Ollama Python library 0.4 with function calling improvements

2024-11-24 · Source: Ollama Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

The Ollama Python library, in its 0.4 release on November 25, 2024, introduces significant enhancements, primarily enabling Python functions to be passed directly as tools to Ollama models. This update simplifies the process of integrating custom logic and external library functions, such as `requests.request`, into model interactions. The library now automatically generates JSON schema for functions by parsing Python type annotations and Google-style docstrings, eliminating the need for manual schema definition. Additionally, the 0.4 release includes full typing support across the library for improved developer experience and updated examples on its GitHub repository.

Key takeaway

For AI Engineers building applications with Ollama, this update streamlines the integration of custom logic and external APIs. You can now directly pass Python functions as tools, reducing boilerplate by eliminating manual JSON schema creation. This simplifies complex workflows, allowing your models to interact with external systems more fluidly and efficiently, enhancing the capabilities of your local LLM deployments.

Key insights

Ollama's 0.4 Python library simplifies tool integration by directly accepting Python functions and auto-generating JSON schemas.

Principles

• Automate schema generation from code.
• Integrate external functions as model tools.

Method

Define Python functions with type annotations and docstrings. Pass function references directly to `ollama.chat` via the `tools` field. Execute tool calls from the model's response.

In practice

• Use `pip install -U ollama` to upgrade.
• Annotate function parameters for schema generation.
• Pass `requests.request` for web interactions.

Topics

• Ollama Python Library
• Function Calling
• LLM Tool Use
• Python Type Hinting
• Tool Schema Generation

Code references

• ollama/ollama-python

Best for: Machine Learning Engineer, AI Engineer, Software Engineer

Related on AIssential

• Python & JavaScript Libraries — Ollama's new Python and JavaScript libraries simplify integrating local LLMs into applications, mirroring the REST API.
• Tool support — Ollama now enables LLMs to interact with external tools for complex tasks via a new API feature.
• Gemma 4 Tool Calling Explained: Build AI Agents with Function Calling (Step-by-Step Guide) — Gemma 4's native tool calling enables local AI agents to reliably interact with external APIs for real-world data.
• Using LLM in the shebang line of a script — LLM CLI can execute generative AI tasks and tool calls directly from a script's shebang line.
• Ollama: Unleash the Power of Local LLMs on Your Machine — Ollama enables private, cost-effective, and low-latency local LLM inference and customization via a Python library.
• 10 Python Libraries for Building LLM Applications — Building LLM applications requires specialized Python libraries for control over model loading, data integration, serving, and evaluation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ollama Blog.