Unable to create Test data with TestsetGenerator #2274
Description
[ ] I have checked the documentation and related resources and couldn't resolve my bug.
Describe the bug
Unable to create Test data with TestsetGenerator.
It's using invalid themes internally: List[Tuple] instead of List[str]
Ragas version: 0.3.4
Python version:3.12.2
Code to Reproduce
def test_dataCreation():
llm=ChatOpenAI(model_name="gpt-4",temperature=0)
langchain_llm=LangchainLLMWrapper(llm)
embed=OpenAIEmbeddings()
generate_embed=LangchainEmbeddingsWrapper(embed)
loader=DirectoryLoader(
path="[**provided a valid path**]",
glob="**/*.docx",
loader_cls=UnstructuredWordDocumentLoader,
)
docs=loader.load()
genertor=TestsetGenerator(llm=langchain_llm,embedding_model=generate_embed)
dataset=genertor.generate_with_langchain_docs(docs,testset_size=5)Error trace
E pydantic_core._pydantic_core.ValidationError: 6 validation errors for ThemesPersonasInput
E themes.0
E Input should be a valid string [type=string_type, input_value=('Rest API Testing', 'REST API'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
E themes.1
E Input should be a valid string [type=string_type, input_value=('postman', 'Postman'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
E themes.2
E Input should be a valid string [type=string_type, input_value=('REST Assured API', 'REST Assured API'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
E themes.3
E Input should be a valid string [type=string_type, input_value=('RestAssured', 'REST Assured API'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
E themes.4
E Input should be a valid string [type=string_type, input_value=('POSTMAN', 'Postman'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
E themes.5
E Input should be a valid string [type=string_type, input_value=('RESTAPI', 'REST API'), input_type=tuple]
E For further information visit https://errors.pydantic.dev/2.11/v/string_type
/opt/anaconda3/lib/python3.12/site-packages/ragas/testset/synthesizers/multi_hop/specific.py:95: ValidationError
**Expected behavior**
Pydantic model expects themes to be a list of strings, but it's being passed a list of tuples.
Additional context
Is there any workaround? Do I need to try with any other version of Ragas?