infiniflow
[Question]: The maximum TopN is 30 ？
Self Checks
- [x] I have searched for existing issues search for existing issues, including closed ones.
- [x] I confirm that I am using English to submit this report (Language Policy).
- [x] Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
- [x] Please do not modify this template :) and fill in all the required fields.
Describe your problem
The maximum TopN is 30, and the processing table records are sometimes much larger than 30 rows, how to solve the problem?
Top N means N chunks which can include many records for each. That depends on the chunk size.
The cvs table slice is a block of one record, TopN 30, LLM can only analyze and process these 30 records, what if I want to analyze more data?
Top Nmeans N chunks which can include many records for each. That depends on the chunk size.
Chunk method
Did you utilize
Generalas chunking method?
chunking method is table, the upload file is csv, after parsing is a record a block, csv table has 2000 records, the number of chunks is 2000, now retrieve one of the 800 records so that the LLM analysis, but the maximum TopN is only 30, I can only analyze the analysis of 800 records in the 30 analysis, resulting in inaccurate results!
I have also encountered the same problem. How can you solve it
My temporary solution is to change the max value in top-n-item.tsx under src, and then run build
I have also encountered the same problem. How can you solve it
My temporary solution is to change the max value in top-n-item.tsx under src, and then run build
If this is done, it will actually cause unnecessary pressure on the LLM model, which is equivalent to passing all the chunks to the LLM network. The communication time is cost, and the actual effect of RAFLOW should not be like this. I think the best way to process Excel and analyze it is MCP protocol, but my current data is stored on Elastic Search, which can be directly connected to the latest version of Elastic. Elastic can call AI directly. The data pulled in this way is full data, and I will no longer use RAFLOW for data analysis and querying