SummaryIndex and DocumentSummaryIndex

SummaryIndex was earlier called List index, It does not keep summary of any node, Based on the query, would fetch the nodes and would generate the response.

Whereas Document SummaryIndex would keep summary of each Document object created and will map the summary with each of the relatd node.

For your usecase, Document SummaryIndex would be best way to go.

You can find more on these here:
https://gpt-index.readthedocs.io/en/latest/examples/index_structs/doc_summary/DocSummary.html

https://gpt-index.readthedocs.io/en/stable/core_modules/data_modules/index/index_guide.html#summary-index-formerly-list-index

Thank you for the information 🙏🏽
I'm not sure I understand why in my case you think DocumentSummaryIndex would be better.. I'm not using many documents.. I got only one.. I need to extract a breakdown/decomposition from it.

As per your requirement, you wanted to create a structured summary for your document right!


So document summary index will do that, it will create the summary and you can access the summary for each document from the index.

I think a ListIndex/SummaryIndex will also work fine too. Although I'm not sure what a "decomposition structured summay" is 😅