def main():
client = MongoClient('mongodb://localhost:27017/')
db = client.cvedb
# Get the collection object
# Here name of the database is "states"
collection = db.cves
# Make a query to list all the documents
cvedicttemp = collection.find()
cvedict = dict()
for key in cvedicttemp:
cvedict[key['id']] = key
vlist = []
carlosplt.pre_paper_plot()
fig = plt.figure()
tester = calc_laplace()
ax = fig.add_subplot(2, 2, 1)
tester.laplace_php([])
ax = fig.add_subplot(2, 2, 2)
tester.laplace_openjdk([])
ax = fig.add_subplot(2, 2, 3)
tester.laplace_wheezy([], False)
ax = fig.add_subplot(2, 2, 4)
tester.laplace_wheezy([], True)
carlosplt.post_paper_plot(True, True, True)
plt.show()
def plot_average_number(self):
average_per_year = [0] * self.plotter.years
for j in range(self.plotter.years):
average_per_year[j] = self.plotter.year_sum[j] / float(
self.plotter.year_num[j])
x_values = list(range(1, self.plotter.years + 1))
slope = np.polyfit(x_values, average_per_year, 1)
print('Slope: ' + str(slope))
n = len(self.plotter.year_sum)
x = range(self.plotter.years)
width = 1 / 2
#plt.bar(x, year_sum, width)
plt.bar(x,
average_per_year,
width,
color='darkblue',
edgecolor='black')
plt.xticks(np.arange(0, n), self.yearsx)
plt.ylabel('Average vulnerabilities per package')
plt.xlabel('Year')
carlosplt.post_paper_plot(True, True, True)
## Linear regression
X = sm.add_constant(x)
y = average_per_year
model = sm.OLS(y, X).fit()
predictions = model.predict(X)
plt.plot(predictions)
plt.show()
print(model.summary())
print(model.summary().as_latex())
def plot_severity_percentage(self):
num_low = [0] * (self.years + 1)
num_med = [0] * (self.years + 1)
num_high = [0] * (self.years + 1)
num_udef = [0] * (self.years + 1)
total = [0] * (self.years + 1)
for i in self.pkg_with_cvss:
for j in range(len(self.src2month[i])):
try:
num_low[j // 12] += self.pkg_with_cvss[i][j][0]
num_med[j // 12] += self.pkg_with_cvss[i][j][1]
num_high[j // 12] += self.pkg_with_cvss[i][j][2]
num_udef[j // 12] += self.pkg_with_cvss[i][j][3]
total[j // 12] += self.pkg_with_cvss[i][j][
3] + self.pkg_with_cvss[i][j][2] + self.pkg_with_cvss[
i][j][1] + self.pkg_with_cvss[i][j][0]
except IndexError:
xx = j // 12
if xx == 19:
continue
else:
print(xx)
#raise IndexError('List index out of bounds')
## Generate percentage
for i in range(self.years + 1):
try:
num_low[i] = num_low[i] / total[i]
num_med[i] = num_med[i] / total[i]
num_high[i] = num_high[i] / total[i]
num_udef[i] = num_udef[i] / total[i]
except ZeroDivisionError:
num_low[i] = 0
num_med[i] = 0
num_high[i] = 0
num_udef[i] = 0
print(num_low)
print(num_high)
carlosplt.pre_paper_plot()
pal = ['#fee8c8', '#fdbb84', '#e34a33', 'grey']
x = range(2001, 2001 + self.years)
labels_cvss = ['low', 'medium', 'high', 'N/A']
h = plt.stackplot(
x, [num_low[1:], num_med[1:], num_high[1:], num_udef[1:]],
colors=pal,
alpha=0.9,
labels=labels_cvss)
plt.xticks(x)
plt.legend(loc='upper left', handles=h[::-1])
carlosplt.post_paper_plot(True, True, True)
plt.show()
def test_slocs(src2month, src2sloccount):
# Remember sloccount is of the form (total, [ansic, cpp, asm, java, python, perl, sh])
ar1 = []
ar2 = []
print(sum(src2month['linux']))
for pkg in src2month:
try:
total_slocs = src2sloccount[pkg][0]
if total_slocs == 0:
continue
else:
ar1.append(int(total_slocs))
except KeyError:
print(pkg + ": no sloccount data found!")
continue
total = sum(src2month[pkg])
if total > 100:
print(pkg + ', ' + str(total) + ', ' + str(total_slocs))
ar2.append(total)
vulns_sorted_slocs_total = [
x for _, x in sorted(zip(ar1, ar2), reverse=True)
]
pop_xaxis = [y for y, _ in sorted(zip(ar1, ar2), reverse=True)]
half_more_slocs = sum(
vulns_sorted_slocs_total[:int(len(vulns_sorted_slocs_total) / 2)])
half_less_slocs = sum(
vulns_sorted_slocs_total[int(len(vulns_sorted_slocs_total) / 2):])
print(half_more_slocs)
print(half_less_slocs)
print(pop_xaxis[0])
print(pop_xaxis[len(pop_xaxis) - 1])
print(spearmanr(ar1, ar2))
carlosplt.pre_paper_plot(True)
plt.plot(vulns_sorted_slocs_total)
plt.ylabel('Number of vulnerabilities')
plt.xlabel('Number of SLOCS ranking')
carlosplt.post_paper_plot(True, True, True)
plt.show()
def plot_num_affected(self):
## Number of affected packages
n = len(self.plotter.year_sum)
for i in range(1, self.plotter.years + 1):
if i % 2 == 1:
self.yearsx.append('\'' + str(i).zfill(2))
else:
self.yearsx.append('')
x = range(self.plotter.years)
width = 1 / 2
plt.bar(x,
self.plotter.year_num,
width,
color='darkblue',
edgecolor='black')
plt.xticks(np.arange(0, n), self.yearsx)
plt.ylabel('Number of affected packages')
plt.xlabel('Year')
carlosplt.post_paper_plot(True, True, True)
plt.show()
def laplace_process_list(self, vlist, pkg, year):
months = len(vlist)
print(pkg)
perhour = 24 * 30 * months
instances = []
laplace_values = []
i = 0
print(pkg)
print(vlist)
for month in vlist:
i += 1
temp = random.sample(range(24 * 30 * (i - 1), 24 * 30 * i), month)
instances += temp
laplace_values.append(lp.laplace_test(instances, 24 * 30 * i))
final_laplace = lp.laplace_test(instances, 24 * 30 * i + 1)
print(final_laplace)
n = len(vlist)
if pkg == 'wheezy':
x = range(n + 12)
else:
x = range(n)
print(n)
yearsx = [
'\'' + str(year + 2000 + i)[-2:]
for i in range(len(vlist) // 12 + 1)
]
carlosplt.post_paper_plot(True, True, True)
#print(x)
if pkg == 'wheezy':
plt.plot(x, [None] * 4 + laplace_values + [None] * 8) #
else:
plt.plot(x, laplace_values)
plt.axhline(y=1.96, linestyle=':', color='orange')
plt.axhline(y=2.33, linestyle=':', color='red')
plt.axhline(y=-1.96, linestyle=':', color='orange')
plt.axhline(y=-2.33, linestyle=':', color='red')
plt.xticks(np.arange(0, n, step=12), yearsx)
plt.ylabel(pkg)
def test_pop(src2month, src2pop):
ar1 = []
ar2 = []
for pkg in src2month:
try:
ar1.append(int(src2pop[pkg]))
except KeyError:
#print(pkg + ": no popularity data found!")
continue
total = sum(src2month[pkg])
if total > 100:
print(pkg + ', ' + str(total) + ', ' + src2pop[pkg])
ar2.append(total)
vulns_sorted_pop = [x for _, x in sorted(zip(ar1, ar2), reverse=True)]
pop_xaxis = [y for y, _ in sorted(zip(ar1, ar2), reverse=True)]
half_more_popular = sum(vulns_sorted_pop[:int(len(vulns_sorted_pop) / 2)])
half_less_popular = sum(vulns_sorted_pop[int(len(vulns_sorted_pop) / 2):])
print(half_more_popular)
print(half_less_popular)
print(pop_xaxis[0])
print(pop_xaxis[len(pop_xaxis) - 1])
print(src2pop)
print(spearmanr(ar1, ar2))
carlosplt.pre_paper_plot(True)
plt.plot(vulns_sorted_pop)
plt.ylabel('Number of vulnerabilities')
plt.xlabel('Popularity ranking')
carlosplt.post_paper_plot(True, True, True)
plt.show()
def plot_wheezy_lts(self):
quarter_num = self.plotter.years * 4
# Get LTS and plot
try:
with open("DLA_sum.txt", "rb") as fp:
ltslist = pickle.load(fp)
with open("src2month_DLA.txt", "rb") as fp:
src2monthDLAs = pickle.load(fp)
with open("DLA_src2month.json", "r") as fp:
src2monthDLA = json.load(fp)
with open("DLA_withcvss.json", "r") as fp:
self.src2monthDLA_cvss = json.load(fp)
# Fix this so it can compute when required
#dla.permonthDLA(src2monthDLAs)
with open("1000.csv", "r") as csvfile:
spamreader = csv.reader(csvfile, delimiter=' ', quotechar='|')
except IOError:
ltslist = dla.getDLAs()
with open("src2month_DLA.txt", "rb") as fp:
src2monthDLAs = pickle.load(fp)
dla.permonthDLA(src2monthDLAs)
return self.plot_wheezy_lts()
## Plot for wheezy
quarter_sum = [0] * quarter_num
DLA_temp = dict()
## Fix src2monthDLA_cvss
for i in self.src2monthDLA_cvss:
temp_list = []
for j in self.src2monthDLA_cvss[i]:
temp_list += j
self.src2monthDLA_cvss[i] = temp_list
## Fix ltslist according to severity
for i in self.src2monthDLA_cvss:
DLA_temp[i] = []
for j in range(len(self.src2monthDLA_cvss[i])):
num_low = self.src2monthDLA_cvss[i][j][0]
num_med = self.src2monthDLA_cvss[i][j][1]
num_high = self.src2monthDLA_cvss[i][j][2]
num_udef = self.src2monthDLA_cvss[i][j][3]
tempp = 0
if self.l:
tempp += num_low
if self.m:
tempp += num_med
if self.h:
tempp += num_high
if self.udef:
tempp += num_udef
DLA_temp[i].append(tempp)
ltslist = []
for m in range((self.plotter.years + 1) * 12):
s = 0
#print(m)
for i in DLA_temp:
s += DLA_temp[i][m]
ltslist.append(s)
totalLTS = ltslist
plt.bar([i for i in range(len(ltslist))], ltslist)
plt.show()
quartersx = []
for i in range(1, self.plotter.years + 1):
for j in range(1, 5):
if j == 1:
quartersx.append('Q' + str(j) + '\'' + str(i).zfill(2))
else:
quartersx.append(' ')
for pkg in self.plotter.src2month_loc:
for j in range(quarter_num):
temp = sum(self.plotter.src2month_loc[pkg][12 + (3 * j):12 +
3 * (j + 1)])
quarter_sum[j] += temp
LTS_quarter = []
for j in range(quarter_num):
temp = sum(totalLTS[12 + (3 * j):12 + 3 * (j + 1)])
LTS_quarter.append(temp)
## Print all LTS
cut = 12 * 4 + 1
n = len(quarter_sum)
x = range(quarter_num)
width = 1 / 2
plt.bar(x,
LTS_quarter,
width,
color='brown',
label='regular support',
edgecolor='black')
plt.xticks(np.arange(0, n), quartersx, rotation="vertical")
plt.ylabel('Vulnerabilities per quarter of Debian LTS')
plt.xlabel('Quarter')
carlosplt.post_paper_plot(True, True, True)
plt.show()
## Filter only wheezy:
quarter_sum_regular = [0] * (12 * 4 + 1) + quarter_sum[12 * 4 +
1:12 * 4 +
9] + [0] * 12
quarter_sum_errors = [0] * (12 * 4 + 9) + quarter_sum[12 * 4 +
9:12 * 4 + 9 +
5] + [0] * 7
LTS_quarter = [0] * (15 * 4 + 2) + LTS_quarter[15 * 4 + 2:-3]
whole_w = quarter_sum_regular[:-12] + quarter_sum_errors[
12 * 4 + 9:-7] + LTS_quarter[15 * 4 + 2:]
#print(quarter_sum_errors)
cut = 12 * 4 + 1
n = len(quarter_sum) - cut
x = range(quarter_num - cut - 3)
width = 1 / 2
#print(len(LTS_quarter))
print(len(x))
print(len(quarter_sum_regular[cut:]))
print(len(quarter_sum_errors[cut:]))
bar1 = plt.bar(x,
quarter_sum_regular[cut:],
width,
color='darkblue',
label='regular',
edgecolor='black')
bar12 = plt.bar(x,
quarter_sum_errors[cut:],
width,
color='darkorange',
label='regular*',
edgecolor='black')
bar2 = plt.bar(x,
LTS_quarter[cut:],
width,
color='darkred',
label='long-term',
edgecolor='black')
plt.legend(handles=[bar1, bar12, bar2])
plt.xticks(np.arange(0, n), quartersx[cut:], rotation="vertical")
plt.ylabel('Vulnerabilities per quarter')
plt.xlabel('Quarter')
carlosplt.post_paper_plot(True, True, True)
## Linear Regression
print(len(x))
print(len(whole_w[cut:]))
X = sm.add_constant(x)
y = whole_w[cut:]
model = sm.OLS(y, X).fit()
predictions = model.predict(X)
plt.plot(predictions)
plt.show()
print(model.summary())
print(model.summary().as_latex())
def plot_total(self):
self.year_sum = [0] * self.years
self.year_num = [0] * self.years
for pkg in self.src2month_loc:
for j in range(self.years):
temp = sum(self.src2month_loc[pkg][12 * (1 + j):12 * (2 + j)])
if (temp > 0):
self.year_num[j] += 1
self.year_sum[j] += temp
## For last 2 years
total = sum(self.src2month_loc[pkg][:])
last_years = sum(self.src2month_loc[pkg][-24:])
#print(pkg + '; ' + str(last_years))
if (total > 1):
self.src2sum[pkg] = total
self.src2lastyears[pkg] = last_years
#calc total
sum_total = 0
one_only = 0
one_plus = 0
for p in self.src2month:
sum_part = sum(self.src2month_loc[p][:])
sum_total += sum_part
if (sum_part == 1):
one_only += 1
elif (sum_part > 1):
one_plus += 1
print('Total = ', sum_total)
print('one_only = ', one_only)
print('one_plus = ', one_plus)
values = sorted(self.src2sum.values(), reverse=True)
#print(values)
keys = list(
sorted(self.src2sum, key=self.src2sum.__getitem__, reverse=True))
n = len(self.year_sum)
yearsx = []
for i in range(1, self.years + 1):
if i % 2 == 1:
yearsx.append('\'' + str(i).zfill(2))
else:
yearsx.append('')
x = range(self.years)
width = 1 / 2
plt.bar(x, self.year_sum, width, color='darkblue', edgecolor='black')
#plt.bar(x, average_per_year, width)
plt.xticks(np.arange(0, n), yearsx)
plt.ylabel('Total vulnerabilities')
plt.xlabel('Year')
carlosplt.post_paper_plot(True, True, True)
sum_all = sum(values)
print("Total: ", sum_all)
## Linear regression model
X = sm.add_constant(x)
y = self.year_sum
model = sm.OLS(y, X).fit()
predictions = model.predict(X)
plt.plot(predictions)
plt.show()
print(model.summary())
print(model.summary().as_latex())
print(results.power_law.discrete)
print('lognormal mu: ', results.lognormal.mu)
print('lognormal sigma: ', results.lognormal.sigma)
#custom_model=[]
#for i in sorted(mydata,reverse=True):
# ccdf =
#fig=results.plot_pdf(color='b', linewidth=2)
carlosplt.pre_paper_plot(True)
fig = results.plot_ccdf(color='darkblue', linestyle='-', label='data')
results.power_law.plot_ccdf(color='darkgreen', ax=fig, label='power-law fit')
#results.truncated_power_law.plot_ccdf(color = 'red', ax=fig)
#results.lognormal_positive.plot_ccdf(color = 'yellow', ax=fig)
#results.lognormal.plot_ccdf(color = 'brown', ax=fig)
#results.exponential.plot_ccdf(color = 'orange', ax=fig)
plt.ylabel('ccdf')
plt.xlabel('Vulnerabilities')
fig.legend()
carlosplt.post_paper_plot(True, True, True)
plt.show()
R, p = results.distribution_compare('power_law', 'exponential')
print('Exponential: ', R, p)
R, p = results.distribution_compare('power_law', 'stretched_exponential')
print('Stretched exponential: ', R, p)
R, p = results.distribution_compare('power_law', 'truncated_power_law')
print('Power law truncated: ', R, p)
R, p = results.distribution_compare('power_law', 'lognormal_positive')
print('Lognormal positive: ', R, p)
R, p = results.distribution_compare('power_law', 'lognormal')
print('Lognormal: ', R, p)
def plot_bounties(ff):
reports_team = dict()
sum_team = dict()
with open("reports_team.json", "r") as fp:
reports_team = json.load(fp)
with open("sum_team.json", "r") as fp:
sum_team = json.load(fp)
if ff < 2:
ibb_list = ['ibb-php', 'ibb-python', 'ibb-data', 'ibb-flash', 'ibb-nginx', 'ibb-perl', 'internet', 'ibb-openssl', 'ibb-apache']
print('list follows')
for j in ibb_list:
print(reports_team[j])
else:
ibb_list = [team for team in reports_team]
most_team = dict()
sum_bounty_team = dict()
for team in ibb_list:
old = 0.0
old_sum = 0.0
for report in reports_team[team]:
try:
new = float(report['total_awarded_bounty_amount'])
old_sum += new
except KeyError:
print('#'*80)
print(report)
print('Report id ', report['id'], ' - bounty not found')
continue
if new > old:
old = new
most_team[team] = old
sum_bounty_team[team] = old_sum
print(most_team)
print(sum_bounty_team)
month2sum = []
month2money = []
month2bountylist = []
#Years: 2001-2018
for i in range(12*18):
month2sum.append(0)
month2money.append(0.0)
month2bountylist.append([])
for team in ibb_list:
for report in reports_team[team]:
datetime_obj = parser.parse(report['latest_disclosable_activity_at'])
print(str(datetime_obj))
month2sum[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += 1
try:
#if report['severity_rating'] == "high":
if (ff==0 or ff ==2) or (report['severity_rating'] == "high") or (report['severity_rating'] == "critical"):
month2money[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += float(report['total_awarded_bounty_amount'])
month2bountylist[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += [float(report['total_awarded_bounty_amount'])]
except KeyError:
continue
print(month2bountylist)
#plt.plot(month2sum[-12*5:])
#plt.show()
#plt.plot(month2money[-12*5:])
#plt.show()
years = 18
quarter_num = years*4
quarter_sum = []
quarter_av = []
carlosplt.pre_paper_plot()
quarter2bountylist = []
quartersx = []
for i in range(1,years+1):
for j in range(1,5):
if j==1:
quartersx.append('Q' + str(j)+'\''+str(i).zfill(2))
else:
quartersx.append(' ')
for j in range(quarter_num):
temp2 = sum(month2money[3*j:3*(j+1)])
temp3 = [item for sublist in month2bountylist[3*j:3*(j+1)] for item in sublist]
temp1 = len(temp3)
print(temp3)
quarter_sum.append(temp1)
quarter2bountylist.append(temp3)
try:
quarter_av.append(temp2/temp1)
except ZeroDivisionError:
quarter_av.append(0)
n = len(quarter_sum[-4*5:])
x = range(len(quarter_sum[-4*5:]))
width = 1/2
#plt.bar(x[-4*5:], quarter_sum[-4*5:], width, color='brown', label='Number', edgecolor='black')
#plt.xticks(np.arange(0,n),quartersx[-4*5:], rotation="vertical")
#plt.ylabel('Number of rewards')
#plt.xlabel('Quarter')
#carlosplt.post_paper_plot(True,True,True)
#plt.show()
#
#plt.bar(x[-4*5:], quarter_av[-4*5:], width, color='darkblue', label='regular support', edgecolor='black')
#
#plt.xticks(np.arange(0,n),quartersx[-4*5:], rotation="vertical")
#plt.ylabel('Average bug price of IBB projects (USD)')
#plt.xlabel('Quarter')
#carlosplt.post_paper_plot(True,True,True)
#plt.show()
#print(quarter2bountylist)
if ff==0:
labeltext = 'IBB-all'
elif ff==1:
labeltext = 'IBB-high'
elif ff==2:
labeltext = 'All-all'
elif ff==3:
labeltext = 'All-high'
## Shapiro normality test for each quarter
## Added powerlaw test
reference = []
for i in quarter2bountylist:
reference+=i
print(reference)
for i in quarter2bountylist:
print(i)
data = i
if len(i)>3:
#sns.distplot(i)
#plt.show()
stat, p = shapiro(data)
print('Statistics=%.3f, p=%.3f' % (stat, p))
# interpret
alpha = 0.01
if p > alpha:
print('Sample looks Gaussian (fail to reject H0)')
else:
print('Sample does not look Gaussian (reject H0)')
w,p = ks_2samp(i,reference)
if p > alpha:
print('Samples look similar')
else:
print('Samples do not look similar')
#mydata = i
#results=powerlaw.Fit(mydata, discrete=False, xmax=5000)
#print('alpha = ',results.power_law.alpha)
#print(results.truncated_power_law.alpha)
#print('xmin = ',results.power_law.xmin)
#print('xmax = ',results.power_law.xmax)
#print('sigma = ',results.power_law.sigma)
#print('D = ',results.power_law.D)
#print(results.truncated_power_law.xmin)
#print('xmax = ', results.truncated_power_law.xmax)
#print(results.power_law.discrete)
#print('lognormal mu: ',results.lognormal.mu)
#print('lognormal sigma: ',results.lognormal.sigma)
## Linear regression of average and median
# Average
xx = []
yy = quarter_av[-4*5:]
y = []
counter=0
for i in yy:
if i!=0:
y.append(i)
xx.append(counter)
counter+=1
X = sm.add_constant(xx)
model = sm.OLS(y,X).fit()
predictions = model.predict(X)
plt.plot(xx,predictions)
print(model.summary())
print(model.summary().as_latex())
xx = []
yy = quarter2bountylist[-4*5:]
y = []
counter=0
for i in yy:
if i!=[]:
y.append(median(i))
xx.append(counter)
counter+=1
X = sm.add_constant(xx)
model = sm.OLS(y,X).fit()
predictions = model.predict(X)
plt.plot(xx,predictions, color='darkred')
print(model.summary())
print(model.summary().as_latex())
## Create box plot
bp = plt.boxplot((quarter2bountylist[-4*5:]), whis = [5,95], patch_artist=True, positions = x)
plt.setp(bp['boxes'], color='black')
plt.setp(bp['whiskers'], color='darkred')
plt.setp(bp['caps'], color='darkred')
plt.setp(bp['fliers'], markersize = 3.0)
plt.yscale('log')
plt.ylim(top=50000)
plt.ylim(bottom=1)
plt.xticks(np.arange(0,n),quartersx[-4*5:], rotation="vertical")
plt.ylabel(labeltext)
plt.xlabel('Quarter')
carlosplt.post_paper_plot(True,True,True)
def plot_demographics(ff):
if ff==0:
labeltext = 'num - IBB'
elif ff==1:
labeltext = 'new - IBB'
elif ff==2:
labeltext = 'num - All'
elif ff==3:
labeltext = 'new - All'
reports_team = dict()
sum_team = dict()
with open("reports_team.json", "r") as fp:
reports_team = json.load(fp)
with open("sum_team.json", "r") as fp:
sum_team = json.load(fp)
if ff < 2:
ibb_list = ['ibb-php', 'ibb-python', 'ibb-data', 'ibb-flash', 'ibb-nginx', 'ibb-perl', 'internet', 'ibb-openssl', 'ibb-apache']
print('list follows')
for j in ibb_list:
print(reports_team[j])
else:
ibb_list = [team for team in reports_team]
most_team = dict()
sum_bounty_team = dict()
for team in ibb_list:
old = 0.0
old_sum = 0.0
for report in reports_team[team]:
try:
new = float(report['total_awarded_bounty_amount'])
old_sum += new
except KeyError:
print('#'*80)
print(report)
print('Report id ', report['id'], ' - bounty not found')
continue
if new > old:
old = new
most_team[team] = old
sum_bounty_team[team] = old_sum
print(most_team)
print(sum_bounty_team)
month2sum = []
month2money = []
month2bountylist = []
month2newreporters = []
repuntilnow = []
#Years: 2001-2018
for i in range(12*18):
month2sum.append(0)
month2newreporters.append(0)
month2money.append(0.0)
month2bountylist.append([])
for team in ibb_list:
for report in reports_team[team]:
datetime_obj = parser.parse(report['latest_disclosable_activity_at'])
print(str(datetime_obj))
month2sum[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += 1
try:
reporter=report['reporter']['id']
#if report['severity_rating'] == "high":
month2money[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += float(report['total_awarded_bounty_amount'])
month2bountylist[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += [float(report['total_awarded_bounty_amount'])]
if reporter not in repuntilnow:
month2newreporters[(int(datetime_obj.year)-2001)*12 + datetime_obj.month] += 1
repuntilnow.append(reporter)
except KeyError:
print('Error with report ', report['id'])
continue
print(month2bountylist)
#plt.plot(month2sum[-12*5:])
#plt.show()
#plt.plot(month2money[-12*5:])
#plt.show()
years = 18
quarter_num = years*4
quarter_sum = []
quarter_av = []
carlosplt.pre_paper_plot()
quarter2bountylist = []
quartersx = []
for i in range(1,years+1):
for j in range(1,5):
if j==1:
quartersx.append('Q' + str(j)+'\''+str(i).zfill(2))
else:
quartersx.append(' ')
for j in range(quarter_num):
temp2 = sum(month2money[3*j:3*(j+1)])
temp4 = sum(month2newreporters[3*j:3*(j+1)])
temp3 = [item for sublist in month2bountylist[3*j:3*(j+1)] for item in sublist]
temp1 = len(temp3)
if ff==1 or ff==3:
quarter_sum.append(temp4)
else:
quarter_sum.append(temp1)
n = len(quarter_sum[-4*5:])
x = range(len(quarter_sum[-4*5:]))
width = 1/2
#print(quarter2bountylist)
reference = []
for i in quarter2bountylist:
reference+=i
print(reference)
## Create bars plot
plt.bar(x[-4*5:], quarter_sum[-4*5:], width, color='darkblue', label='Number', edgecolor='black')
plt.xticks(np.arange(0,n),quartersx[-4*5:], rotation="vertical")
plt.ylabel(labeltext)
plt.xlabel('Quarter')
carlosplt.post_paper_plot(True,True,True)