156 lines
3.5 KiB
Markdown

# Tu-Zi API Proxy Service Documentation
## Overview
This API provides a queueing proxy service for requests to Tu-Zi services. It allows asynchronous processing of requests with authentication management, concurrency control, and status tracking.
## Base URL
```
/api/tu_zi
```
## Endpoints
### Submit API Request
Submits a request to a Tu-Zi API endpoint and returns a token for tracking.
**Endpoint:** `POST /api/tu_zi`
**Headers:**
- `X-Api-Target` (required): The full URL of the Tu-Zi API endpoint to call
**Request Body:**
- The raw body content to be forwarded to the target API
**Response:**
```json
{
"token": "550e8400-e29b-41d4-a716-446655440000"
}
```
**Example:**
```bash
curl -X POST "https://animator-gg-api.ca2324.servep2p.com:8443/api/tu_zi" \
-H "X-Api-Target: https://api.tu-zi.com/v1/chat/completions" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4-gizmo-g-2fkFE8rbu",
"messages": [
{
"role": "user",
"content": "Who are you? What can you do?"
}
],
"stream": false
}'
```
### Get Request Status
Retrieves the status and result of an API request by token.
**Endpoint:** `GET /api/tu_zi?token={token}`
**Query Parameters:**
- `token` (required): The execution token returned by the POST request
**Response:**
```json
{
"status": "completed",
"result": {
"taskId": "550e8400-e29b-41d4-a716-446655440000",
"success": true,
"response": "{ ... response from Tu-Zi API ... }"
},
"timestamp": 1743991194237
}
```
**Status Values:**
- `pending`: Request is waiting in the queue
- `running`: Request is currently being processed
- `completed`: Request has completed successfully
- `failed`: Request failed to complete
**Example:**
```bash
curl -X GET "https://animator-gg-api.ca2324.servep2p.com:8443/api/tu_zi?token=550e8400-e29b-41d4-a716-446655440000"
```
### Get Queue Statistics
Retrieves information about the current request queue.
**Endpoint:** `GET /api/tu_zi/queue`
**Response:**
```json
{
"size": 3,
"pending": 2,
"isPaused": false,
"activeExecutions": 5
}
```
**Response Fields:**
- `size`: Number of requests in the queue
- `pending`: Number of pending requests
- `isPaused`: Whether the queue is paused
- `activeExecutions`: Total number of active executions (including completed but not yet expired)
**Example:**
```bash
curl -X GET "https://animator-gg-api.ca2324.servep2p.com:8443/api/tu_zi/queue"
```
## Error Handling
The API returns appropriate HTTP status codes:
- `200 OK`: Request processed successfully
- `400 Bad Request`: Missing required parameters
- `500 Internal Server Error`: Server-side error during request execution
Error responses include a JSON body with an error message:
```json
{
"error": "Error message description"
}
```
## Authentication
The service automatically adds appropriate authentication headers to requests based on the target hostname. Currently, the service supports:
- Tu-Zi domains (`*.tu-zi.com`)
## Request Lifecycle
- The service uses a queue to manage concurrent requests (default: 2)
- Each request result is stored for 3 hours after completion
- After this time, the result will be purged, and querying the token will return an "Invalid execution token or result expired" error
## Client Usage Pattern
1. Submit a request to the Tu-Zi API through the proxy
2. Receive a token immediately
3. Poll the status endpoint with the token until the status is "completed" or "failed"
4. Process the response or handle the error