litellm-rs 0.6.0

A high-performance AI Gateway written in Rust, providing OpenAI-compatible APIs with intelligent routing, load balancing, and enterprise features
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

use super::*;

// ============================================================================
// OpenAI Thinking Tests
// ============================================================================

#[test]
fn test_openai_thinking_detection() {
    assert!(openai_thinking::supports_thinking("o1"));
    assert!(openai_thinking::supports_thinking("o1-preview"));
    assert!(openai_thinking::supports_thinking("o3-mini"));
    assert!(openai_thinking::supports_thinking("O1-PREVIEW")); // Case insensitive
    assert!(openai_thinking::supports_thinking("o4"));
    assert!(openai_thinking::supports_thinking("openai/o1-preview")); // With prefix
    assert!(!openai_thinking::supports_thinking("gpt-4"));
    assert!(!openai_thinking::supports_thinking("gpt-4o"));
}

#[test]
fn test_openai_capabilities() {
    let caps = openai_thinking::capabilities("o1-preview");
    assert!(caps.supports_thinking);
    assert!(!caps.supports_streaming_thinking);
    assert_eq!(caps.max_thinking_tokens, Some(20_000));
    assert_eq!(caps.supported_efforts.len(), 3);
    assert!(caps.can_return_thinking);
    assert!(!caps.thinking_always_on);
}

#[test]
fn test_openai_capabilities_non_thinking_model() {
    let caps = openai_thinking::capabilities("gpt-4");
    assert!(!caps.supports_thinking);
}

#[test]
fn test_openai_config_transform() {
    let config = ThinkingConfig {
        enabled: true,
        budget_tokens: Some(10000),
        effort: Some(ThinkingEffort::High),
        include_thinking: true,
        extra_params: Default::default(),
    };

    let result = openai_thinking::transform_config(&config, "o1").unwrap();
    assert!(result.get("max_reasoning_tokens").is_some());
    assert_eq!(
        result.get("max_reasoning_tokens").unwrap().as_u64(),
        Some(10000)
    );
    assert!(result.get("include_reasoning").is_some());
    assert_eq!(
        result.get("reasoning_effort").unwrap().as_str(),
        Some("high")
    );
}

#[test]
fn test_openai_config_transform_budget_capping() {
    let config = ThinkingConfig {
        enabled: true,
        budget_tokens: Some(50000), // Over the 20k limit
        effort: Some(ThinkingEffort::Medium),
        include_thinking: false,
        extra_params: Default::default(),
    };

    let result = openai_thinking::transform_config(&config, "o1").unwrap();
    assert_eq!(
        result.get("max_reasoning_tokens").unwrap().as_u64(),
        Some(20000)
    );
    assert_eq!(
        result.get("reasoning_effort").unwrap().as_str(),
        Some("medium")
    );
    assert!(result.get("include_reasoning").is_none());
}

#[test]
fn test_openai_config_transform_minimal() {
    let config = ThinkingConfig {
        enabled: true,
        budget_tokens: None,
        effort: None,
        include_thinking: false,
        extra_params: Default::default(),
    };

    let result = openai_thinking::transform_config(&config, "o1").unwrap();
    assert!(result.get("max_reasoning_tokens").is_none());
    assert!(result.get("reasoning_effort").is_none());
    assert!(result.get("include_reasoning").is_none());
}

#[test]
fn test_openai_config_transform_low_effort() {
    let config = ThinkingConfig {
        enabled: true,
        budget_tokens: None,
        effort: Some(ThinkingEffort::Low),
        include_thinking: false,
        extra_params: Default::default(),
    };

    let result = openai_thinking::transform_config(&config, "o1").unwrap();
    assert_eq!(
        result.get("reasoning_effort").unwrap().as_str(),
        Some("low")
    );
}

#[test]
fn test_openai_thinking_extraction() {
    let openai_response = serde_json::json!({
        "choices": [{
            "message": {
                "content": "The answer is 42.",
                "reasoning": "Let me think about this step by step..."
            }
        }],
        "usage": {
            "reasoning_tokens": 150
        }
    });

    let thinking = openai_thinking::extract_thinking(&openai_response);
    assert!(thinking.is_some());
    if let Some(ThinkingContent::Text { text, .. }) = thinking {
        assert!(text.contains("step by step"));
    }
}

#[test]
fn test_openai_thinking_extraction_missing() {
    let response = serde_json::json!({
        "choices": [{
            "message": {
                "content": "The answer is 42."
            }
        }]
    });

    let thinking = openai_thinking::extract_thinking(&response);
    assert!(thinking.is_none());
}

#[test]
fn test_openai_usage_extraction() {
    let openai_response = serde_json::json!({
        "usage": {
            "reasoning_tokens": 150
        }
    });

    let usage = openai_thinking::extract_usage(&openai_response);
    assert!(usage.is_some());
    let usage = usage.unwrap();
    assert_eq!(usage.thinking_tokens, Some(150));
    assert_eq!(usage.provider, Some("openai".to_string()));
}

#[test]
fn test_openai_usage_extraction_missing() {
    let response = serde_json::json!({
        "usage": {
            "total_tokens": 100
        }
    });

    let usage = openai_thinking::extract_usage(&response);
    assert!(usage.is_none());
}