Files

2025-08-20 09:32:35 +02:00

2.6 KiB

Raw Permalink Blame History

🧪 Mock OpenAI Chat API Server

This project simulates the OpenAI /v1/chat/completions API using a local FastAPI server.
It logs all incoming requests and returns a dummy response using "Lorem Ipsum..." content for debugging and development purposes.

✅ Features

Local server that mimics OpenAI ChatCompletion endpoint
Logs full request (headers + body) to requests_log.json
Returns a dummy OpenAI-style response
Plug-in replacement for openai.ChatCompletion.create(...)

⚙️ Requirements

Python 3.7+
FastAPI
Uvicorn

📦 Installation & Setup

1. Clone or copy this project into a local folder

2. Create and activate a virtual environment

python -m venv venv

Then activate it:

Windows:

venv\Scripts\activate

macOS/Linux:

source venv/bin/activate

3. Install dependencies

pip install fastapi uvicorn

🚀 How to Start the Server

Use this command to launch the mock API server:

python -m uvicorn mock_openai_server:app --reload --port 8000

It will be available at:

http://localhost:8000/v1/chat/completions

🧪 How to Use with OpenAI Python SDK

Update your client to point to the mock server:

import openai

openai.api_key = "sk-dummy"
openai.api_base = "http://localhost:8000"

response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[
        {"role": "user", "content": "Hello!"}
    ]
)

print(response)

📄 Example Dummy Response

ID:            chatcmpl-mock123
Model:         gpt-4
Created:       <current timestamp>
Role:          assistant
Message:       Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Finish reason: stop
Tokens used:   prompt=10, completion=12, total=22

📝 Output Logs

All requests and responses are logged to requests_log.json.
Each log entry contains:

Timestamp
Headers (sent by client)
JSON body (messages, model, etc.)
Dummy response

You can use this log to debug what your app sends to OpenAI.

📂 Project Structure

.
├── mock_openai_server.py    # FastAPI mock server
├── requests_log.json        # Saved request/response logs
└── README.md                # Documentation

🧰 Optional: Save Installed Packages

To track your environment:

pip freeze > requirements.txt

Later, you can restore it with:

pip install -r requirements.txt

📌 Notes

No actual API calls are made to OpenAI
Great for debugging payload formats, headers, and SDK integrations
For local development only

🪪 License

MIT License

2.6 KiB Raw Permalink Blame History