batch_processing/
batch_processing.rs

1/// AI-lib 批处理示例 - AI-lib batch processing example
2use ai_lib::types::common::Content;
3use ai_lib::{AiClient, ChatCompletionRequest, Message, Provider, Role};
4
5#[tokio::main]
6async fn main() -> Result<(), Box<dyn std::error::Error>> {
7    println!("🚀 AI-lib Batch Processing Example");
8    println!("==================================");
9
10    // Create client
11    let client = AiClient::new(Provider::Groq)?;
12    println!(
13        "✅ Created client with provider: {:?}",
14        client.current_provider()
15    );
16
17    // Prepare multiple requests
18    let requests = vec![
19        ChatCompletionRequest::new(
20            "llama3-8b-8192".to_string(),
21            vec![Message {
22                role: Role::User,
23                content: Content::Text("What is the capital of France?".to_string()),
24                function_call: None,
25            }],
26        )
27        .with_temperature(0.7)
28        .with_max_tokens(50),
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        ChatCompletionRequest::new(
40            "llama3-8b-8192".to_string(),
41            vec![Message {
42                role: Role::User,
43                content: Content::Text("Tell me a short joke.".to_string()),
44                function_call: None,
45            }],
46        )
47        .with_temperature(0.9)
48        .with_max_tokens(100),
49        ChatCompletionRequest::new(
50            "llama3-8b-8192".to_string(),
51            vec![Message {
52                role: Role::User,
53                content: Content::Text(
54                    "What is the largest planet in our solar system?".to_string(),
55                ),
56                function_call: None,
57            }],
58        )
59        .with_temperature(0.5)
60        .with_max_tokens(60),
61    ];
62
63    println!(
64        "📤 Prepared {} requests for batch processing",
65        requests.len()
66    );
67
68    // Method 1: Batch processing with concurrency limit
69    println!("\n🔄 Method 1: Batch processing with concurrency limit (2)");
70    let start_time = std::time::Instant::now();
71
72    let responses = client
73        .chat_completion_batch(requests.clone(), Some(2))
74        .await?;
75
76    let duration = start_time.elapsed();
77    println!("⏱️  Batch processing completed in {:?}", duration);
78
79    // Process responses
80    for (i, response) in responses.iter().enumerate() {
81        match response {
82            Ok(resp) => {
83                println!(
84                    "✅ Request {}: {}",
85                    i + 1,
86                    resp.choices[0].message.content.as_text()
87                );
88            }
89            Err(e) => {
90                println!("❌ Request {} failed: {}", i + 1, e);
91            }
92        }
93    }
94
95    // Method 2: Smart batch processing (auto-select strategy)
96    println!("\n🧠 Method 2: Smart batch processing");
97    let start_time = std::time::Instant::now();
98
99    let responses = client.chat_completion_batch_smart(requests.clone()).await?;
100
101    let duration = start_time.elapsed();
102    println!("⏱️  Smart batch processing completed in {:?}", duration);
103
104    // Count successes and failures
105    let successful: Vec<_> = responses.iter().filter_map(|r| r.as_ref().ok()).collect();
106    let failed: Vec<_> = responses
107        .iter()
108        .enumerate()
109        .filter_map(|(i, r)| r.as_ref().err().map(|e| (i, e)))
110        .collect();
111
112    println!("📊 Results:");
113    println!("   ✅ Successful: {}/{}", successful.len(), responses.len());
114    println!("   ❌ Failed: {}/{}", failed.len(), responses.len());
115    println!(
116        "   📈 Success rate: {:.1}%",
117        (successful.len() as f64 / responses.len() as f64) * 100.0
118    );
119
120    // Method 3: Unlimited concurrent batch processing
121    println!("\n🚀 Method 3: Unlimited concurrent batch processing");
122    let start_time = std::time::Instant::now();
123
124    let responses = client.chat_completion_batch(requests, None).await?;
125
126    let duration = start_time.elapsed();
127    println!(
128        "⏱️  Unlimited concurrent processing completed in {:?}",
129        duration
130    );
131
132    // Display all responses
133    for (i, response) in responses.iter().enumerate() {
134        match response {
135            Ok(resp) => {
136                println!(
137                    "✅ Request {}: {}",
138                    i + 1,
139                    resp.choices[0].message.content.as_text()
140                );
141            }
142            Err(e) => {
143                println!("❌ Request {} failed: {}", i + 1, e);
144            }
145        }
146    }
147
148    println!("\n🎉 Batch processing example completed successfully!");
149    Ok(())
150}
batch_processing/batch_processing.rs

batch_processing/
batch_processing.rs
