = match.group('flavor')
full_category = match.group('category')
category = [i for i in sv_categories if i in full_category][0]
tree = tfile.Get('ttree')
for cut in itercuts():
nentries = float(tree.GetEntries(cut))
if cut not in yields[flavor][category]:
yields[flavor][category][cut] = 0
yields[flavor][category][cut] += nentries
with open('../data_trees/%s_yields.json' % args.sample, 'w') as out:
out.write(prettyjson.dumps(yields))
import localsettings as site
import features
parser = ArgumentParser()
parser.add_argument('--flavour', default='C', help='the flavour fto be used for drawing the correlation matrix')
parser.add_argument('--category', default='*', help='category to be used for drawing correlation matrix (POSIX regex)')
parser.add_argument('--pickEvery', type=int, default=10, help='pick one event every to draw the correlation matrix')
parser.add_argument('--batch', action='store_true', help='batch mode')
args = parser.parse_args()
args_dict = deepcopy(args.__dict__)
current_file = os.path.abspath(inspect.getfile(inspect.currentframe()))
watermark = TObjString(prettyjson.dumps(args_dict))
codeset = open(current_file).read() #zlib.compress(open(current_file).read()) compressing does not work well with root...
codemark = TObjString(
codeset
)
#
# CORRELATION MATRIX
#
scripts_dir = os.path.join(os.environ['CTRAIN'], 'scripts')
fname_regex = re.compile('[a-zA-Z_0-9]+_(?P<category>[a-zA-Z]+)_(?P<flavor>[A-Z]+)\.root')
qcd_txt_path= os.path.join(scripts_dir, 'data/flat_trees/qcd_flat.list')
if args.input:
jconf = prettyjson.loads(open(args.input[-1]).read())
file_names = [tuple(i) for i in jconf['file_names']]
output = jconf['output']
graph_path = jconf['graph_path']
#dump configuration in json for bookkeeping
jconf = {
'file_names' : file_names,
'output' : output,
'graph_path' : graph_path,
}
jout = output.split('.')[0]
with open('%s.json' % jout, 'w') as out:
out.write(prettyjson.dumps(jconf))
canvas = plt.Canvas(800, 800)
canvas.SetLogy();
canvas.SetGridx();
canvas.SetGridy();
max_txt_len = max(len(i) for _, i in file_names)
legend = plt.Legend(
len(file_names),
leftmargin=0.18+(30-max_txt_len)*0.016,
rightmargin=0.005,
topmargin=0.60-0.057*(len(file_names)-3),
entrysep=0,
margin=0.1+0.006*(30-max_txt_len)
)
for imin, imax in zip(mins, maxes):
offsets.append(-1*imin)
delta = imax-imin
scales.append(100./delta if delta else 1.)
for row in X:
for idx in range(len(row)):
row[idx] = scales[idx]*(row[idx]+offsets[idx])
if row[idx] < 0: set_trace()
log.info('... ranking ...')
chi2_ranking = SelectKBest(chi2, 'all')
chi2_ranking.fit(X, y)
log.info('... done')
anova_scores = zip(variables, anova_ranking.scores_)
anova_scores.sort(key=lambda x: x[1], reverse=True)
chi2_scores = zip(variables, chi2_ranking.scores_)
chi2_scores.sort(key=lambda x: x[1], reverse=True)
scores = [
{'method' : 'ANOVA', 'scores' : anova_scores},
{'method' : 'chi2', 'scores' : anova_scores}
]
if '.json' in args.out:
with open(args.out, 'w') as out:
out.write(prettyjson.dumps(scores))
else:
pickle.dump(scores, open(args.out, "wb"))
log.setLevel(rootpy.log.DEBUG)
input_files = [i.strip() for i in open('data/inputs/%s.list' % args.sample)]
fname_regex = re.compile('[a-zA-Z_0-9\/]*\/?[a-zA-Z_0-9]+_(?P<category>[a-zA-Z]+)_(?P<flavor>[A-Z]+)\.root')
yields = {}
for flavor in flavors:
yields[flavor] = {}
for category in sv_categories:
yields[flavor][category] = {}
for fname in input_files:
log.info('processing file %s' % fname)
with io.root_open(fname) as tfile:
match = fname_regex.match(fname)
if not match:
raise ValueError("Could not match the regex to the file %s" % fname)
flavor = match.group('flavor')
full_category = match.group('category')
category = [i for i in sv_categories if i in full_category][0]
tree = tfile.Get(full_category)
for cut in itercuts():
nentries = float(tree.GetEntries(cut))
if cut not in yields[flavor][category]:
yields[flavor][category][cut] = 0
yields[flavor][category][cut] += nentries
with open('data/%s_yields.json' % args.sample, 'w') as out:
out.write(prettyjson.dumps(yields))
## many, many checks!
for flavour in flavours:
assert(is_flat(bin_yields(final_qcd[flavour])))
for bin in bins:
bias_qcds = [final_qcd[flavour][i][bin] for i in categories]
ttjs = [ttj_yields[flavour][i][bin] for i in categories]
assert(ratios_are_ok(bias_qcds, ttjs))
##
## STORE IN H2D and jsons
##
with open('data/qcd_bin_weights.json', 'w') as out:
out.write(
prettyjson.dumps(bin_weight)
)
with io.root_open('data/qcd_weights.root', 'w') as out:
fweights = ROOT.TObjString(prettyjson.dumps(flav_weights))
out.WriteTObject(fweights, 'flavour_weights')
for flavour in flavours:
fdir = out.mkdir(flavour)
for category in categories:
cdir = fdir.mkdir(category)
h2d_category = Hist2D(binning.pt_bins, binning.eta_bins) #category bias weights
h2d_bin = Hist2D(binning.pt_bins, binning.eta_bins) #pt/eta bin weights
for cat_bin, bin_bin in zip(h2d_category, h2d_bin):
cat_cut = binning.cut_from_bin(cat_bin)
bin_cut = binning.cut_from_bin(bin_bin)
assert(cat_cut == bin_cut)
scales.append(100. / delta if delta else 1.)
for row in X:
for idx in range(len(row)):
row[idx] = scales[idx] * (row[idx] + offsets[idx])
if row[idx] < 0: set_trace()
log.info('... ranking ...')
chi2_ranking = SelectKBest(chi2, 'all')
chi2_ranking.fit(X, y)
log.info('... done')
anova_scores = zip(variables, anova_ranking.scores_)
anova_scores.sort(key=lambda x: x[1], reverse=True)
chi2_scores = zip(variables, chi2_ranking.scores_)
chi2_scores.sort(key=lambda x: x[1], reverse=True)
scores = [{
'method': 'ANOVA',
'scores': anova_scores
}, {
'method': 'chi2',
'scores': anova_scores
}]
if '.json' in args.out:
with open(args.out, 'w') as out:
out.write(prettyjson.dumps(scores))
else:
pickle.dump(scores, open(args.out, "wb"))
if args.input:
jconf = prettyjson.loads(open(args.input[-1]).read())
file_names = [tuple(i) for i in jconf['file_names']]
output = jconf['output']
graph_path = jconf['graph_path']
#dump configuration in json for bookkeeping
jconf = {
'file_names': file_names,
'output': output,
'graph_path': graph_path,
}
jout = output.split('.')[0]
with open('%s.json' % jout, 'w') as out:
out.write(prettyjson.dumps(jconf))
canvas = plt.Canvas(800, 800)
canvas.SetLogy()
canvas.SetGridx()
canvas.SetGridy()
max_txt_len = max(len(i) for _, i in file_names)
legend = plt.Legend(len(file_names),
leftmargin=0.18 + (30 - max_txt_len) * 0.016,
rightmargin=0.005,
topmargin=0.60 - 0.057 * (len(file_names) - 3),
entrysep=0,
margin=0.1 + 0.006 * (30 - max_txt_len))
legend.SetTextSize(legend.GetTextSize() * 0.8)
legend.SetFillColor(0)
