v2.3.2 Astro Bilingual Publishing Integration Demo

This is a temporary article for verifying the local end-to-end link of the Obsidian Engine Publish plugin. It will test the complete path from proofreading metadata on the Obsidian side, creating an English mirror draft, publishing via wxengine, pushing to GitHub, triggering Astro build and deployment, to the CDN page being visible.

This paragraph contains some explicit classification keywords: AI, Large Language Model, LLM, ChatGPT. The plugin’s auto-classification and tagging logic should be able to assign it to ai-era or generate relevant tags. If the classification does not hit, the current plugin falls back to life, which also needs to be documented in the integration notes.

The goal of this article is not long-term retention, but to confirm that the bilingual publishing pipeline works in a real environment. After the Chinese draft is published successfully, the English mirror draft should reuse the same source_id, slug, and categories, changing only the language, title, description, tags, and body content.

Integration checkpoints include:

• Whether metadata proofreading fills in the necessary frontmatter
• Whether the English mirror draft is generated in the same directory
• Whether the translation task goes through the queued -> running -> ok/failed status flow
• Whether Astro bilingual sync publishing returns zh/en two routes
• Whether the online site displays both Chinese and English articles after the propagation window