def _prep_repo(self, path_repo: str) -> Generator[Git, None, None]:
local_path_repo = path_repo
if self._is_remote(path_repo):
local_path_repo = self._clone_remote_repo(self._clone_folder(),
path_repo)
local_path_repo = str(Path(local_path_repo).expanduser().resolve())
# when multiple repos are given in input, this variable will serve as a reminder
# of which one we are currently analyzing
self._conf.set_value('path_to_repo', local_path_repo)
self.git = Git(local_path_repo, self._conf)
# saving the Git object for further use
self._conf.set_value("git", self.git)
# checking that the filters are set correctly
self._conf.sanity_check_filters()
yield self.git
# cleaning, this is necessary since GitPython issues on memory leaks
self._conf.set_value("git", None)
self.git.clear()
self.git = None # type: ignore
# delete the temporary directory if created
if self._is_remote(path_repo) and self._cleanup is True:
assert self._tmp_dir is not None
try:
self._tmp_dir.cleanup()
except PermissionError:
# on Windows, Python 3.5, 3.6, 3.7 are not able to delete
# git directories because of read-only files.
# In this case, just ignore the errors.
shutil.rmtree(self._tmp_dir.name, ignore_errors=True)
def test_commit_in_master_branch(repo: Git):
assert repo.get_head().hash == '29e929fbc5dc6a2e9c620069b24e2a143af4285f'
repo.checkout('8986af2a679759e5a15794f6d56e6d46c3f302f1')
git_to_change_head = Git('test-repos/branches_merged')
commit = git_to_change_head.get_commit('8169f76a3d7add54b4fc7bca7160d1f1eede6eda')
assert commit.in_main_branch is False
repo.reset()
assert repo.get_head().hash == '29e929fbc5dc6a2e9c620069b24e2a143af4285f'
def test_changed_methods():
gr = Git("test-repos/diff")
# add a new method
mod = gr.get_commit(
'ea95227e0fd128aa69c7ab6a8ac485f72251b3ed').modified_files[0]
assert len(mod.changed_methods) == 1
assert mod.changed_methods[0].name == 'GitRepository::singleProjectThirdMethod'
# add 2 new methods
mod = gr.get_commit(
'd8eb8e80b671246a43c98d97b05f6d1c5ada14fb').modified_files[0]
assert len(mod.changed_methods) == 2
# remove one method
mod = gr.get_commit(
'0c8f9fdec926785198b399a2c49adb5884aa952c').modified_files[0]
assert len(mod.changed_methods) == 1
# add and remove one one method at different locations
mod = gr.get_commit(
'd8bb142c5616041b71cbfaa11eeb768d9a1a296e').modified_files[0]
assert len(mod.changed_methods) == 2
# add and remove one one method at the same location
# this is equivalent to replacing a method - although we expect 2 methods
mod = gr.get_commit(
'9e9473d5ca310b7663e9df93c402302b6b7f24aa').modified_files[0]
assert len(mod.changed_methods) == 2
# update a method
mod = gr.get_commit(
'b267a14e0503fdac36d280422f16360d1f661f12').modified_files[0]
assert len(mod.changed_methods) == 1
# update and add a new method
mod = gr.get_commit(
'2489099dfd90edb99ddc2c82b62524b66c07c687').modified_files[0]
assert len(mod.changed_methods) == 2
# update and delete methods
mod = gr.get_commit(
'5aebeb30e0238543a93e5bed806639481460cd9a').modified_files[0]
assert len(mod.changed_methods) == 2
# delete 3 methods (test cleanup - revert the test file to its
# initial set of methods)
mod = gr.get_commit(
'9f6ddc2aac740a257af59a76860590cb8a84c77b').modified_files[0]
assert len(mod.changed_methods) == 3
def repo(request):
gr = Git(request.param)
yield gr
gr.clear()
def extract(self,
labeled_files: List[FailureProneFile],
product: bool = True,
process: bool = True,
delta: bool = False):
""" Extract metrics from labeled files.
Parameters
----------
labeled_files : List[FailureProneFile]
The list of FailureProneFile objects that are used to label a script as failure-prone (1) or clean (0).
product: bool
Whether to extract product metrics.
process: bool
Whether to extract process metrics.
delta: bool
Whether to extract delta metrics between two successive releases or commits.
"""
self.dataset = pd.DataFrame()
git_repo = Git(self.path_to_repo)
metrics_previous_release = dict() # Values for iac metrics in the last release
for commit in Repository(self.path_to_repo, order='date-order', num_workers=1).traverse_commits():
# To handle renaming in metrics_previous_release
for modified_file in commit.modified_files:
old_path = modified_file.old_path
new_path = modified_file.new_path
if old_path != new_path and old_path in metrics_previous_release:
# Rename key old_path wit new_path
metrics_previous_release[new_path] = metrics_previous_release.pop(old_path)
if commit.hash not in self.commits_at:
continue
# Else
git_repo.checkout(commit.hash)
process_metrics = {}
if process:
# Extract process metrics
i = self.commits_at.index(commit.hash)
from_previous_commit = commit.hash if i == 0 else self.commits_at[i - 1]
to_current_commit = commit.hash # = self.commits_at[i]
process_metrics = self.get_process_metrics(from_previous_commit, to_current_commit)
for filepath in self.get_files():
file_content = get_content(os.path.join(self.path_to_repo, filepath))
if not file_content or self.ignore_file(filepath, file_content):
continue
tmp = FailureProneFile(filepath=filepath, commit=commit.hash, fixing_commit='')
if tmp not in labeled_files:
label = 0 # clean
else:
label = 1 # failure-prone
metrics = dict(
filepath=filepath,
commit=commit.hash,
committed_at=str(commit.committer_date),
failure_prone=label
)
if process_metrics:
metrics['change_set_max'] = process_metrics['dict_change_set_max']
metrics['change_set_avg'] = process_metrics['dict_change_set_avg']
metrics['code_churn_count'] = process_metrics['dict_code_churn_count'].get(filepath, 0)
metrics['code_churn_max'] = process_metrics['dict_code_churn_max'].get(filepath, 0)
metrics['code_churn_avg'] = process_metrics['dict_code_churn_avg'].get(filepath, 0)
metrics['commits_count'] = process_metrics['dict_commits_count'].get(filepath, 0)
metrics['contributors_count'] = process_metrics['dict_contributors_count'].get(filepath, 0)
metrics['minor_contributors_count'] = process_metrics['dict_minor_contributors_count'].get(filepath, 0)
metrics['highest_contributor_experience'] = process_metrics[
'dict_highest_contributor_experience'].get(filepath, 0)
metrics['hunks_median'] = process_metrics['dict_hunks_median'].get(filepath, 0)
metrics['additions'] = process_metrics['dict_additions'].get(filepath, 0)
metrics['additions_max'] = process_metrics['dict_additions_max'].get(filepath, 0)
metrics['additions_avg'] = process_metrics['dict_additions_avg'].get(filepath, 0)
metrics['deletions'] = process_metrics['dict_deletions'].get(filepath, 0)
metrics['deletions_max'] = process_metrics['dict_deletions_max'].get(filepath, 0)
metrics['deletions_avg'] = process_metrics['dict_deletions_avg'].get(filepath, 0)
if product:
metrics.update(self.get_product_metrics(file_content))
if delta:
delta_metrics = dict()
previous = metrics_previous_release.get(filepath, dict())
for metric, value in previous.items():
if metric in ('filepath', 'commit', 'committed_at', 'failure_prone'):
continue
difference = metrics.get(metric, 0) - value
delta_metrics[f'delta_{metric}'] = round(difference, 3)
metrics_previous_release[filepath] = metrics.copy()
metrics.update(delta_metrics)
self.dataset = self.dataset.append(metrics, ignore_index=True)
git_repo.reset()
def repo():
path = "test-repos/dmm-test-repo"
gr = Git(path)
yield gr
gr.clear()
