OPENAI_API_KEY missing error when running the End-to-end example from Azure OpenAI docs

[yashworlikar]

I encountered an error while running the End-to-end example from the Azure OpenAI documentation. The error indicates that the OPENAI_API_KEY is not set .

Error Message:

Traceback (most recent call last):
  File "C:\Applications\Application\autoLLM\demo.py", line 21, in <module>
    world_model = WorldModel()
                  ^^^^^^^^^^^^
  File "C:\Applications\Application\autoLLM\venv\Lib\site-packages\lavague\core\world_model.py", line 377, in __init__
    mm_llm = get_default_context().mm_llm
             ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Applications\Application\autoLLM\venv\Lib\site-packages\lavague\core\context.py", line 35, in get_default_context
    return OpenaiContext()
           ^^^^^^^^^^^^^^^
  File "C:\Applications\Application\autoLLM\venv\Lib\site-packages\lavague\contexts\openai\base.py", line 23, in __init__
    raise ValueError("OPENAI_API_KEY is not set")
ValueError: OPENAI_API_KEY is not set

Steps to Reproduce:

1. Follow the End-to-end example from the Azure OpenAI documentation.
2. Fill in the required Azure configs
3. Run the provided code snippet.

Code Snippet:

from lavague.core import WorldModel, ActionEngine
from lavague.core.agents import WebAgent
from lavague.drivers.selenium import SeleniumDriver
from lavague.contexts.openai import AzureOpenaiContext

# Initialize context with our custom elements
context = AzureOpenaiContext(
    api_key="your_api_key",
    deployment="deployment_name",
    llm="<model_name>",
    mm_llm="<model_name>",
    endpoint="<your_endpoint>",
    embedding="embedding_model_name",
    embedding_deployment="embedding_deployment_name"
)

selenium_driver = SeleniumDriver()

# Build Action Engine and World Model from Context
action_engine = ActionEngine.from_context(context=context, driver=selenium_driver)
world_model = WorldModel()

# Build agent & run query
agent = WebAgent(world_model, action_engine)
agent.get("https://huggingface.co/")
agent.run("What is this week's top Space of the week?")

Expected Behavior: The code should run without errors and perform the intended actions.

Actual Behavior: The code raises a ValueError indicating that the OPENAI_API_KEY is not set.

Passing the mm_llm from the context to the WorldModel when creating an instance of it fixes it.

world_model = WorldModel(mm_llm=context.mm_llm) 

[hi0001234d]

have you tried "export OPENAI_API_KEY=[your open ai key]" in the terminal

just wondering if that's the case