karma/crab
1
0
Fork 0
mirror of https://github.com/karma-riuk/crab.git synced 2025-07-05 13:38:12 +02:00
Code Issues Packages Projects Releases Wiki Activity

made so that binary files content are ignored

Browse Source
This commit is contained in:
Karma Riuk
2025-03-28 18:18:36 +01:00
parent 69bf557a61
commit d7cba34e3d
1 changed files with 26 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
dataset.py
View File

@ -142,6 +142,21 @@ def migrate(dataset: Dataset) -> Dataset_new:
ret.entries.append(new_entry) ret.entries.append(new_entry)
return ret return ret
def try_decode(content: bytes) -> str:
try:
return content.decode()
except UnicodeDecodeError:
return "Binary file (from API), to be ignored"
def try_read_file(fname: str) -> str:
if not os.path.exists(fname):
return "" # file was removed after the PR
try:
with open(fname, "r", encoding="utf-8") as f:
return f.read()
except FileNotFoundError:
return "Binary file (from filesystem), to be ignored"
def new_files(repo: Repository, pr: PullRequest, new_metadata: Metadata_new, old_entry: DatasetEntry, repo_path: str) -> dict[str, FileData_new]: def new_files(repo: Repository, pr: PullRequest, new_metadata: Metadata_new, old_entry: DatasetEntry, repo_path: str) -> dict[str, FileData_new]:
review_comments = list(pr.get_review_comments()) review_comments = list(pr.get_review_comments())
if not review_comments: if not review_comments:
@ -161,16 +176,23 @@ def new_files(repo: Repository, pr: PullRequest, new_metadata: Metadata_new, old
assert isinstance( assert isinstance(
contents, ContentFile contents, ContentFile
), f"Multiple files with the same name {fname} in base sha {comment_commit_id} ({contents})" ), f"Multiple files with the same name {fname} in base sha {comment_commit_id} ({contents})"
content_before = contents.decoded_content.decode() content_before = try_decode(contents.decoded_content)
except Exception as e: except Exception as e:
content_before = "" # file didn't exist before the PR content_before = "" # file didn't exist before the PR
if old_entry.metadata.reason_for_failure == "Couldn't fetch the PR's merge commit": if old_entry.metadata.reason_for_failure == "Couldn't fetch the PR's merge commit":
content_after = "" content_after = ""
else: else:
run_git_cmd(["checkout", pr.merge_commit_sha], repo_path) try:
with open(os.path.join(repo_path, fname), "r") as f: contents = repo.get_contents(fname, ref=pr.merge_commit_sha)
content_after = f.read() assert isinstance(
contents, ContentFile
), f"Multiple files with the same name {fname} in base sha {comment_commit_id} ({contents})"
content_after = try_decode(contents.decoded_content)
except Exception as e:
run_git_cmd(["checkout", pr.merge_commit_sha], repo_path)
content_after = try_read_file(os.path.join(repo_path, fname))
ret[fname] = FileData_new( ret[fname] = FileData_new(
is_code_related=fname.endswith('.java'), is_code_related=fname.endswith('.java'),