Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Commit 6e682c8

Browse files
ishepardishepard
committed
start working on using git filters. major changes need to be done for this
1 parent 293b02c commit 6e682c8

6 files changed

Lines changed: 70 additions & 51 deletions

File tree

‎.github/workflows/continuous-integration-workflow.yml‎

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -45,7 +45,5 @@ jobs:
4545
if: matrix.os == 'ubuntu-latest' || matrix.os == 'macOS-latest'
4646
uses: codecov/codecov-action@v1
4747
with:
48-
token: ${{ secrets.CODECOV_TOKEN }}
4948
file: ./coverage.xml
5049
yml: .codecov.yml
51-
fail_ci_if_error: true

‎.pylintrc‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -244,7 +244,7 @@ indent-after-paren=4
244244
indent-string=' '
245245

246246
# Maximum number of characters on a single line.
247-
max-line-length=79
247+
max-line-length=150
248248

249249
# Maximum number of lines in a module.
250250
max-module-lines=1000

‎pydriller/git_repository.py‎

Lines changed: 8 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -121,16 +121,20 @@ def get_head(self) -> Commit:
121121
head_commit = self.repo.head.commit
122122
return Commit(head_commit, self._conf)
123123

124-
def get_list_commits(self, branch: str = None,
125-
reverse_order: bool = True) \
126-
-> Generator[Commit, None, None]:
124+
def get_list_commits(self, args: List = None) -> Generator[Commit, None, None]:
127125
"""
128126
Return a generator of commits of all the commits in the repo.
129127
130128
:return: Generator[Commit], the generator of all the commits in the
131129
repo
132130
"""
133-
for commit in self.repo.iter_commits(branch, reverse=reverse_order):
131+
# With no arguments, just analyze all the commits until the HEAD,
132+
# in reverse order
133+
if args is None or not args:
134+
args = [self.repo.head.commit.hexsha, '--reverse']
135+
136+
print(args)
137+
for commit in self.repo.iter_commits(args):
134138
yield self.get_commit_from_gitpython(commit)
135139

136140
def get_commit(self, commit_id: str) -> Commit:

‎pydriller/repository_mining.py‎

Lines changed: 10 additions & 17 deletions
Original file line numberDiff line numberDiff line change
@@ -139,14 +139,11 @@ def traverse_commits(self) -> Generator[Commit, None, None]:
139139
if self._conf.get('clone_repo_to'):
140140
clone_folder = str(Path(self._conf.get('clone_repo_to')))
141141
if not os.path.isdir(clone_folder):
142-
raise Exception("Not a directory: " \
143-
"{0}".format(clone_folder))
144-
path_repo = self._clone_remote_repos(clone_folder,
145-
path_repo)
142+
raise Exception("Not a directory: {0}".format(clone_folder))
143+
path_repo = self._clone_remote_repos(clone_folder, path_repo)
146144
else:
147145
tmp_folder = tempfile.TemporaryDirectory()
148-
path_repo = self._clone_remote_repos(tmp_folder.name,
149-
path_repo)
146+
path_repo = self._clone_remote_repos(tmp_folder.name, path_repo)
150147

151148
git_repo = GitRepository(path_repo, self._conf)
152149
self._conf.set_value("git_repo", git_repo)
@@ -155,19 +152,15 @@ def traverse_commits(self) -> Generator[Commit, None, None]:
155152
logger.info('Analyzing git repository in %s', git_repo.path)
156153

157154
if self._conf.get('filepath') is not None:
158-
self._conf.set_value('filepath_commits',
159-
git_repo.get_commits_modified_file(
160-
self._conf.get('filepath')))
155+
self._conf.set_value('filepath_commits', git_repo.get_commits_modified_file(self._conf.get('filepath')))
161156

162157
if self._conf.get('only_releases'):
163-
self._conf.set_value('tagged_commits',
164-
git_repo.get_tagged_commits())
165-
166-
for commit in git_repo.get_list_commits(self._conf.get(
167-
'only_in_branch'), not self._conf.get('reversed_order')):
168-
logger.info('Commit #%s in %s from %s', commit.hash,
169-
commit.committer_date,
170-
commit.author.name)
158+
self._conf.set_value('tagged_commits', git_repo.get_tagged_commits())
159+
160+
args = self._conf.build_args()
161+
162+
for commit in git_repo.get_list_commits(args):
163+
logger.info('Commit #%s in %s from %s', commit.hash, commit.committer_date, commit.author.name)
171164

172165
if self._conf.is_commit_filtered(commit):
173166
logger.info('Commit #%s filtered', commit.hash)

‎pydriller/utils/conf.py‎

Lines changed: 44 additions & 27 deletions
Original file line numberDiff line numberDiff line change
@@ -69,6 +69,7 @@ def sanity_check_filters(self):
6969
Check if the values passed by the user are correct.
7070
7171
"""
72+
self._check_correct_filters_order()
7273
self.check_starting_commit()
7374
self.check_ending_commit()
7475
self._check_timezones()
@@ -83,8 +84,6 @@ def sanity_check_filters(self):
8384
self.set_value("to_commit", None)
8485
self.set_value("single", single)
8586

86-
self._check_correct_filters_order()
87-
8887
if self.get('single') is not None:
8988
if any([self.get('since'),
9089
self.get('to'),
@@ -113,9 +112,7 @@ def _check_correct_filters_order(self):
113112
self.get('git_repo').get_commit(self.get('from_commit')),
114113
self.get('git_repo').get_commit(self.get('to_commit')))
115114

116-
if self.get('reversed_order') and chronological_order:
117-
self._swap_commit_fiters()
118-
elif not self.get('reversed_order') and not chronological_order:
115+
if not chronological_order:
119116
self._swap_commit_fiters()
120117

121118
def _swap_commit_fiters(self):
@@ -140,12 +137,17 @@ def check_starting_commit(self):
140137
raise Exception('You can only specify one between since, '
141138
'from_tag and from_commit')
142139
if self.get('from_tag') is not None:
143-
self.set_value('from_commit', self.get(
144-
"git_repo").get_commit_from_tag(self.get('from_tag')).hash)
140+
self.set_value('from_commit', self.get("git_repo").get_commit_from_tag(self.get('from_tag')).hash)
145141
if self.get('from_commit'):
146142
try:
147-
self.set_value('from_commit', self.get("git_repo").get_commit(
148-
self.get('from_commit')).hash)
143+
commit = self.get("git_repo").get_commit(self.get('from_commit'))
144+
if len(commit.parents) == 0:
145+
self.set_value('from_commit', [commit.hash])
146+
elif len(commit.parents) == 1:
147+
self.set_value('from_commit', ['^' + commit.hash + '^'])
148+
else:
149+
commits = ['^' + x for x in commit.parents]
150+
self.set_value('from_commit', commits)
149151
except Exception:
150152
raise Exception("The commit {} defined in the 'from_tag' "
151153
"or 'from_commit' filter does "
@@ -165,8 +167,8 @@ def check_ending_commit(self):
165167
"git_repo").get_commit_from_tag(self.get('to_tag')).hash)
166168
if self.get('to_commit'):
167169
try:
168-
self.set_value('to_commit', self.get("git_repo").get_commit(
169-
self.get('to_commit')).hash)
170+
commit = self.get("git_repo").get_commit(self.get('to_commit'))
171+
self.set_value('to_commit', commit.hash)
170172
except Exception:
171173
raise Exception("The commit {} defined in the 'to_tag' "
172174
"or 'to_commit' filter does "
@@ -182,6 +184,37 @@ def only_one_filter(arr):
182184
"""
183185
return len([x for x in arr if x is not None]) <= 1
184186

187+
def build_args(self):
188+
from_commit = self.get('from_commit')
189+
to_commit = self.get('to_commit')
190+
branch = self.get('only_in_branch')
191+
192+
args = []
193+
194+
if from_commit is not None or to_commit is not None:
195+
if from_commit is not None and to_commit is not None:
196+
args.extend(from_commit)
197+
args.append(to_commit)
198+
elif from_commit is not None:
199+
args.extend(from_commit)
200+
args.append('HEAD')
201+
else:
202+
args.append(to_commit)
203+
else:
204+
args.append('HEAD')
205+
206+
if branch is not None:
207+
args.append(branch)
208+
209+
if self.get('only_no_merge'):
210+
args.append('--no-merges')
211+
212+
if not self.get('reversed_order'):
213+
args.append('--reverse')
214+
215+
print(f'Returning args: {args}')
216+
return args
217+
185218
def is_commit_filtered(self, commit: Commit):
186219
# pylint: disable=too-many-branches,too-many-return-statements
187220
"""
@@ -201,22 +234,6 @@ def is_commit_filtered(self, commit: Commit):
201234
(self.get('to') is not None and
202235
commit.committer_date > self.get('to')):
203236
return True
204-
if self.get('from_commit') is not None and \
205-
self.get('from_commit') == commit.hash:
206-
self.set_value('from_commit_started', True)
207-
return False
208-
if self.get('from_commit') is not None and \
209-
self.get('from_commit') != commit.hash and \
210-
self.get('from_commit_started') is None:
211-
return True
212-
if self.get('to_commit') is not None and \
213-
self.get('to_commit') != commit.hash and \
214-
self.get('to_commit_reached') is not None:
215-
return True
216-
if self.get('to_commit') is not None and \
217-
self.get('to_commit') == commit.hash:
218-
self.set_value('to_commit_reached', True)
219-
return False
220237
if self.get('only_modifications_with_file_types') is not None:
221238
if not self._has_modification_with_file_type(commit):
222239
logger.debug('Commit filtered for modification types')

‎tests/integration/test_ranges.py‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -138,6 +138,13 @@ def test_from_and_to_commit(repository_mining_cc, expected_commits):
138138
assert len(repository_mining_cc) == expected_commits
139139

140140

141+
def test_from_and_to_commit_with_merge_commit():
142+
commits = RepositoryMining('test-repos/pydriller',
143+
from_commit="015f7144641a418f6a9fae4d024286ec17fd7ce8",
144+
to_commit="01d2f2fbeb6980cc5568825d008017ca8ca767d6").traverse_commits()
145+
assert len(list(commits)) == 3
146+
147+
141148
# FROM AND TO TAG
142149
@pytest.mark.parametrize('to_tag,expected_commits', [
143150
('v1.4', 3),

0 commit comments

Comments
 (0)