Knowledge configuration throwing error missing 2 required keyword-only arguments: 'response' and 'body' . Using azure openai and azure embedding-ada-2

CrewAI Community Support
1

Hi,
I’m using Azure OpenAI and Azure text-embedding-ada-2 models.
Below is the code

import os;
from crewai import Agent, Task, Crew, Process, LLM
from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource

Create a knowledge source

string_source = StringKnowledgeSource(
content=“Users name is John. He is 30 years old and lives in San Francisco.”
)

os.environ[‘AZURE_OPENAI_API_KEY’] =“azure_open_ai_key”
os.environ[‘AZURE_OPENAI_ENDPOINT’] =“azure open ai endpoint”
os.environ[‘AZURE_API_VERSION’] = “2024-08-01-preview”

Create an LLM with a temperature of 0 to ensure deterministic outputs

llm = LLM(
model=“azure/gpt-4o-mini”,
api_key=‘azure_open_ai_key’,
base_url=‘aisprdoai.openai.azure.com/’,
api_version=‘2024-08-01-preview’
)

embedder={
“provider”: “azure”,
“config”: {
“api_key”: “azure_open_ai_key”,
“api_base”: “aisprdoai.openai.azure.com”,
“api_version”: “2023-05-15”,
“model_name”: “aisprdoai-embedding-ada-2”
}
}

Create an agent with the knowledge store

agent = Agent(
role=“About User”,
goal=“You know everything about the user and provide the information from source only.”,
backstory=“”“You are a master at understanding people and their preferences.”“”,
verbose=True,
allow_delegation=False,
llm=llm,
knowledge_sources=[string_source],
embedder=embedder
)
task = Task(
description=“Answer the following questions about the user: {question}”,
expected_output=“An answer to the question.”,
agent=agent,
verbose=True
)

crew = Crew(
agents=[agent],
tasks=[task],
verbose=True,
process=Process.sequential,
)

result = crew.kickoff(inputs={“question”: “Who is john and how old is he?”})

when i run this getting below error.
File “C:\Users\xxxxx\latest_ai_development\src\latest_ai_development\knowledge1base.py”, line 54, in
agent = Agent(
File “C:\Users\xxxx\anaconda3\envs\crewai-env\lib\site-packages\pydantic\main.py”, line 214, in init
validated_self = self.pydantic_validator.validate_python(data, self_instance=self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for Agent
Value error, Invalid Knowledge Configuration: APIStatusError.init() missing 2 required keyword-only arguments: ‘response’ and ‘body’ [type=value_error, input_value={‘role’: ‘About User’, ‘g…evai-embedding-ada-2’}}}, input_type=dict]
For further information visit errors.pydantic.dev/2.10/v/value_error

2

I have the same issue. Does anyone have any idea how to address this issue?

3

I’m facing the same issue. Does anyone have an idea on how to solve this?

4

crewai tools doesn’t work with third party llms, I’ve tried to get it to work a lot and never succeeded. The workaround is to use ChatOpenAI which acts as a wrapper around your llm to expose it as a open ai llm to the tool, but this also doesnt work well as the tool needs openai models to work.

5

knowledge is like a tool in crewai which is heavily dependent on open ai llms, so if you can get an api key directly, It will work

6

As with any problem (whether it’s a cake recipe or building a supersonic jet), let’s start with a small enough example to make understanding and troubleshooting easier, just like this code here in the CrewAI documentation.

But before the code, a few points, @Suneel_Kumar:

  • In your embedder["config"], where it says model_name, I believe it should be model.
  • In your embedder["config"], where it says api_base, I believe it should be base_url.
  • Considering Azure’s documentation, your AZURE_API_BASE env variable should have the full format https://<SOMETHING-HERE>.openai.azure.com/.

For the small example below, I used OpenRouter as the provider, Qwen as the LLM, and Google as the embedder, to evaluate if “crewai is heavily dependent on open ai llms.” I ran it (version 0.105.0) and everything went very well. So, I suggest you first adapt this functional example, ensure it’s not a configuration error, and if the error persists, provide the complete error message so someone can effectively help.

First things first, make sure to have a clean knowledge db:

$ crewai reset-memories --knowledge

Hope the code helps:

from crewai import Agent, Task, Crew, Process, LLM
from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource
import os

from crewai.utilities.paths import db_storage_path
from pathlib import Path

print(
    '! Knowledge Source location is: '
    f'{str(Path(db_storage_path()) / "knowledge")}'
)

# My env:
os.environ["OPENROUTER_API_KEY"] = ""
os.environ["GEMINI_API_KEY"] = ""
# Your env:
os.environ["AZURE_API_KEY"] = ""  # "your-azure-api-key"
os.environ["AZURE_API_BASE"] = ""  # "https://example-endpoint.openai.azure.com"
os.environ["AZURE_API_VERSION"] = ""  # "2029-01-01"

content = "Users name is John. He is 108 years old and lives in Alabama."
string_source = StringKnowledgeSource(
    content=content,
)

openrouter_llm = LLM(
    model='openrouter/qwen/qwq-32b:free',
    api_key=os.environ["OPENROUTER_API_KEY"],
    temperature=0
)

google_embedder = {
    "provider": "google",
    "config": {
        "model": "models/text-embedding-004",
        "api_key": os.environ["GEMINI_API_KEY"]
    }
}

agent = Agent(
    role="About User",
    goal="You know everything about the user.",
    backstory="You are a master at understanding people and their preferences.",
    verbose=True,
    allow_delegation=False,
    llm=openrouter_llm
)

task = Task(
    description="Answer the following questions about the user: {question}",
    expected_output="An answer to the question.",
    agent=agent,
)

crew = Crew(
    agents=[agent],
    tasks=[task],
    verbose=True,
    process=Process.sequential,
    knowledge_sources=[string_source],
    embedder=google_embedder
)

result = crew.kickoff(
    inputs={
        "question": "Where does John live and how old is he?"
    }
)