flyllm 0.4.1

A Rust library for unifying LLM backends as an abstraction layer with load balancing.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198

//! Tests for load balancing strategies.

use flyllm::{LlmManager, ProviderType, TaskDefinition};
use flyllm::load_balancer::strategies::{
    LoadBalancingStrategy,
    LeastRecentlyUsedStrategy,
    LowestLatencyStrategy,
    RandomStrategy
};

// ============================================================================
// Strategy Creation Tests
// ============================================================================

#[test]
fn test_lru_strategy_creation() {
    // Just verify that creation doesn't panic
    let _strategy = LeastRecentlyUsedStrategy::new();
}

#[test]
fn test_lowest_latency_strategy_creation() {
    let _strategy = LowestLatencyStrategy::new();
}

#[test]
fn test_random_strategy_creation() {
    let _strategy = RandomStrategy::new();
}

// ============================================================================
// Builder Strategy Configuration Tests
// ============================================================================

#[tokio::test]
async fn test_builder_with_lru_strategy() {
    let manager = LlmManager::builder()
        .define_task(TaskDefinition::new("test"))
        .add_instance(ProviderType::OpenAI, "gpt-4", "key")
        .supports("test")
        .strategy(Box::new(LeastRecentlyUsedStrategy::new()))
        .build()
        .await
        .unwrap();

    assert_eq!(manager.get_provider_count().await, 1);
}

#[tokio::test]
async fn test_builder_with_lowest_latency_strategy() {
    let manager = LlmManager::builder()
        .define_task(TaskDefinition::new("test"))
        .add_instance(ProviderType::OpenAI, "gpt-4", "key")
        .supports("test")
        .strategy(Box::new(LowestLatencyStrategy::new()))
        .build()
        .await
        .unwrap();

    assert_eq!(manager.get_provider_count().await, 1);
}

#[tokio::test]
async fn test_builder_with_random_strategy() {
    let manager = LlmManager::builder()
        .define_task(TaskDefinition::new("test"))
        .add_instance(ProviderType::OpenAI, "gpt-4", "key")
        .supports("test")
        .strategy(Box::new(RandomStrategy::new()))
        .build()
        .await
        .unwrap();

    assert_eq!(manager.get_provider_count().await, 1);
}

// ============================================================================
// TOML Strategy Configuration Tests
// ============================================================================

#[tokio::test]
async fn test_config_lru_strategy() {
    let toml = r#"
[settings]
strategy = "lru"

[[tasks]]
name = "test"

[[providers]]
type = "openai"
model = "gpt-4"
api_key = "key"
tasks = ["test"]
"#;

    let manager = LlmManager::from_config_str(toml).await.unwrap();
    assert_eq!(manager.get_provider_count().await, 1);
}

#[tokio::test]
async fn test_config_lowest_latency_strategy() {
    let toml = r#"
[settings]
strategy = "lowest_latency"

[[tasks]]
name = "test"

[[providers]]
type = "openai"
model = "gpt-4"
api_key = "key"
tasks = ["test"]
"#;

    let manager = LlmManager::from_config_str(toml).await.unwrap();
    assert_eq!(manager.get_provider_count().await, 1);
}

#[tokio::test]
async fn test_config_random_strategy() {
    let toml = r#"
[settings]
strategy = "random"

[[tasks]]
name = "test"

[[providers]]
type = "openai"
model = "gpt-4"
api_key = "key"
tasks = ["test"]
"#;

    let manager = LlmManager::from_config_str(toml).await.unwrap();
    assert_eq!(manager.get_provider_count().await, 1);
}

#[tokio::test]
async fn test_config_default_strategy() {
    let toml = r#"
[[tasks]]
name = "test"

[[providers]]
type = "openai"
model = "gpt-4"
api_key = "key"
tasks = ["test"]
"#;

    // Should default to LRU
    let manager = LlmManager::from_config_str(toml).await.unwrap();
    assert_eq!(manager.get_provider_count().await, 1);
}

// ============================================================================
// Multi-Provider Strategy Tests
// ============================================================================

#[tokio::test]
async fn test_multiple_providers_with_strategy() {
    let manager = LlmManager::builder()
        .define_task(TaskDefinition::new("chat"))
        .add_instance(ProviderType::OpenAI, "gpt-4", "key1")
        .supports("chat")
        .add_instance(ProviderType::Anthropic, "claude-3", "key2")
        .supports("chat")
        .add_instance(ProviderType::Mistral, "mistral-large", "key3")
        .supports("chat")
        .strategy(Box::new(LeastRecentlyUsedStrategy::new()))
        .build()
        .await
        .unwrap();

    // All three providers should be available for load balancing
    assert_eq!(manager.get_provider_count().await, 3);
}

#[tokio::test]
async fn test_strategy_with_disabled_provider() {
    let manager = LlmManager::builder()
        .define_task(TaskDefinition::new("chat"))
        .add_instance(ProviderType::OpenAI, "gpt-4", "key1")
        .supports("chat")
        .add_instance(ProviderType::Anthropic, "claude-3", "key2")
        .supports("chat")
        .enabled(false) // This one is disabled
        .strategy(Box::new(RandomStrategy::new()))
        .build()
        .await
        .unwrap();

    // Both providers are added, but one is disabled
    assert_eq!(manager.get_provider_count().await, 2);
}