SuperSCC.rag.ConnectRAG

class SuperSCC.rag.ConnectRAG(host, api_key, location, url, collection_name, embedding_model)

Establishes a connection to the Qdrant vector store using the provided details.

Parameters:

• (dict (encode_kwargs) – Keyword arguments for loading the Hugging Face embedding model. Defaults to {“device”: “cpu”}.

• optional) – Keyword arguments for loading the Hugging Face embedding model. Defaults to {“device”: “cpu”}.

• (dict – Keyword arguments for the embedding model’s encoding process. Defaults to {“normalize_embeddings”: True}.

• optional) – Keyword arguments for the embedding model’s encoding process. Defaults to {“normalize_embeddings”: True}.

• (int (timeout) – The request timeout in seconds for the Qdrant client. Defaults to 1000.

• optional) – The request timeout in seconds for the Qdrant client. Defaults to 1000.

__init__(host, api_key, location, url, collection_name, embedding_model)

Methods

__init__(host, api_key, location, url, ...)
add_documents(file_path, file_type[, ...])
change_text_embedding(model_name[, ...])
connect_client([model_kwargs, ...])	Establishes a connection to the Qdrant vector store using the provided details.
create_rag_chain(vector_store, model, ...[, ...])
data_loader(file_path[, mode, metadata_columns])
format_docs(docs)
get_all_ids()
get_answer(gene_list[, query, ...])	The main entry point for asking a question.
get_relevant_segments()
highlight_docs()
hybrid_search([hierarchy_search, key, value])
recursive_search(path[, type])
refine_query()
rerank([model, top_n])
run_rag(llm_model, llm_api_key, llm_base_url)	Initializes the RAG chain using the established connection to the vector store.
score_documents([docs])
summary_res(res)
text_encode(text, model_name, location[, ...])
text_split(docs[, chunk_size, ...])
translator([query])
update_rag_chain([model, api_key, base_url, ...])	Updates components of the existing RAG chain, such as the LLM or prompt.