From 9ecfadcb7c2903a611d91cdd7d99c174d061bdf3 Mon Sep 17 00:00:00 2001
From: watsonk1998 <1515673657@qq.com>
Date: Mon, 13 Apr 2026 22:27:30 +0800
Subject: [PATCH] docs: add CJK language showcase example

---
 examples/cjk-showcase/README.md                    | 14 ++++++++++++++
 examples/cjk-showcase/raw/2026-04-13-reflection.md |  7 +++++++
 2 files changed, 21 insertions(+)
 create mode 100644 examples/cjk-showcase/README.md
 create mode 100644 examples/cjk-showcase/raw/2026-04-13-reflection.md

diff --git a/examples/cjk-showcase/README.md b/examples/cjk-showcase/README.md
new file mode 100644
index 0000000..ef104ab
--- /dev/null
+++ b/examples/cjk-showcase/README.md
@@ -0,0 +1,14 @@
+# CJK Showcase (Chinese Language Example)
+
+This directory demonstrates how LLM Wiki Agent performs with Non-English (CJK) languages.
+
+The agent naturally supports processing Chinese content. With the CJK query bug fixed, you can ingest, query, and linguistically search across Chinese entries without any language-specific configuration. 
+
+## Files included in this showcase:
+
+- `raw/2026-04-13-reflection.md`: A sample source document (a personal reflection on career transition).
+- `wiki/sources/2026-04-13-reflection.md`: The parsed structured source page.
+- `wiki/entities/杨帆.md`: Auto-extracted Chinese entity page.
+- `wiki/concepts/AI转型.md`: Auto-extracted Chinese concept page.
+
+Try running `python tools/query.py "关于AI转型的建议"` from the root directory after moving these to your main knowledge base to see how semantic extraction and keyword matching behave in non-English contexts!
diff --git a/examples/cjk-showcase/raw/2026-04-13-reflection.md b/examples/cjk-showcase/raw/2026-04-13-reflection.md
new file mode 100644
index 0000000..3316642
--- /dev/null
+++ b/examples/cjk-showcase/raw/2026-04-13-reflection.md
@@ -0,0 +1,7 @@
+# 2026-04-13 关于AI转型的复盘总结
+
+今天和杨帆深入讨论了土木工程转向AI产品经理的路径。他提到最大的陷阱是“工具旅游（Tool Tourism）”——很多非技术背景的人沉迷于尝试各种AI工具，却忽略了业务本质和产品交付。
+
+真正的破局点在于将大模型视为一种新的计算范式，而不是魔术。我们需要关注模型稳定性、成本、并发以及长上下文的召回率。同时，我也在思考目前个人的技术栈，从玩提示词到掌握Agentic Workflow框架（如LangChain或自定义多Agent系统），这是一个质的飞跃。
+
+决定下一步：减少看泛科普文章，直接深入开源社区，比如通过贡献代码或者提出架构Issue来积累实际影响力。