20250312 2

[yuji38kwmt]

No description provided.

[kci-pr-agent]

Title

20250312 2

---

Description

• download_annotation_archiveにジョブ完了待機機能を追加

• HTTP 409/WOKER_ALREADY_INVOKED時にリトライ実装

• 例外キャッチを細分化してエラー処理改善

• Locationヘッダ参照箇所を整理

---

Changes walkthrough 📝

	Relevant files
Enhancement	wrapper.py download_annotation_archiveの待機・例外処理強化                                        annofabapi/wrapper.py download_annotation_archiveにjob_access_intervalとmax_job_access引数を追加 HTTP 409エラーでWORKER_ALREADY_INVOKED検知し待機後リトライ wait_until_job_finished呼び出しによる更新ジョブ完了待機 例外キャッチをValueError, KeyError, TypeErrorなどに限定 +38/-2

---

Need help?

Type /help how to ... in the comments thread for any questions about PR-Agent usage.

Check out the documentation for more information.

[kci-pr-agent]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review 例外処理のカバレッジ get_annotation_archive が HTTPError 以外のステータスコード 409 を返す場合や、リトライ後にも同様のエラーが発生したときの挙動が不明瞭です。意図した例外が確実に伝播し、無限ループや未処理の状態にならないか確認してください。 def download_annotation_archive( self, project_id: str, dest_path: str | Path, *, job_access_interval: int = 60, max_job_access: int = 360, ) -> str: """ simpleアノテーションZIPをダウンロードする。 アノテーションZIPの更新中（HTTP 409, WORKER_ALREADY_INVOKED）の場合は、 更新ジョブ（gen-annotation）が完了するまで待ってからダウンロードする。 Args: project_id: プロジェクトID dest_path: ダウンロード先のファイルパス job_access_interval: アノテーションZIPの更新ジョブにアクセスする間隔[秒] max_job_access: アノテーションZIPの更新ジョブに最大何回アクセスするか Returns: ダウンロード元のURL """ try: _, response = self.api.get_annotation_archive(project_id) except requests.exceptions.HTTPError as e: if e.response is not None and e.response.status_code == requests.codes.conflict: try: error_codes = [err["error_code"] for err in e.response.json()["errors"]] except (ValueError, KeyError, TypeError): error_codes = [] if "WORKER_ALREADY_INVOKED" in error_codes: logger.info( "project_id='%s' :: アノテーションZIPの更新中のため、更新ジョブ(gen-annotation)が完了するまで待ちます。", project_id, ) self.wait_until_job_finished( project_id, ProjectJobType.GEN_ANNOTATION, job_access_interval=job_access_interval, max_job_access=max_job_access, ) _, response = self.api.get_annotation_archive(project_id) else: raise else: raise ロガー参照の明示 メソッド内で logger を使用していますが、モジュールの先頭で適切に定義・インポートされているか確認をお願いします。ローカルスコープでの未定義エラーを防止してください。 logger.info( "project_id='%s' :: アノテーションZIPの更新中のため、更新ジョブ(gen-annotation)が完了するまで待ちます。", project_id, )

[Copilot]

Copilot wasn't able to review any files in this pull request.

---

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

The defaults allow blocking for up to 6 hours (60s * 360) inside a synchronous download method, which can be problematic for CLI/CI usage and services (long-running requests, worker timeouts). Consider using a single timeout_seconds parameter (derived internally), lowering the defaults, and/or documenting prominently that this call may block for a long time by default.

[Copilot]

After waiting, the second get_annotation_archive(project_id) call is not protected by the same 409/WORKER_ALREADY_INVOKED handling. If the worker is still running (or a new run starts), this will raise HTTPError and bypass the intended wait/retry behavior. Consider wrapping the retry in the same handling (or using a small bounded loop that retries on 409/WORKER_ALREADY_INVOKED until the polling budget is exhausted).

Suggested change

	max_retry_on_conflict = 2
	retry_count = 0
	while True:
	try:
	_, response = self.api.get_annotation_archive(project_id)
	break
	except requests.exceptions.HTTPError as e:
	if e.response is not None and e.response.status_code == requests.codes.conflict:
	try:
	error_codes = [err["error_code"] for err in e.response.json()["errors"]]
	except (ValueError, KeyError, TypeError):
	error_codes = []
	if "WORKER_ALREADY_INVOKED" in error_codes and retry_count < max_retry_on_conflict:
	logger.info(
	"project_id='%s' :: アノテーションZIPの更新中のため、更新ジョブ(gen-annotation)が完了するまで待ちます。",
	project_id,
	)
	self.wait_until_job_finished(
	project_id,
	ProjectJobType.GEN_ANNOTATION,
	job_access_interval=job_access_interval,
	max_job_access=max_job_access,
	)
	retry_count += 1
	continue

[Copilot]

Accessing response.headers[\"Location\"] will raise a raw KeyError if the header is missing, which is hard to diagnose. Consider using an explicit check (e.g., get) and raising a clearer exception that includes project_id, status code, and available headers (or a short subset) to aid debugging.