ERROR: Result' object has no attribute 'top_k

[ashwinzyx]

Facing issue with cross-encoder rerankers

Loading TransformerRanker model mixedbread-ai/mxbai-rerank-large-v1
No device set
Using device mps
No dtype set
Using dtype torch.float16
Loaded model mixedbread-ai/mxbai-rerank-large-v1
Using device mps.
Using dtype torch.float16.
/Users/ashwinaravind/.pyenv/versions/3.12.5/lib/python3.12/site-packages/langchain/hub.py:86: DeprecationWarning: The `langchainhub sdk` is deprecated.
Please use the `langsmith sdk` instead:
  pip install langsmith
Use the `pull_prompt` method.
  res_dict = client.pull_repo(owner_repo_commit)
/Users/ashwinaravind/.pyenv/versions/3.12.5/lib/python3.12/site-packages/posthog/client.py:356: DeprecationWarning: datetime.datetime.utcnow() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.datetime.now(datetime.UTC).
  timestamp = datetime.utcnow().replace(tzinfo=tzutc())
Number of requested results 100 is greater than number of elements in index 1, updating n_results = 1
/Users/ashwinaravind/.pyenv/versions/3.12.5/lib/python3.12/site-packages/posthog/request.py:40: DeprecationWarning: datetime.datetime.utcnow() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.datetime.now(datetime.UTC).
  body["sentAt"] = datetime.utcnow().replace(tzinfo=tzutc()).isoformat()
An error occurred: 'Result' object has no attribute 'top_k'

sample code below

from langchain_openai import ChatOpenAI

from langchain_community.document_loaders import *
from langchain_openai import OpenAIEmbeddings
from langchain.text_splitter import MarkdownHeaderTextSplitter
from langchain_chroma import Chroma
import chromadb

from rerankers import Reranker
from langchain.retrievers.document_compressors import EmbeddingsFilter
from langchain.retrievers import ContextualCompressionRetriever
import os
from operator import itemgetter
from langchain import hub
from langchain_core.output_parsers import StrOutputParser
from langchain_core.runnables import RunnablePassthrough, RunnableParallel, RunnableLambda
from langchain.retrievers import MergerRetriever,EnsembleRetriever
from langchain.retrievers.document_compressors import DocumentCompressorPipeline

def rag_pipeline():
    try:
        def format_docs(docs):
            return "\n".join(doc.page_content for doc in docs)

        llm=ChatOpenAI(model='gpt-4o-mini')

        loader = WebBaseLoader('https://ashwinaravind.github.io/')
        docs = loader.load()

        embedding=OpenAIEmbeddings(model='text-embedding-3-small')

        headers_to_split_on = [
            ("#", "Header 1"),
            ("##", "Header 2"),
            ("###", "Header 3")]
        splitter = MarkdownHeaderTextSplitter(headers_to_split_on=headers_to_split_on, strip_headers=False)
        splits=splitter.split_text(docs[0].page_content)
        c=Chroma.from_documents(documents=splits, embedding=embedding, collection_name='testindex-ragbuilder-1727764109091', client_settings=chromadb.config.Settings(allow_reset=True))
        retrievers=[]
        retriever=c.as_retriever(search_type='similarity', search_kwargs={'k': 100})
        retrievers.append(retriever)
        retriever=EnsembleRetriever(retrievers=retrievers)
        arr_comp=[]
        ranker = Reranker(model_name='mixedbread-ai/mxbai-rerank-large-v1', model_type='cross-encoder')
        compressor = ranker.as_langchain_compressor(k=1)
        arr_comp.append(compressor) 
        retriever=ContextualCompressionRetriever(base_retriever=retriever,base_compressor=compressor)
        prompt = hub.pull("rlm/rag-prompt")
        rag_chain = (
            RunnableParallel(context=retriever, question=RunnablePassthrough())
                .assign(context=itemgetter("context") | RunnableLambda(format_docs))
                .assign(answer=prompt | llm | StrOutputParser())
                # .pick(["answer", "context"])
                )
        print(rag_chain.invoke('How does rag work'))
    except Exception as e:
        print(f"An error occurred: {e}") 
rag_pipeline()