if i manually add embeddings for a

Yes it will make embeddings for the test_docs

again?

is there a way to skip that process and still make the index?

It doesn't generate any embeddings, i tested witha smaller batch

However, could you help me with something else?

You can make an index with an empty list also
index= VectorStoreIndex.from_documents([],...)

This way it will create an index but it won't have any content in it

The number of test_docs is a lot different from the number of nodes

is there a way to make them similar?

I want a node to have the same content as a document

Test_docs are of type Document
Which are usually one page per document.


Nodes are created to have uniformity in the text size for readability and other stuff.

If you want you can directly create nodes of your own size and then in place of using from_documents you'll have to use node method.

Typing is taking time, as I'm away from the system at present 😆

Now worries

i just use from-node? I do not care if they are non uniform

Yep you'll have to use this method: https://github.com/run-llama/llama_index/blob/main/llama-index-core%2Fllama_index%2Fcore%2Findices%2Fbase.py#L180

thanks bro