Is their a simple implementation of that

At a glance

Is their a simple implementation of that in isolation? This codebase is quite messy for head as a starting point. I would like to understand the basic concept

11 comments

ggangadhar

whats your use case

AAbhimanyu Aryan 🧞

right now the use case is just replicate the it

ggangadhar

so you want to make a evctor store for local files?

AAbhimanyu Aryan 🧞

exactly and see how to do it

AAbhimanyu Aryan 🧞

and then pass it to OpenAI API

AAbhimanyu Aryan 🧞

without using llamaindex

AAbhimanyu Aryan 🧞

maybe a simple http request with payload

ggangadhar

i'd say dont rely on llamaindex. Do this-
1) Make a chromadb (preferred ) or qdrant vector collection locally.
2) Use any embeddings and chunking size based on that to make the vector records for db.
3) load the vector in app with something like this-

Plain Text

def get_vector_store(client):
    vector_store = QdrantVectorStore(
    client=client, collection_name=""
    )
    return vector_store
storage_context = StorageContext.from_defaults(vector_store=vector_store)
embed_model = # your embedding model used to create vector db
    index = 
VectorStoreIndex.from_vector_store(storage_context=storage_context, embed_model=embed_model)

you can further query the vector store with chat_engine using OpenAI. Otherwise I'd recommend using standalone code for retrieving documents and passing it OAI api

ggangadhar

dm me if you have questions

AAbhimanyu Aryan 🧞

sure

AAbhimanyu Aryan 🧞

thanks

Add a reply

Find answers from the community

Is their a simple implementation of that