向量知识库添加内容错误，但没有报错，显示为×

[lceCre4m]

请问有大佬遇到这种情况吗？该怎么解决 1、源文件和向量库都为×

2、检索上传过的文档一模一样的内容查找不到

3、对话一直为“未找到相关文档,该回答为大模型自身能力解答！”

[lceCre4m]

而且都启动和文档上传及添加向量数据库过程都没有报错

[dosubot[bot]]

To address the issues you're facing with the Langchain-Chatchat application, here are focused solutions:

1. Adding Content to the Vector Knowledge Base Failing:

   • Verify that all necessary models and modules, such as sentence-transformers and modelscope, are correctly installed and accessible.
   • A workaround mentioned in similar issues suggests commenting out lines 202 and 203 in knowledge_base.py. This might help but proceed with caution and consider the implications on your setup.
   • Use the recreate_vector_store endpoint after uploading your documents to ensure they are vectorized and recognized by the system.

2. Issues Retrieving Uploaded Documents:

   • Confirm that your documents are correctly uploaded by checking if the system recognizes the CSV file. The system should acknowledge the file's presence for it to be searchable.
   • Review the logs for any errors during the document upload process to ensure there were no failures or issues that might prevent document retrieval.
   • If retrieval issues persist, consider adjusting the distance_strategy to MAX_INNER_PRODUCT as it resolved similar issues for other users.

These steps are derived from solutions to similar problems reported in the repository. If these do not resolve your issues, providing more details about the documents (format, size, content) and any specific error messages or logs could help further diagnose the problem.

Details

Sources

• [BUG] 上传文件后，知识库文件加载失败
• [BUG] 知识库上传csv没有进入向量库
• [BUG] 上传json/jsonl文件且向量化后，网页端看不到划分后的知识库并且模型找不到相应的检索知识库
• zhipu-api轻量版本，知识库问答报错‘未找到相关文档,该回答为大模型自身能力解答！’
• [BUG] 无法使用zhipu的在线embedding
• [BUG] v0.2.8 知识库上传文件会生成两份文件记录
• server/knowledge_base/kb_service/default_kb_service.py
• server/knowledge_base/kb_service/es_kb_service.py
• server/api.py
• server/chat/file_chat.py
• .gitmodules
• knowledge_base/samples/content/test_files/langchain-ChatGLM_closed.jsonl
• knowledge_base/samples/content/test_files/langchain-ChatGLM_open.jsonl

About Dosu This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

To continue the conversation, mention @dosu.

[zRzRzRzRzRzRzR]

能提问正常回答就是向量化了，否则后台会报错

[HanqiaoZheng]

能提问正常回答就是向量化了，否则后台会报错

不行，我也遇到同样问题了，回答知识库点开显示红色字体，未找到相关文档,该回答为大模型自身能力解答！

[HanqiaoZheng]

楼主解决了吗，我也遇到这个问题了

[mrkingsun]

我也遇到了，0.2.10版本，上传后源文件和向量库都是x，这什么bug？

[lwh8915]

我也遇到了，请问解决了吗

[ZG688]

我也遇到了 一点错误没有 就是找不到文档

[danny-zhu]

我这边也遇到相同的问题，可以试试把model_config中用的模型路径写完整就可以解决了。MODEL_ROOT_PATH这参数可能是有的地方没有自动拼接上，我还没来得及去逐行排查代码，你们遇到可以试试看。

[ZG688]

我这边也遇到相同的问题，可以试试把model_config中用的模型路径写完整就可以解决了。MODEL_ROOT_PATH这参数可能是有的地方没有自动拼接上，我还没来得及去逐行排查代码，你们遇到可以试试看。

model_config.py这个文件里面的模型路径绝对的相对的都能用，LLM问答里面没有问题，构建向量库也没有报错，就是知识库问答找不到相关文档，后台也没错误，成功加载向量库 麻了

[ReverseSystem001]

我这边也遇到相同的问题，可以试试把model_config中用的模型路径写完整就可以解决了。MODEL_ROOT_PATH这参数可能是有的地方没有自动拼接上，我还没来得及去逐行排查代码，你们遇到可以试试看。

model_config.py这个文件里面的模型路径绝对的相对的都能用，LLM问答里面没有问题，构建向量库也没有报错，就是知识库问答找不到相关文档，后台也没错误，成功加载向量库 麻了

问题解决了吗

哥们，问题解决了吗？