def get_winners_info():
aom = {}
aoy = {}
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
author, mail = parts[4].split('<', 1)
author = author.rstrip()
date = datetime.datetime.fromtimestamp(float(stamp))
# author of the month/year
yymm = date.strftime('%Y-%m')
if yymm in aom:
aom[yymm][author] = aom[yymm].get(author, 0) + 1
else:
aom[yymm] = {}
aom[yymm][author] = 1
yy = date.year
if yy in aoy:
aoy[yy][author] = aoy[yy].get(author, 0) + 1
else:
aoy[yy] = {}
aoy[yy][author] = 1
return aom, aoy
def get_weekly_hourly_activity():
activity_by_hour_of_week = {}
activity_by_hour_of_week_busiest = 0
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
date = datetime.datetime.fromtimestamp(float(stamp))
# activity
# hour
hour = date.hour
# day of week
day = date.weekday()
# hour of week
if day not in activity_by_hour_of_week:
activity_by_hour_of_week[day] = {}
activity_by_hour_of_week[day][hour] = activity_by_hour_of_week[day].get(hour, 0) + 1
# most active hour?
if activity_by_hour_of_week[day][hour] > activity_by_hour_of_week_busiest:
activity_by_hour_of_week_busiest = activity_by_hour_of_week[day][hour]
return activity_by_hour_of_week, activity_by_hour_of_week_busiest
def get_tags_info():
tags = {}
lines = get_pipe_output(['git show-ref --tags']).split('\n')
for line in lines:
if len(line) == 0:
continue
(hash, tag) = line.split(' ')
tag = tag.replace('refs/tags/', '')
output = get_pipe_output(
['git log "%s" --pretty=format:"%%at %%aN" -n 1' % hash])
if len(output) > 0:
parts = output.split(' ')
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
tags[tag] = {
'stamp': int(stamp),
'hash': hash,
'date':
datetime.datetime.fromtimestamp(stamp).strftime('%Y-%m-%d'),
'commits': 0,
'authors': {}
}
# collect info on tags, starting from latest
tags_sorted_by_date_desc = [
el[1] for el in reversed(
sorted([(el[1]['date'], el[0]) for el in tags.items()]))
]
prev = None
for tag in reversed(tags_sorted_by_date_desc):
cmd = 'git shortlog -s "%s"' % tag
if prev is not None:
cmd += ' "^%s"' % prev
output = get_pipe_output([cmd])
if len(output) == 0:
continue
prev = tag
for line in output.split('\n'):
parts = re.split('\s+', line, 2)
commits = int(parts[1])
author = parts[2]
tags[tag]['commits'] += commits
tags[tag]['authors'][author] = commits
return tags
def get_gnuplot_version(self):
if self.gnuplot_version_string is None:
reg = re.compile(r"(\d+)\.(\d+)\.?(\d+)?")
version_str = get_pipe_output(['%s --version' % self.gnuplot_executable]).split('\n')[0]
match = reg.search(version_str)
if match:
self.gnuplot_version_string = version_str[match.span()[0]:match.span()[1]]
return self.gnuplot_version_string
def get_num_files_in_revision(time_rev):
"""
Get number of files at a given revision
"""
time, rev = time_rev
return int(time), rev, int(
get_pipe_output(['git ls-tree -r --name-only "%s"' % rev,
'wc -l']).split('\n')[0])
def process_gnuplot_scripts(self, scripts_path, data_path, output_images_path):
scripts = glob.glob(os.path.join(scripts_path, '*.plot'))
os.chdir(output_images_path)
for script in scripts:
gnuplot_command = '%s -e "data_folder=\'%s\'" "%s"' % (self.configuration.gnuplot_executable, data_path, script)
out = get_pipe_output([gnuplot_command])
if len(out) > 0:
print(out)
def get_authors_history():
lines = get_pipe_output([
'git log --shortstat --date-order --pretty=format:"%%at %%aN" %s' %
(getlogrange('HEAD'))
]).split('\n')
lines.reverse()
inserted = 0
stamp = 0
tmp_authors = {}
changes_by_date_by_author = {}
for line in lines:
if len(line) == 0:
continue
# <stamp> <author>
if re.search('files? changed', line) is None:
pos = line.find(' ')
if pos != -1:
try:
oldstamp = stamp
(stamp, author) = (int(line[:pos]), line[pos + 1:])
if oldstamp > stamp:
# clock skew, keep old timestamp to avoid having ugly graph
# FIXME: in the original version the clock skew was used to avoid splashes on the graphs
# due to cherry-picking, rebases and amendments. So far the skew is switched of and supposed to
# be fixed using commiter time in future
# stamp = oldstamp
pass
if author not in tmp_authors:
tmp_authors[author] = {'lines_added': 0, 'commits': 0}
tmp_authors[author]['commits'] = tmp_authors[author].get(
'commits', 0) + 1
tmp_authors[author][
'lines_added'] = tmp_authors[author].get(
'lines_added', 0) + inserted
if stamp not in changes_by_date_by_author:
changes_by_date_by_author[stamp] = {}
if author not in changes_by_date_by_author[stamp]:
changes_by_date_by_author[stamp][author] = {}
changes_by_date_by_author[stamp][author][
'lines_added'] = tmp_authors[author]['lines_added']
changes_by_date_by_author[stamp][author][
'commits'] = tmp_authors[author]['commits']
files, inserted, deleted = 0, 0, 0
except ValueError:
print('Warning: unexpected line "%s"' % line)
else:
print('Warning: unexpected line "%s"' % line)
else:
numbers = get_stat_summary_counts(line)
if len(numbers) == 3:
(files, inserted, deleted) = [int(el) for el in numbers]
else:
print('Warning: failed to handle line "%s"' % line)
(files, inserted, deleted) = (0, 0, 0)
return changes_by_date_by_author
def get_timezone_info():
timezones = {}
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
timezone = parts[3]
timezones[timezone] = timezones.get(timezone, 0) + 1
return timezones
def test_rev_list(self):
revlines = get_pipe_output([
'git rev-list --pretty=format:"%%at %%T" %s' % getlogrange('HEAD'),
'grep -v ^commit'
]).strip().split('\n')
expected_data = []
for line in revlines:
ts, tree_id = line.split(' ')
expected_data.append((int(ts), tree_id))
actual_data = []
for t, r in self.gs.get_revisions():
actual_data.append((t, r))
self.assertListEqual(expected_data, actual_data)
def get_active_days_info():
active_days = set()
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
date = datetime.datetime.fromtimestamp(float(stamp))
yymmdd = date.strftime('%Y-%m-%d')
# project: active days
active_days.add(yymmdd)
return active_days
def get_monthly_activity_info():
activity_by_month_of_year = {}
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
date = datetime.datetime.fromtimestamp(float(stamp))
# activity
# month of year
month = date.month
activity_by_month_of_year[month] = activity_by_month_of_year.get(month, 0) + 1
return activity_by_month_of_year
def get_commits_count_change_timeline():
commits_by_month = {}
commits_by_year = {}
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
date = datetime.datetime.fromtimestamp(float(stamp))
yymm = date.strftime('%Y-%m')
commits_by_month[yymm] = commits_by_month.get(yymm, 0) + 1
yy = date.year
commits_by_year[yy] = commits_by_year.get(yy, 0) + 1
return commits_by_month, commits_by_year
def get_domain_info():
domains = {}
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
author, mail = parts[4].split('<', 1)
mail = mail.rstrip('>')
domain = '?'
if mail.find('@') != -1:
domain = mail.rsplit('@', 1)[1]
# domain stats
if domain not in domains:
domains[domain] = {}
# commits
domains[domain]['commits'] = domains[domain].get('commits', 0) + 1
return domains
def get_activity_by_year_week():
activity_by_year_week = {}
activity_by_year_week_peak = 0
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
date = datetime.datetime.fromtimestamp(float(stamp))
# activity
# yearly/weekly activity
yyw = date.strftime('%Y-%W')
activity_by_year_week[yyw] = activity_by_year_week.get(yyw, 0) + 1
if activity_by_year_week_peak < activity_by_year_week[yyw]:
activity_by_year_week_peak = activity_by_year_week[yyw]
return activity_by_year_week, activity_by_year_week_peak
def test_missing_authors_email(self):
import tempfile
import os
from tools import shellhelper
cwd = os.getcwd()
with tempfile.TemporaryDirectory() as tmpdir:
# change working directory
os.chdir(tmpdir)
shellhelper.get_pipe_output(['git init'])
shellhelper.get_pipe_output(['touch file.txt'])
shellhelper.get_pipe_output(['git add file.txt'])
# apparently it is not possible to create pygit.Signature with empty author's email (and name)
# but commits with no author's email can be created via git
shellhelper.get_pipe_output([
'git commit -m "No-email author" --author "Author NoEmail <>"'
])
try:
# Commit without author's email does not crash statistics calculation
GitStatistics(tmpdir)
except Exception as e:
self.fail(str(e))
os.chdir(cwd)
def get_authors_info():
authors = {}
# Collect revision statistics
# Outputs "<stamp> <date> <time> <timezone> <author> '<' <mail> '>'"
lines = get_pipe_output(
['git rev-list --pretty=format:"%%at %%ai %%aN <%%aE>" %s' % getlogrange('HEAD'), 'grep -v ^commit']).split(
'\n')
for line in lines:
parts = line.split(' ', 4)
try:
stamp = int(parts[0])
except ValueError:
stamp = 0
author, mail = parts[4].split('<', 1)
author = author.rstrip()
date = datetime.datetime.fromtimestamp(float(stamp))
# author stats
if author not in authors:
authors[author] = {}
# commits, note again that commits may be in any date order because of cherry-picking and patches
if 'last_commit_stamp' not in authors[author]:
authors[author]['last_commit_stamp'] = stamp
if stamp > authors[author]['last_commit_stamp']:
authors[author]['last_commit_stamp'] = stamp
if 'first_commit_stamp' not in authors[author]:
authors[author]['first_commit_stamp'] = stamp
if stamp < authors[author]['first_commit_stamp']:
authors[author]['first_commit_stamp'] = stamp
# authors: active days
yymmdd = date.strftime('%Y-%m-%d')
if 'last_active_day' not in authors[author]:
authors[author]['last_active_day'] = yymmdd
authors[author]['active_days'] = {yymmdd}
elif yymmdd != authors[author]['last_active_day']:
authors[author]['last_active_day'] = yymmdd
authors[author]['active_days'].add(yymmdd)
lines = get_pipe_output(
['git log --shortstat --date-order --pretty=format:"%%at %%aN" %s' % (getlogrange('HEAD'))]).split('\n')
lines.reverse()
inserted = 0
deleted = 0
stamp = 0
for line in lines:
if len(line) == 0:
continue
# <stamp> <author>
if re.search('files? changed', line) is None:
pos = line.find(' ')
if pos != -1:
try:
oldstamp = stamp
(stamp, author) = (int(line[:pos]), line[pos + 1:])
if oldstamp > stamp:
# clock skew, keep old timestamp to avoid having ugly graph
stamp = oldstamp
if author not in authors:
authors[author] = {'lines_added': 0, 'lines_removed': 0, 'commits': 0}
authors[author]['commits'] = authors[author].get('commits', 0) + 1
authors[author]['lines_added'] = authors[author].get('lines_added', 0) + inserted
authors[author]['lines_removed'] = authors[author].get('lines_removed', 0) + deleted
files, inserted, deleted = 0, 0, 0
except ValueError:
print('Warning: unexpected line "%s"' % line)
else:
print('Warning: unexpected line "%s"' % line)
else:
numbers = get_stat_summary_counts(line)
if len(numbers) == 3:
(files, inserted, deleted) = [int(el) for el in numbers]
else:
print('Warning: failed to handle line "%s"' % line)
(files, inserted, deleted) = (0, 0, 0)
return authors
def get_total_changes_timeline():
# line statistics
# outputs:
# N files changed, N insertions (+), N deletions(-)
# <stamp> <author>
changes_by_date = {} # stamp -> { files, ins, del }
lines_added_by_month = {}
lines_removed_by_month = {}
lines_added_by_year = {}
lines_removed_by_year = {}
total_lines_added = 0
total_lines_removed = 0
# computation of lines of code by date is better done
# on a linear history.
extra = ''
if conf['linear_linestats']:
extra = '--first-parent -m'
lines = get_pipe_output(
['git log --shortstat %s --pretty=format:"%%at %%aN" %s' % (extra, getlogrange('HEAD'))]).split('\n')
lines.reverse()
files = 0
inserted = 0
deleted = 0
total_lines = 0
for line in lines:
if len(line) == 0:
continue
# <stamp> <author>
if re.search('files? changed', line) is None:
pos = line.find(' ')
if pos != -1:
try:
(stamp, author) = (int(line[:pos]), line[pos + 1:])
changes_by_date[stamp] = {u'files': files, u'ins': inserted, u'del': deleted, u'lines': total_lines}
date = datetime.datetime.fromtimestamp(stamp)
yymm = date.strftime('%Y-%m')
lines_added_by_month[yymm] = lines_added_by_month.get(yymm, 0) + inserted
lines_removed_by_month[yymm] = lines_removed_by_month.get(yymm, 0) + deleted
yy = date.year
lines_added_by_year[yy] = lines_added_by_year.get(yy, 0) + inserted
lines_removed_by_year[yy] = lines_removed_by_year.get(yy, 0) + deleted
files, inserted, deleted = 0, 0, 0
except ValueError:
print('Warning: unexpected line "%s"' % line)
else:
print('Warning: unexpected line "%s"' % line)
else:
numbers = get_stat_summary_counts(line)
if len(numbers) == 3:
(files, inserted, deleted) = [int(el) for el in numbers]
total_lines += inserted
total_lines -= deleted
total_lines_added += inserted
total_lines_removed += deleted
else:
print('Warning: failed to handle line "%s"' % line)
(files, inserted, deleted) = (0, 0, 0)
return changes_by_date, total_lines_added, total_lines_removed
def query_gnuplot_version(self):
query_str = get_pipe_output(['%s --version' % self.gnuplot_executable
]).split('\n')[0]
return query_str
