TypeError while calling langchain runnables

[Tharun1718]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

embedding_function = OpenAIEmbeddings(openai_api_key=OPENAI_API_KEY, model="text-embedding-ada-002")
vectorstore = Chroma.from_documents(documents, embedding_function)
retriever = vectorstore.as_retriever()

retriever_chain = (
    {"context": retriever, "question": RunnablePassthrough()}
    | retriever_prompt
    | llm
    | StrOutputParser()
)

final_chain = (RunnablePassthrough.assign(context=retriever_chain, table_names_to_assign = table_chain) | generate_sql_query)
final_chain.invoke({"question": "Name the agencies are present"})

Description

I am trying to pass a question to langchain runnable and I am getting the below error 68 try:
---> 69 return self._core_bpe.encode_ordinary(text)
70 except UnicodeEncodeError:
71 # See comment in encode

TypeError: argument 'text': 'dict' object cannot be converted to 'PyString'

while executing retriever_chain.invoke("question_string") then question is provided as string and it works, but while invoking final_chain and passing question as a dict ({"question":""}). I am getting the above error

System Info

Python 3.11.9
Langchain 0.2.8

[dosubot[bot]]

Hey @Tharun1718! 👋 I'm here to help you with any bugs, questions, or contributions you have in mind. If you need assistance, feel free to ask.

To resolve the TypeError when passing the question as a dictionary to final_chain, ensure that the dictionary key matches the input_key expected by the FalkorDBQAChain. The input_key is set to "query" by default. Therefore, the dictionary should have a key "query" with the question as its value.

Here is an example of how to correctly pass the question to final_chain:

question_dict = {"query": "Name the agencies are present"}
result = final_chain.invoke(question_dict)

Make sure that the key in the dictionary is "query" as defined by the input_key property in the FalkorDBQAChain class ^[1].

Additionally, if you are working with a SQL chain, the input dictionary should have a key "question" as defined by the SQLInput class. Here is an example of how to correctly pass the question to a SQL chain:

question_dict = {"question": "Name the agencies are present"}
response = final_chain.invoke(question_dict)

Ensure that the key in the dictionary is "question" as required by the SQLInput class ^[2].

To continue talking to Dosu, mention @dosu.

[Tharun1718]

@dosu I have passed the dictionary correctly with "question" key, I am still getting the type error I have mentioned.

[dosubot[bot]]

To resolve the TypeError when executing final_chain.invoke({"question": "Name the agencies are present"}), ensure that the invoke method is called with the correct type of input. According to the provided code, the invoke method expects an Input type, which can be a string or another type defined by the InputType property of the Runnable.

Here is the corrected code:

final_chain.invoke({"query": "Name the agencies are present"})

This should resolve the TypeError since the invoke method will now receive the expected input type. If final_chain is configured to accept a dictionary, ensure that the InputType property is correctly set to handle such input ^[1][2][3].

To continue talking to Dosu, mention @dosu.

[Tharun1718]

@dosu I am still getting the same error

[davidcorigliano]

Hey @dosu @Tharun1718, have you had any luck getting this to work? I have the same issue and can't figure out how to fix it. Appreciate any other ideas you might have - thanks

[dosubot[bot]]

To resolve the TypeError when passing a question to a LangChain runnable, ensure that the input you provide to the invoke method is correctly structured. The context should be a string or a list of strings/Documents, and the question should be a string.

Here is an example of how to correctly structure the input:

from langchain_ai21.contextual_answers import AI21ContextualAnswers

# Create an instance of the AI21ContextualAnswers runnable
contextual_answers_runnable = AI21ContextualAnswers()

# Define the input with both context and question
input_data = {
    "context": "This is the context for the question.",
    "question": "What is the context about?"
}

# Invoke the runnable with the input data
try:
    answer = contextual_answers_runnable.invoke(input=input_data)
    print(answer)
except ValueError as e:
    print(f"Error: {e}")

Additionally, ensure that your input data matches the expected input type defined in your chain. For example, if you are using a custom chain, the input type should be defined using a Pydantic model:

from langchain_core.pydantic_v1 import BaseModel
from langchain_core.runnables import RunnablePassthrough

class InputType(BaseModel):
    question: str

chain = RunnablePassthrough().with_types(input_type=InputType)

Ensure that your input data matches this structure. This should help resolve the TypeError you are encountering ^[1][2].

To continue talking to Dosu, mention @dosu.