Error When Running Ollama LLM Test Script for GPTR

[hayjohnny2000]

Describe the bug
When I run the basic test script at https://docs.gptr.dev/docs/gpt-researcher/llms/running-with-ollama#run-llm-test-script-for-gptr, it produces the following error:

Traceback (most recent call last):
  File "D:\Users\Chickens\Documents\code_projects\gpt-researcher\tests\test-ollama.py", line 14, in <module>
    llm_provider = get_llm(
        "ollama",
    ...<4 lines>...
        verify_ssl=False  # Add this line
    )
  File "D:\ProgramData\.conda_envs\gpt_researcher\Lib\site-packages\gpt_researcher\utils\llm.py", line 19, in get_llm
    return GenericLLMProvider.from_provider(llm_provider, **kwargs)
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\ProgramData\.conda_envs\gpt_researcher\Lib\site-packages\gpt_researcher\llm_provider\generic\base.py", line 77, in from_provider
    llm = ChatOllama(base_url=os.environ["OLLAMA_BASE_URL"], **kwargs)
                              ~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
  File "<frozen os>", line 716, in __getitem__
KeyError: 'OLLAMA_BASE_URL'

I'm running this on Windows 11, with a 3.13.1 Python interpreter in a Conda virtual environment.

To Reproduce
Steps to reproduce the behavior:

1. conda create -p D:\ProgramData\.conda_envs\gpt_researcher
2. conda activate D:\ProgramData\.conda_envs\gpt_researcher
3. conda install conda-forge::pip
4. cd D:\Users\Chickens\Documents\code_projects
5. git clone https://github.com/assafelovic/gpt-researcher.git
6. cd gpt-researcher
7. pip install -r .\requirements.txt
8. Copy the script contents from https://docs.gptr.dev/docs/gpt-researcher/llms/running-with-ollama#run-llm-test-script-for-gptr into ./tests/test-ollama.py
9. python ./tests/test-ollama.py

Expected behavior
The base_url property defined in the test script should be used to initialise the ChatOllama object.

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Windows 11
• Python version: 3.13.1

Additional context
When I specify the os.environ["OLLAMA_BASE_URL"] = OLLAMA_BASE_URL in the test script, it produces another error:

Traceback (most recent call last):
  File "D:\Users\Chickens\Documents\code_projects\gpt-researcher\tests\test-ollama.py", line 14, in <module>
    llm_provider = get_llm(
        "ollama",
    ...<4 lines>...
        verify_ssl=False  # Add this line
    )
  File "D:\ProgramData\.conda_envs\gpt_researcher\Lib\site-packages\gpt_researcher\utils\llm.py", line 19, in get_llm
    return GenericLLMProvider.from_provider(llm_provider, **kwargs)
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\ProgramData\.conda_envs\gpt_researcher\Lib\site-packages\gpt_researcher\llm_provider\generic\base.py", line 77, in from_provider
    llm = ChatOllama(base_url=os.environ["OLLAMA_BASE_URL"], **kwargs)
TypeError: langchain_ollama.chat_models.ChatOllama() got multiple values for keyword argument 'base_url'

[assafelovic]

@hayjohnny2000 does this still occur? @ElishaKay perhaps you can help with this?

[ElishaKay]

Welcome @hayjohnny2000

It's possible there were breaking changes during a recent refactoring.

I tested the Ollama script in the Docs as well recently and didn't succeed in getting it working.

Would love to see a PR if you crack it

[krokosik]

The solution is to simply remove the OLLAMA_BASE_URL variable from the script and function argument and simply use the one from the .env file