Error in Testset Generation - ExceptionInRunner: The runner thread which was running the jobs raised an exeception.

[jaymon0703]

[X] I have checked the documentation and related resources and couldn't resolve my bug.

Describe the bug test_generator.generate_with_langchain_docs() returns RuntimeError.

Please note i am using VertexAI models.

RuntimeError: Task <Task pending name='Task-13277' coro=<BaseChatModel._agenerate_with_cache() running at [c:\Users\MASKED\.venv\lib\site-packages\langchain_core\language_models\chat_models.py:617](file:///C:/Users/MASKED/.venv/lib/site-packages/langchain_core/language_models/chat_models.py:617)> cb=[gather.<locals>._done_callback() at [C:\Users\MASKED\AppData\Local\Programs\Python\Python310\lib\asyncio\tasks.py:718](file:///C:/Users/MASKED/AppData/Local/Programs/Python/Python310/lib/asyncio/tasks.py:718)]> got Future <Task pending name='Task-13278' coro=<UnaryUnaryCall._invoke() running at [c:\Users\MASKED\.venv\lib\site-packages\grpc\aio\_call.py:566](file:///C:/Users/MASKED/.venv/lib/site-packages/grpc/aio/_call.py:566)>> attached to a different loop

Ragas version: 0.1.3 Python version: 3.10.0

Code to Reproduce

from langchain.document_loaders import DirectoryLoader
loader = DirectoryLoader("<your_diretory_here>")
documents = loader.load()

from google.cloud import aiplatform
from langchain.llms import VertexAI
from langchain.chat_models import ChatVertexAI
from langchain.chains import APIChain
from langchain.embeddings import VertexAIEmbeddings

aiplatform.init(
    # your Google Cloud Project ID or number
    # environment default used is not set
    project='<your_project_id_here>',
    location="us-central1"
)

# create Langchain LLM and Embeddings
ragas_vertexai_llm = ChatVertexAI()
vertexai_embeddings = VertexAIEmbeddings()

from ragas.llms import LangchainLLMWrapper
from ragas.embeddings.base import LangchainEmbeddingsWrapper

ragas_vertexai_llm = LangchainLLMWrapper(ragas_vertexai_llm)
vertexai_embeddings = LangchainEmbeddingsWrapper(vertexai_embeddings)

from ragas.testset.extractor import KeyphraseExtractor
from langchain.text_splitter import TokenTextSplitter
from ragas.testset.docstore import InMemoryDocumentStore

splitter = TokenTextSplitter(chunk_size=1000, chunk_overlap=100)
keyphrase_extractor = KeyphraseExtractor(llm=ragas_vertexai_llm)

docstore = InMemoryDocumentStore(
    splitter=splitter,
    embeddings=vertexai_embeddings,
    extractor=keyphrase_extractor,
)

from ragas.testset import TestsetGenerator
from ragas.testset.evolutions import simple, reasoning, multi_context

test_generator = TestsetGenerator(
    generator_llm=ragas_vertexai_llm,
    critic_llm=ragas_vertexai_llm,
    embeddings=vertexai_embeddings,
    docstore=docstore,
)

distributions = {simple: 0.5, reasoning: 0.25, multi_context: 0.25}

# use generator.generate_with_llamaindex_docs if you use llama-index as document loader
testset = test_generator.generate_with_langchain_docs(
    documents=documents, test_size=10, distributions=distributions
)

[shahules786]

Can you share the OS info @jaymon0703 too? Are you facing this issue everytime?

[jaymon0703]

Thank you @shahules786

Error manifests every time.

{'platform': 'Windows', 'platform-release': '10', 'platform-version': '10.0.19045', 'architecture': 'AMD64', 'processor': 'Intel64 Family 6 Model 142 Stepping 12, GenuineIntel', 'ram': '16 GB'}

[HRUSHI1212]

I am also facing the same issue and i am using GPT models with python 3.10 on collab.

[jaymon0703]

If anyone has VertexAI examples for testset generation and evaluation it would be great to see it. Currently getting errors on both, with above for testset generation.

[bdeck8317]

I am also getting this same error. Any resolution? @jaymon0703

[HRUSHI1212]

while using gpt use ragas==0.1.4 langchain==0.1.5 it worked for me

[jaymon0703]

Sorry what is "GPT" you are referring to @HRUSHI1212?

[HRUSHI1212]

Sorry what is "GPT" you are referring to @HRUSHI1212?

While generating synthetic dataset I am using GPT-4 and gpt 3.5 turbo which is the default one.with the versions ragas==0.1.4 langchain==0.1.5

[jaymon0703]

Hey @shahules786 is it possible to update VertexAI docs with working code for generating testset? Thanks!

[wahidur028]

Hi there,

I am also facing the same error: "The runner thread which was running the jobs raised an exception."

When I use two different models for generator_llm and critic_llm and set test_size=2, the code works. However, in other cases, such as using the same model for both generator_llm and critic_llm or setting test_size to 3, 5, or 10, the code breaks and generates the error.

[jjmachan]

was addressed with #1093 do check out the latest versions, it should be fixed