how to use llama CPP server in llamaindex . I launched openapi compatible llamacpp server.

from llama_index.llms.llama_cpp import LlamaCPP

model_url = "<https://huggingface.co/TheBloke/Llama-2-13B-chat-GGML/resolve/main/llama-2-13b-chat.ggmlv3.q4_0.bin">

llm = LlamaCPP(
    # You can pass in the URL to a GGML model to download it automatically
    model_url=model_url,
    # optionally, you can set the path to a pre-downloaded model instead of model_url
    model_path=None,
    temperature=0.1,
    max_new_tokens=256,
    # llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
    context_window=3900,
    # kwargs to pass to __call__()
    generate_kwargs={},
    # kwargs to pass to __init__()
    # set to at least 1 to use GPU
    model_kwargs={"n_gpu_layers": 1},
    verbose=True,
)

from llama_index.tools.openapi.base import OpenAPIToolSpec
from llama_index.tools.requests.base import RequestsToolSpec
from llama_index.tools.tool_spec.load_and_search.base import LoadAndSearchToolSpec

# Load the OpenAPI spec for your server
open_api_spec = ...  # Load your OpenAPI spec here

open_spec = OpenAPIToolSpec(open_api_spec)

requests_spec = RequestsToolSpec(
    {
        "your-server-url": {
            "Authorization": "Bearer your-key",
            "Content-Type": "application/json",
        }
    }
)

# OpenAPI spec is too large for content, wrap the tool to separate loading and searching
wrapped_tools = LoadAndSearchToolSpec.from_defaults(
    open_spec.to_tool_list()[0],
).to_tool_list()

agent = OpenAIAgent.from_tools(
    [*wrapped_tools, *requests_spec.to_tool_list()], verbose=True
)

from llama_index.llms.openai_like import OpenAILike

llm = OpenAILike(model="my model", api_base="<http://localhost:port",> api_key="fake")

response = llm.complete("Hello World!")
print(str(response))

from llama_index.core.base.query_pipeline.query import BaseComponent

class MyCustomComponent(BaseComponent):
    # Your custom code here
    pass