Prompt Template Issue in modules/lumberchunker.py

[dzbao]

Hello, during reading the source code of lmchunker, especially the realization for LumberChunker method, I find its Chinese prompt is:

system_prompt_zh = """你将被赋予一个英文文档，其中段落用'ID XXXX: <text>'标识。

任务：找出第一个段落（不是第一个），其中内容与前面的段落相比较，显着地发生变化。

输出：以如下格式返回段落的ID：'Answer: ID XXXX'。

其他注意事项：避免形成过长的段落集群。考虑句子逻辑和语义内容。寻找一个平衡，以识别内容变化并保持段落集群长度适中。"""

(line 32 to line 38 in modules/lumberchunker.py

I suppose that using LumberChunker with Chinese LLM for Chinese Documents processing should be accompanied by "你将被赋予一个中文文档" instead of "你将被赋予一个英文文档".

[Robot2050]

We are deeply grateful for your profound interest in this project and for raising such meticulous questions. We feel honored and will promptly refine and update the relevant content in response to your inquiry. Thank you once again for your invaluable feedback！