swan-macro 0.2.4

A declarative HTTP client library for Rust
Documentation

Swan Macro

Crates.io Documentation License: MIT

🌏 Languages: English | δΈ­ζ–‡

Swan Macro is the procedural macro component of the Swan HTTP library, providing declarative HTTP client definition syntax.

🌟 Core Features

  • Declarative Client Definition: Define HTTP clients and methods using macro annotations
  • Automatic Code Generation: Generate high-performance HTTP client code at compile time
  • Smart Retry Mechanism: Method-level progressive exponential backoff retry
  • Interceptor Integration: Seamless integration of global and method-level interceptors
  • Dynamic Parameter Support: Parameter placeholders in URLs and headers
  • State Injection: Axum-like application state management

πŸ“¦ Installation

Add the following to your Cargo.toml:

[dependencies]
swan-macro = "0.2"
swan-common = "0.2"  # Required runtime dependency
serde = { version = "1.0", features = ["derive"] }
anyhow = "1.0"
tokio = { version = "1.0", features = ["macros", "rt-multi-thread"] }

πŸš€ Quick Start

Basic Usage

use serde::{Deserialize, Serialize};
use swan_macro::{http_client, get, post};

#[derive(Deserialize)]
struct User {
    id: u32,
    name: String,
    email: String,
}

#[derive(Serialize)]
struct CreateUserRequest {
    name: String,
    email: String,
}

// Define HTTP client
#[http_client(base_url = "https://api.example.com")]
struct ApiClient;

impl ApiClient {
    // GET request
    #[get(url = "/users/{id}")]
    async fn get_user(&self, id: u32) -> anyhow::Result<User> {}

    // POST request
    #[post(url = "/users", content_type = json)]
    async fn create_user(&self, body: CreateUserRequest) -> anyhow::Result<User> {}
    
    // Request with retry
    #[get(url = "/users/{id}", retry = "exponential(3, 100ms)")]
    async fn get_user_with_retry(&self, id: u32) -> anyhow::Result<User> {}
}

#[tokio::main]
async fn main() -> anyhow::Result<()> {
    let client = ApiClient::new();
    
    // Use client
    let user = client.get_user(1).await?;
    println!("User: {}", user.name);
    
    Ok(())
}

πŸ”§ Supported Macros

#[http_client]

Define HTTP client struct:

#[http_client(
    base_url = "https://api.example.com",
    interceptor = MyInterceptor,  // Optional: global interceptor
    state = AppState              // Optional: application state type
)]
struct ApiClient;

HTTP Method Macros

Supported HTTP methods:

  • #[get(url = "...")] - GET requests
  • #[post(url = "...", content_type = json)] - POST requests
  • #[put(url = "...", content_type = json)] - PUT requests
  • #[delete(url = "...")] - DELETE requests

Method Parameters

impl ApiClient {
    #[get(
        url = "/users/{id}",                    // Path parameters
        header = "Authorization: Bearer {token}", // Dynamic headers
        retry = "exponential(3, 100ms)",        // Retry strategy
        interceptor = MethodLevelInterceptor    // Method-level interceptor
    )]
    async fn get_user(&self, id: u32, token: String) -> anyhow::Result<User> {}
}

πŸ”„ Retry Mechanism

Retry Strategy Types

// Exponential backoff retry
#[get(url = "/api", retry = "exponential(3, 100ms)")]
#[get(url = "/api", retry = "exponential(
    max_attempts=5,
    base_delay=200ms,
    max_delay=30s,
    exponential_base=2.0,
    jitter_ratio=0.1
)")]

// Fixed delay retry
#[get(url = "/api", retry = "fixed(max_attempts=3, delay=1s)")]

Automatic Retry Conditions

  • 5xx Server Errors (500-599)
  • 429 Too Many Requests (rate limiting)
  • 408 Request Timeout (timeout)
  • Network Connection Errors

Idempotency Protection

By default, only safe HTTP methods are retried:

#[get(url = "/data")]     // βœ… Auto retry
#[put(url = "/data")]     // βœ… Auto retry  
#[delete(url = "/data")]  // βœ… Auto retry
#[post(url = "/data")]    // ❌ No retry by default (non-idempotent)

// Force retry for non-idempotent methods (use with caution)
#[post(url = "/idempotent", retry = "exponential(
    max_attempts=3,
    base_delay=100ms,
    idempotent_only=false
)")]

🌐 Dynamic Parameters

URL Parameters

// Path parameters
#[get(url = "/users/{user_id}/posts/{post_id}")]
async fn get_user_post(&self, user_id: u32, post_id: u32) -> anyhow::Result<Post> {}

// Query parameters
#[get(url = "/search?q={query}&page={page}")]
async fn search(&self, query: String, page: u32) -> anyhow::Result<Vec<Post>> {}

// Positional parameter reference
#[get(url = "/posts?author={param0}&category={param1}")]
async fn search_by_position(&self, author: String, category: String) -> anyhow::Result<Vec<Post>> {}

Dynamic Headers

#[get(
    url = "/protected",
    header = "Authorization: Bearer {token}",
    header = "X-User-ID: {user_id}"
)]
async fn get_protected_data(&self, token: String, user_id: u32) -> anyhow::Result<Data> {}

πŸ”Œ Interceptor Integration

use async_trait::async_trait;
use swan_common::SwanInterceptor;

#[derive(Default)]
struct AuthInterceptor;

#[async_trait]
impl SwanInterceptor for AuthInterceptor {
    async fn before_request<'a>(
        &self,
        request: reqwest::RequestBuilder,
        request_body: &'a [u8],
        _context: Option<&(dyn std::any::Any + Send + Sync)>,
    ) -> anyhow::Result<(reqwest::RequestBuilder, std::borrow::Cow<'a, [u8]>)> {
        let request = request.header("Authorization", "Bearer token");
        Ok((request, std::borrow::Cow::Borrowed(request_body)))
    }

    async fn after_response(
        &self,
        response: reqwest::Response,
        _context: Option<&(dyn std::any::Any + Send + Sync)>,
    ) -> anyhow::Result<reqwest::Response> {
        println!("Response status: {}", response.status());
        Ok(response)
    }
}

// Use interceptor
#[http_client(base_url = "https://api.example.com", interceptor = AuthInterceptor)]
struct SecureApiClient;

🏷️ Content Types

Supported content types:

  • content_type = json - application/json
  • content_type = form_urlencoded - application/x-www-form-urlencoded
  • content_type = form_multipart - multipart/form-data

⚑ Compile-Time Optimization

Swan Macro generates highly optimized code at compile time:

  • Zero Runtime Overhead: All configuration determined at compile time
  • Inline Optimization: Automatically inline small function calls
  • Conditional Compilation: Remove debug code in release mode
  • Smart Caching: Interceptor instance reuse

πŸ§ͺ Testing

Run tests:

cargo test --lib

πŸ“– Documentation

Detailed API documentation:

cargo doc --open

🀝 Use with Swan Common

Swan Macro depends on Swan Common for runtime support:

[dependencies]
swan-macro = "0.2"
swan-common = "0.2"

πŸ“„ License

This project is licensed under the MIT License. See the LICENSE file for details.