Viewing File: /home/ubuntu/codegamaai-test/efimarket_bot/src/retriver_chunk.py

from langchain.vectorstores import Qdrant
import qdrant_client
from src.constants import *
from langchain.embeddings import OpenAIEmbeddings
import os
# from pronouns import *
os.environ["OPENAI_API_KEY"] = os.environ["OPENAI_API_KEY"]
embedding = OpenAIEmbeddings()

client = qdrant_client.QdrantClient(
    qdrant_url,
    api_key=None )
def vectore_retrival_algo(vectordb,query,n_doc,score,algo):
    query = str(query)
    if algo == "similarity_score_threshold":
        return vectordb.as_retriever(search_type="similarity_score_threshold", search_kwargs={"score_threshold": score, "k": n_doc}).get_relevant_documents(query=query)
    else:
        return vectordb.as_retriever(search_type="mmr",search_kwargs={'k': n_doc, 'lambda_mult': score}).get_relevant_documents(query=query)

def retirve_method_1(user_id,bot_id,history,query,n_history_questions,n_doc_retirve,lambda_mult,algo):
    try:
        collection_id = str(user_id) + "_" + str(bot_id)
        split_query = []
        # if check_conjunctions_in_sentence(query) == True:
        #     split_query =  preprocess_query(query)
        vectordb = Qdrant(client=client,embeddings=embedding, collection_name=collection_id)
        doc = vectore_retrival_algo(vectordb,query,n_doc_retirve,lambda_mult,algo)
        # # n_contents = [item["content"] for item in history if item["role"] == "user"][-n_history_questions:]
        # n_contents = n_contents + split_query
        # for i in range(len(n_contents)):
        #     last_user_value = n_contents[i]
        #     doc1 = vectore_retrival_algo(vectordb,last_user_value,2,lambda_mult,algo)
            # doc.extend(doc1)
        return doc
    except:
        print("retirve_method_1 exception")
        return []