Tighten the analyze prompts (runtime + recommendation platform) so the
model returns 8-10 question-style search queries (哪家好/有哪些/怎么选/
价格/排名/避坑 …) instead of generic tag words, and lift the keyword cap
in normalizeAnalyzeResult from 5 to 10 to keep them. Add tests covering
the new ten-keyword window and the question-like guidance in the seeded
analyze prompt.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Added support for external markdown synchronization in ArticleEditorCanvas.vue.
- Implemented a new function to sync markdown from props, ensuring the editor reflects external changes.
- Introduced a removeOutlineSection function in TemplateWizardView.vue to manage outline sections dynamically.
- Updated UI components to improve user experience, including replacing a-input with a-textarea for better text handling.
- Refactored CSS styles for a cleaner layout and improved responsiveness in TemplateWizardView.vue.
- Enhanced LLM generation requests to include response format options in ark.go and client.go.
- Updated template assist logic to enforce structured outline outputs in template_assist.go.
- Added tests for outline result decoding and prompt generation to ensure robustness.
- Created migration scripts to update prompt templates with new writing requirements for top X articles.
root