TestOllamaAPI Workspace

Project Description

TestOllamaAPI is an exploratory, full-stack workspace designed to build and test local LLM chat applications using Ollama. It provides a unified React frontend that can interface with a choice of two fully featured backend APIs: one written in Python (FastAPI) and the other in C# (ASP.NET Core).

The primary goals of this project are to demonstrate how to:

• Seamlessly connect to a local Ollama service to generate text and handle chat streams.
• Manage conversational state and chat history.
• Expose a robust REST API for LLM interactions.
• Build a responsive web UI to chat with local AI models.

The workspace is divided into three main components: a frontend application built with React, and two different backend implementations for the API layer. This allows developers to experiment with or learn from different tech stacks side-by-side.

Project Structure

• chat-app/: The frontend React web application.
• TestOllamaAPI/: The C# ASP.NET Core API implementation.
• ollama-api-python/: The Python FastAPI implementation.

---

1. React Web Application (chat-app/)

A modern web client to interface with the Ollama chat APIs.

• Stack: React 18, Vite, TypeScript
• Setup & Run:

  cd chat-app
  npm install
  npm run dev

• Default URL: Usually http://localhost:5173 (Vite's default)

---

2. Python FastAPI (ollama-api-python/)

A lightweight, fast, and highly concurrent Python backend for managing chat conversations and interfacing with the Ollama service.

• Stack: Python 3.x, FastAPI, Uvicorn, SQLite
• Features:
  • Conversation history tracking using SQLite
  • Cross-Origin Resource Sharing (CORS) configured for the web frontend
• Setup & Run:

  cd ollama-api-python
  python -m venv venv
  # Activate venv:
  # Windows: venv\Scripts\activate
  # Mac/Linux: source venv/bin/activate
  pip install -r requirements.txt
  python main.py

• Default URL: http://localhost:3003
• API Documentation: http://localhost:3003/docs (Swagger UI)

---

3. ASP.NET Core API (TestOllamaAPI/)

An alternative backend implementation built with the robust Microsoft .NET ecosystem.

• Stack: C#, .NET, ASP.NET Core, Scalar (for API documentation)
• Features:
  • Strongly-typed services (IOllamaService, IConversationService)
  • Configured CORS for frontend access
• Setup & Run:

  cd TestOllamaAPI
  dotnet build
  dotnet run

• Default URLs:
  • HTTPS: https://localhost:7201
  • HTTP: http://localhost:5220
• API Documentation: https://localhost:7201/scalar/v1 (Scalar API Reference)

---

Getting Started

1. Start a Backend: Choose either the Python backend or the ASP.NET backend and follow its setup instructions to start the API server. Ensure Ollama is running on your machine.
2. Start the Frontend: Navigate to the chat-app directory, install the dependencies, and start the development server.
3. Chat: Open your browser to the Vite dev server URL and start interacting with your local Ollama models!