fix(translate): pipeline 写库用 service 返回的 engine(之前写死 tencent)

pipeline.translate_article 调 translation_service.translate() 后,
自己写 engine_label='tencent',忽略 service 实际返回的 engine。
效果:即使 service 内部按 spark 优先链路跑成功了,
DB 里 translation_engine 也只显示 tencent。

修法: 改用 tr_title.engine(or 默认 tencent 兜底)。

附带: deploy_pull.py 加 sys.stdout.reconfigure(encoding='utf-8'),
Windows GBK 终端下 Unicode 字符不会再 UnicodeEncodeError。
This commit is contained in:
Mavis
2026-06-10 23:31:03 +08:00
parent b27643123e
commit a8e93cf7c7
2 changed files with 20 additions and 12 deletions

View File

@@ -182,7 +182,8 @@ async def translate_article(article_id: int) -> None:
translated_chunks.append(tr.text)
tr_body = "\n\n".join(translated_chunks)
engine_label = "tencent"
# 用 service 返回的 engine 标签(spark / tencent / tencent_maas / agnes / nllb / cache)
engine_label = tr_title.engine or "tencent"
status = "ok" if (tr_title.text and tr_body) else "partial"
except Exception as e:
logger.exception("translate article %s failed: %s", article_id, e)