How can I use llamacppapi (llama_cpp.server)

[vmajor]

I can select it in agent settings and give it the localhost URL where the server is waiting for queries, but my agent never makes a query and instead times out with:

agixt-backend-streamlit-1  | 2023-05-26 15:26:54.896 Uncaught app exception
agixt-backend-streamlit-1  | Traceback (most recent call last):
agixt-backend-streamlit-1  |   File "/usr/local/lib/python3.10/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 565, in _run_script
agixt-backend-streamlit-1  |     exec(code, module.__dict__)
agixt-backend-streamlit-1  |   File "/agixt/pages/3-Instructions.py", line 66, in <module>
agixt-backend-streamlit-1  |     response = agent.smart_instruct(
agixt-backend-streamlit-1  |   File "/agixt/AGiXT.py", line 218, in smart_instruct
agixt-backend-streamlit-1  |     self.run(
agixt-backend-streamlit-1  |   File "/agixt/AGiXT.py", line 142, in run
agixt-backend-streamlit-1  |     loop.run_until_complete(
agixt-backend-streamlit-1  |   File "/usr/local/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
agixt-backend-streamlit-1  |     return future.result()
agixt-backend-streamlit-1  |   File "/agixt/AGiXT.py", line 495, in websearch_agent
agixt-backend-streamlit-1  |     results = results.split("\n")
agixt-backend-streamlit-1  | AttributeError: 'NoneType' object has no attribute 'split' 

Ah right... I think I know what is happening. I am running AGiXT in docker and it has its own network so http://localhost:8000 cannot work since nothing is in fact running inside the docker image's localhost - at least I think that's what the problem is.

Similarly, using local models from the docker instance ought to be impossible or at least difficult to get to work as I would need to either copy the model to docker storage or mount local storage in docker and thus eliminating the protection that docker provides.