batch_processing/
batch_processing.rs

1use ai_lib::types::common::Content;
2use ai_lib::{AiClient, ChatCompletionRequest, Message, Provider, Role};
3
4#[tokio::main]
5async fn main() -> Result<(), Box<dyn std::error::Error>> {
6    println!("🚀 AI-lib Batch Processing Example");
7    println!("==================================");
8
9    // 创建客户端
10    let client = AiClient::new(Provider::Groq)?;
11    println!(
12        "✅ Created client with provider: {:?}",
13        client.current_provider()
14    );
15
16    // 准备多个请求
17    let requests = vec![
18        ChatCompletionRequest::new(
19            "llama3-8b-8192".to_string(),
20            vec![Message {
21                role: Role::User,
22                content: Content::Text("What is the capital of France?".to_string()),
23                function_call: None,
24            }],
25        )
26        .with_temperature(0.7)
27        .with_max_tokens(50),
28
29        ChatCompletionRequest::new(
30            "llama3-8b-8192".to_string(),
31            vec![Message {
32                role: Role::User,
33                content: Content::Text("What is 2 + 2?".to_string()),
34                function_call: None,
35            }],
36        )
37        .with_temperature(0.1)
38        .with_max_tokens(20),
39
40        ChatCompletionRequest::new(
41            "llama3-8b-8192".to_string(),
42            vec![Message {
43                role: Role::User,
44                content: Content::Text("Tell me a short joke.".to_string()),
45                function_call: None,
46            }],
47        )
48        .with_temperature(0.9)
49        .with_max_tokens(100),
50
51        ChatCompletionRequest::new(
52            "llama3-8b-8192".to_string(),
53            vec![Message {
54                role: Role::User,
55                content: Content::Text("What is the largest planet in our solar system?".to_string()),
56                function_call: None,
57            }],
58        )
59        .with_temperature(0.5)
60        .with_max_tokens(60),
61    ];
62
63    println!("📤 Prepared {} requests for batch processing", requests.len());
64
65    // 方法1: 使用并发限制的批处理
66    println!("\n🔄 Method 1: Batch processing with concurrency limit (2)");
67    let start_time = std::time::Instant::now();
68    
69    let responses = client.chat_completion_batch(requests.clone(), Some(2)).await?;
70    
71    let duration = start_time.elapsed();
72    println!("⏱️  Batch processing completed in {:?}", duration);
73
74    // 处理响应
75    for (i, response) in responses.iter().enumerate() {
76        match response {
77            Ok(resp) => {
78                println!(
79                    "✅ Request {}: {}",
80                    i + 1,
81                    resp.choices[0].message.content.as_text()
82                );
83            }
84            Err(e) => {
85                println!("❌ Request {} failed: {}", i + 1, e);
86            }
87        }
88    }
89
90    // 方法2: 使用智能批处理（自动选择策略）
91    println!("\n🧠 Method 2: Smart batch processing");
92    let start_time = std::time::Instant::now();
93    
94    let responses = client.chat_completion_batch_smart(requests.clone()).await?;
95    
96    let duration = start_time.elapsed();
97    println!("⏱️  Smart batch processing completed in {:?}", duration);
98
99    // 统计成功和失败
100    let successful: Vec<_> = responses.iter().filter_map(|r| r.as_ref().ok()).collect();
101    let failed: Vec<_> = responses.iter().enumerate().filter_map(|(i, r)| {
102        r.as_ref().err().map(|e| (i, e))
103    }).collect();
104
105    println!("📊 Results:");
106    println!("   ✅ Successful: {}/{}", successful.len(), responses.len());
107    println!("   ❌ Failed: {}/{}", failed.len(), responses.len());
108    println!("   📈 Success rate: {:.1}%", (successful.len() as f64 / responses.len() as f64) * 100.0);
109
110    // 方法3: 无限制并发批处理
111    println!("\n🚀 Method 3: Unlimited concurrent batch processing");
112    let start_time = std::time::Instant::now();
113    
114    let responses = client.chat_completion_batch(requests, None).await?;
115    
116    let duration = start_time.elapsed();
117    println!("⏱️  Unlimited concurrent processing completed in {:?}", duration);
118
119    // 显示所有响应
120    for (i, response) in responses.iter().enumerate() {
121        match response {
122            Ok(resp) => {
123                println!(
124                    "✅ Request {}: {}",
125                    i + 1,
126                    resp.choices[0].message.content.as_text()
127                );
128            }
129            Err(e) => {
130                println!("❌ Request {} failed: {}", i + 1, e);
131            }
132        }
133    }
134
135    println!("\n🎉 Batch processing example completed successfully!");
136    Ok(())
137}
batch_processing/batch_processing.rs

batch_processing/
batch_processing.rs
