A unified chat completions gateway that now forwards requests to the internal new-api service (default http://localhost:3000) while preserving the same client-facing schema.
An asynchronous data import analysis pipeline that orchestrates LLM calls to produce structured metadata and processing recommendations

The following instructions cover environment setup, dependency installation, and running the backend service.

Prerequisites

Python 3.11 (recommended) or newer
Git
uv package manager (used for Python dependency management)

Install uv

# Linux / macOS
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows (PowerShell)
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

After installation, ensure uv is on your PATH:

uv --version

Install Python Dependencies

Create (or activate) a virtual environment, then install project dependencies with uv:

# Create a virtualenv named .venv if it doesn't exist
uv venv .venv

# Activate the virtualenv (Linux/macOS)
source .venv/bin/activate

# On Windows PowerShell:
# .\.venv\Scripts\Activate.ps1

# Install dependencies from requirements.txt
uv pip install -r requirements.txt

If you prefer native pip, replace the last command with pip install -r requirements.txt.

Environment Variables

Copy .env.example to .env (if provided) or edit .env to supply API keys and configuration values:

OPENAI_API_KEY, ANTHROPIC_API_KEY, OPENROUTER_API_KEY, etc.
HTTP_CLIENT_TIMEOUT, IMPORT_CHAT_TIMEOUT_SECONDS
LOG_LEVEL, LOG_FORMAT for logging
NEW_API_BASE_URL (defaults to http://localhost:3000) and optional NEW_API_AUTH_TOKEN if the new-api component enforces authentication.

Run the Backend Service

Start the FastAPI application using uvicorn:

uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

--reload enables auto-restart during development.
Access the interactive API docs at http://127.0.0.1:8000/docs.

To keep it running in the background (Unix-like systems):

nohup uvicorn app.main:app --host 0.0.0.0 --port 8000 > server.log 2>&1 &

Or use a process manager such as pm2, supervisor, or systemd for production deployments.

API List

导入分析schema接口 http://localhost:8000/v1/import/analyze

Additional Commands

Run the data import analysis example: python test/data_import_analysis_example.py
Test the OpenRouter demo: python test/openrouter_chat_example.py
Send a DeepSeek chat request script: python scripts/deepseek_request.py