Hey - 我在这里给出了一些整体反馈：

• 在 resolve_lang_ids 中，新的“相同行”消歧逻辑依赖全局的 LANG_ORDER，并且会为每个存在歧义的文本重新计算 rows；建议将 LANG_ORDER（或预先按语言 ID 计算好的行映射）作为参数传入，这样更容易测试，也能避免重复的表查找。
• 在 expand_expected_from_ids 中，seen 集合在所有 lang_ids 之间是共享的，这会把行为改成“全局去重文本”；如果本意只是避免单行内重复，请将 seen 移到循环内部，这样它会在每个 lang_id 上重新初始化。

Please address the comments from this code review:

## Overall Comments
- In `resolve_lang_ids`, the new identical-row disambiguation logic relies on the global `LANG_ORDER` and recomputes `rows` for each ambiguous text; consider passing `LANG_ORDER` (or a precomputed per-lang-id row map) as an argument to make this easier to test and avoid repeated table lookups.
- The `seen` set in `expand_expected_from_ids` is shared across all `lang_ids`, which changes behavior to globally deduplicate texts; if the intention was only to avoid duplicates within a single row, move `seen` inside the loop so it resets for each `lang_id`.

Please address the comments from this code review:

## Overall Comments
- In `resolve_lang_ids`, the new identical-row disambiguation logic relies on the global `LANG_ORDER` and recomputes `rows` for each ambiguous text; consider passing `LANG_ORDER` (or a precomputed per-lang-id row map) as an argument to make this easier to test and avoid repeated table lookups.
- The `seen` set in `expand_expected_from_ids` is shared across all `lang_ids`, which changes behavior to globally deduplicate texts; if the intention was only to avoid duplicates within a single row, move `seen` inside the loop so it resets for each `lang_id`.

Pull request overview

本 PR 主要用于批量更新各 pipeline OCR 节点的 expected 文本，以适配更多语言/表述，并增强 sync_ocr_expected.py 在反查 i18n ID 时的兜底策略与输出去重，降低重复文本带来的识别噪声。

Changes:

• 增强 tools/i18n/sync_ocr_expected.py：歧义 ID 新增“多语言整行一致则自动选取”兜底，并对展开后的 expected 做去重。
• 大范围更新 assets/resource* 下 pipeline 的 OCR expected 列表：补充/调整中繁英日/韩等文本，清理重复项、修正顺序与覆盖范围。
• 同步修正部分节点的 expected 内容（如 “Refresh/更新/刷新”、“探索等级/Exploration Level”等）以提升跨语言识别命中率。

Reviewed changes

Copilot reviewed 51 out of 51 changed files in this pull request and generated 2 comments.