`documents SimpleDirectoryReader data

At a glance

documents = SimpleDirectoryReader('data').load_data()
parser = SimpleNodeParser.from_defaults()
nodes = parser.get_nodes_from_documents(documents)
index = VectorStoreIndex(nodes)
query_engine = index.as_query_engine()
response = query_engine.query("What is this book all about?")
print(response)

with this code i'm facing RateLimitError

if i do this without nodes
like this...

documents = SimpleDirectoryReader('data').load_data()
index = VectorStoreIndex.from_documents(documents)
query_engine = index.as_query_engine()
response = query_engine.query("Short Summery of book?")
print(response)

i'm not getting RateLimitError

why so any work around? 🤨

9 comments

LLogan M

I think it might be a weird coincidence 😅

Do you have a paid openai account? The free tokens are severly rate limited

HHimalayPatel

nope. free 🥲

LLogan M

Try using local embeddings for now

LLogan M

Plain Text

from llama_index.service_context import ServiceContext, set_global_service_context

service_context = ServiceContext.from_defaults(embed_model="local:BAAI/bge-small-en-v1.5")

set_global_service_context(service_context)

LLogan M

It will download that model (~200MB) and run it locally for embeddings

LLogan M

but still call openai for generating the responses

LLogan M

should help get around the rate limits

MMaximus

@Logan M is the simpledirectoryloader also creating/parsing the nodes? I'm confused by that part.

LLogan M

It's creating documents (which is basically the same as nodes tbh). Different loaders will create documents differently

Add a reply

Find answers from the community

`documents SimpleDirectoryReader data