lsimons-llm

Shared LLM client library for OpenAI-compatible APIs.

Installation

As a dependency in other projects

Add to your pyproject.toml:

dependencies = [
    "lsimons-llm @ git+https://github.com/lsimons-bot/lsimons-llm.git",
]

# For async support
dependencies = [
    "lsimons-llm[async] @ git+https://github.com/lsimons-bot/lsimons-llm.git",
]

Local development setup

When developing lsimons-llm alongside dependent projects, use editable installs to see changes immediately:

# In each dependent project directory (lsimons-bot, quarto4sbp, lsimons-agent)
uv pip install -e ../lsimons-llm

This overrides the git dependency with your local copy. Changes to lsimons-llm are available without pushing.

Projects using this library:

• lsimons-bot - Slack bot (async client)
• quarto4sbp - Quarto document tool (sync client)
• lsimons-agent - AI coding agent (sync client)

Configuration

Set environment variables:

Variable	Default	Description
LLM_API_KEY	(required)	API key for authentication
LLM_BASE_URL	https://litellm.sbp.ai/v1	API endpoint
LLM_MODEL	azure/gpt-4o-mini	Model name
LLM_MAX_TOKENS	4096	Maximum tokens per request
LLM_TEMPERATURE	0.7	Sampling temperature
LLM_TIMEOUT	120	Request timeout (seconds)
LLM_MAX_RETRIES	3	Maximum retry attempts

Usage

Simple usage

from lsimons_llm import chat

response = chat([{"role": "user", "content": "Hello!"}])
print(response)

With client instance

from lsimons_llm import LLMClient, load_config

config = load_config()
with LLMClient(config) as client:
    response = client.chat([
        {"role": "system", "content": "You are helpful."},
        {"role": "user", "content": "Hello!"}
    ])
    print(response)

Async client

from lsimons_llm.async_client import AsyncLLMClient
from lsimons_llm import load_config

config = load_config()
async with AsyncLLMClient(config) as client:
    response = await client.chat([{"role": "user", "content": "Hello!"}])
    print(response)

Configuration override

from lsimons_llm import LLMClient, load_config

# Override via load_config
config = load_config(model="gpt-4", temperature=0.9)

# Or override per-request
client = LLMClient(config)
response = client.chat(
    messages=[{"role": "user", "content": "Hello!"}],
    model="gpt-4-turbo",
    max_tokens=1000,
)

Tool calling

from lsimons_llm import LLMClient, load_config

config = load_config()
client = LLMClient(config)

tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get weather for a location",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {"type": "string"}
            }
        }
    }
}]

# Use chat_raw to get full response including tool calls
response = client.chat_raw(
    messages=[{"role": "user", "content": "What's the weather in Amsterdam?"}],
    tools=tools,
)

Development

# Setup
uv venv && uv sync --all-groups

# Run tests
uv run pytest

# Lint and format
uv run ruff check .
uv run ruff format .

# Type check
uv run basedpyright

License

See LICENSE.md.