[Bug] rag功能不是很好使,请指点下
sparkssssssss opened this issue · comments
Bug Description
没有如作者图中调用插件;
存储中可以看到向量文件,但是都是乱码
对话如下
pdf内容如下
Steps to Reproduce
暂无
Expected Behavior
暂无
Screenshots
No response
Deployment Method
- Docker
- Vercel
- Server
Desktop OS
No response
Desktop Browser
No response
Desktop Browser Version
No response
Smartphone Device
No response
Smartphone OS
No response
Smartphone Browser
No response
Smartphone Browser Version
No response
Additional Logs
No response
Bot detected the issue body's language is not English, translate it automatically.
Title: [Bug] The rag function is not very easy to use, please give me some advice.
要不你把 pdf 发出来我试试?
Bot detected the issue body's language is not English, translate it automatically.
How about you send the pdf so I can try it?
我感觉不是pdf的问题,我之前传过几次,也有正常显示的,但是我从来没有出现过调用插件rag那个提示,所以想在请教下,正确的姿势?
目前我尝试过:
1-关闭所有插件,无效
2-关闭所有插件,除了那个pdf查看的,无效
3-不管插件,无效
我使用的是最新night的镜像
Bot detected the issue body's language is not English, translate it automatically.
I don’t think it’s a problem with PDF. I’ve uploaded it several times before and it was displayed normally. However, I’ve never had the prompt to call the plug-in rag, so I’d like to ask for advice on the correct posture?
So far I've tried:
1-Close all plug-ins, invalid
2-Close all plug-ins, except the one for PDF viewing, which is invalid.
3- Regardless of the plug-in, it is invalid
I am using the latest night image
Bot detected the issue body's language is not English, translate it automatically.
I tried the following PDF. This PDF file should have been specially processed. The text copied directly from the file is also garbled. This kind of file cannot correctly parse the text. ![image](https://private-user-images.githubusercontent.com/14031260/337267597-1b67047c-e8f2-43fd-b1f5-02437945d995.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3Mi OiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3MTc3MzY5ODYsIm5iZiI6MTcxNzczNjY4NiwicGF0aCI6Ii8xNDAzMTI2MC 8zMzcyNjc1OTctMWI2NzA0N2MtZThmMi00M2ZkLWIxZjUtMDI0Mzc5NDVkOTk1LnBuZz9YLUFtei1 BbGdvcml0aG09QVdTNC1ITUFDLVNIQTI1NiZYLUFtei1DcmVkZW50aWFsPUFLSUFWQ09EWUxTQTUz UFFLNFpBJTJGMjAyNDA2MDclMkZ1cy1lYXN0LTElMkZzMyUyRmF3czRfcmVxdWVzdCZYLUFtei1EYXRlPTIwMjQwNjA3VDA1MDQ0NlomWC1BbXotRXhwaXJlcz0zMDAmWC1BbXotU2lnbmF0dXJlPW Q3NWNhMmM1MjI4MjcxODAwMmQwNmZkY2ZiMTMyZWFhNDAzNDM4MDIwNGVkMWZlMDQzYzA3MDUxOTJkNTdlY2UmWC1BbXotU2lnbmVkSGVhZGVycz1ob3N0JmFjdG9yX2lkPTAma2V5X2lkPTAmcmV wb19pZD0wIn0.Ws3KWs67hYuWKlQZEPLYOoRxSEREFKt-JkRt0RdaWQA)
I uploaded a random pdf and generated a vector file, but I can’t figure out why.
由于现在是基于插件的形式进行检索的,所以需要引导 gpt 使用 rag-search 才会返回相关上下文。
这个月事情有点多,等待后面的优化吧。
Bot detected the issue body's language is not English, translate it automatically.
Since the retrieval is now based on a plug-in, you need to guide gpt to use rag-search to return the relevant context.
There are a lot of things going on this month, let’s wait for the optimization later.