How to use pre-chunked documents as input to graphrag?

[zhangyuanwang777]

I have a pre-chunked document with only 100 tokens per text chunk, and I want to use grapgrag to analyze it. I found that there is an option in the source code that defaults to False for prechunked. How can I change it to True in the configuration file? And what data format should I use to input my document data?

[zhangyuanwang777]

This problem can be solved by passing data in csv format

[lyl2000zy]

What about the code change?Can you give me a specific process,please ?

[adhazel]

I found this to be most helpful: https://github.com/microsoft/graphrag/blob/main/docs/config/env_vars.md#input-data

Essentially, prechunked is set by the system and not something you need to set in your vars manually. Instead, just use the CSV format for input and that in of itself will end up setting your prechunked to True.

Hope this helps.