HELLO! I AM SUZUKI!!!

I have no money and would like to enjoy LlamaIndex using acheong08/ChatGPT which is free and does not require machine specs.
I am aware that I need an LLM implementation for this, but what would be the minimum required methods for the LLM class?
I would like to know if there are any templates used to implement the LLM class.
I have a hard time with python where there is no interface.

Regarding retrieving similarity data per document, not per node
I know that "index.as_retriever(similarity_top_k=5)" can retrieve data for each node, but I would like to retrieve data for each document.
Due to the large number of tokens in a document, it is not possible to have a one-to-one relationship between a document and a node.
Is there any option or functionality that would allow us to retrieve per document?

I'm having trouble understanding a certain aspect, and I would greatly appreciate your clarification.

I've grasped that the fundamental structure of LlamaIndex encompasses three stages: Indexing, Retrieval, and Generation. However, I'm puzzled as to why the query inputs for both the Retrieval and Generation phases are identical.

To illustrate, when I use the command index.as_query_engine().query("Summarize hogehoge in 300 words or less."), it appears to function effectively. Yet, it seems illogical for the Retrieval phase to incorporate the specification "in 300 words or less" as part of its query. My assumption is that only "Summarize hogehoge" should be directed towards Retrieval, and the complete phrase "Summarize hogehoge in 300 words or less" should be reserved for Generation. Could you please confirm if my understanding is correct?

Furthermore, if my interpretation is accurate, is there a way to implement this in practice? Ideally, I envision a scenario where one could perform generation based on nodes retrieved via a command like nodes = index.as_retriever().retrieve("Summarize hogehoge").

Thankyou.

LlamaIndex's Index doesn't seem to be able to scratch the itch.
ListIndex: (good if you want to get the latest information once, but too costly because you have to dump the whole text)
VectorIndex: (useless without embedding API)
TableKeywordIndex: (can't limit the number of queries)

Given the above, I came up with the idea of using ElasticSearch to extract N highly relevant nodes and throw them into a ListIndex.

What do you think about this?
Is it already there?
Or is it already there?
Let me know if you have any concerns.

Hello!

For VectorStoreIndex, there is an argument similarity_top_k to limit the number of queries, but is there any way to limit the number of queries for KeywordTableIndex?

I know there is max_keywords_per_chunk, but this does not limit the number of queries, does it?

Thankyou