realizar 0.8.4

Pure Rust ML inference engine built from scratch - model serving for GGUF and safetensors
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


#[cfg(test)]
mod tests {
    use super::*;
    use crate::apr::{HEADER_SIZE, MAGIC};

    #[test]
    #[ignore = "Mock APR format incompatible with real AprModel::from_bytes parser"]
    fn test_batch_handler_with_errors() {
        let model_bytes = create_sum_model(2);
        let handler = LambdaHandler::from_bytes(model_bytes).expect("test");

        let request = BatchLambdaRequest {
            instances: vec![
                LambdaRequest {
                    features: vec![1.0, 2.0],
                    model_id: None,
                },
                LambdaRequest {
                    features: vec![], // Empty features will fail
                    model_id: None,
                },
                LambdaRequest {
                    features: vec![5.0, 0.0],
                    model_id: None,
                },
            ],
            max_parallelism: None,
        };

        let response = handler.handle_batch(&request).expect("test");

        assert_eq!(response.predictions.len(), 3);
        assert_eq!(response.success_count, 2);
        assert_eq!(response.error_count, 1);

        // Check successful predictions
        assert!((response.predictions[0].prediction - 3.0).abs() < 0.001);
        assert!(response.predictions[1].prediction.is_nan()); // Error placeholder
        assert!((response.predictions[2].prediction - 5.0).abs() < 0.001);
    }

    #[test]
    fn test_batch_handler_rejects_empty_batch() {
        let model_bytes = create_sum_model(2);
        let handler = LambdaHandler::from_bytes(model_bytes).expect("test");

        let request = BatchLambdaRequest {
            instances: vec![],
            max_parallelism: None,
        };

        let result = handler.handle_batch(&request);
        assert!(result.is_err());
        assert_eq!(result.unwrap_err(), LambdaError::EmptyBatch);
    }

    // --------------------------------------------------------------------------
    // Test: Prometheus metrics (PROD-001)
    // Per spec ยง11.3
    // --------------------------------------------------------------------------

    #[test]
    fn test_metrics_new() {
        let metrics = LambdaMetrics::new();
        assert_eq!(metrics.requests_total, 0);
        assert_eq!(metrics.requests_success, 0);
        assert_eq!(metrics.requests_failed, 0);
        assert_eq!(metrics.cold_starts, 0);
        assert_eq!(metrics.batch_requests, 0);
    }

    #[test]
    fn test_metrics_record_success() {
        let mut metrics = LambdaMetrics::new();

        metrics.record_success(5.0, true);
        assert_eq!(metrics.requests_total, 1);
        assert_eq!(metrics.requests_success, 1);
        assert_eq!(metrics.cold_starts, 1);
        assert!((metrics.latency_total_ms - 5.0).abs() < 0.001);

        metrics.record_success(3.0, false);
        assert_eq!(metrics.requests_total, 2);
        assert_eq!(metrics.requests_success, 2);
        assert_eq!(metrics.cold_starts, 1); // No new cold start
        assert!((metrics.latency_total_ms - 8.0).abs() < 0.001);
    }

    #[test]
    fn test_metrics_record_failure() {
        let mut metrics = LambdaMetrics::new();

        metrics.record_failure();
        assert_eq!(metrics.requests_total, 1);
        assert_eq!(metrics.requests_failed, 1);
        assert_eq!(metrics.requests_success, 0);
    }

    #[test]
    fn test_metrics_record_batch() {
        let mut metrics = LambdaMetrics::new();

        metrics.record_batch(5, 2, 10.0);
        assert_eq!(metrics.batch_requests, 1);
        assert_eq!(metrics.requests_total, 7); // 5 + 2
        assert_eq!(metrics.requests_success, 5);
        assert_eq!(metrics.requests_failed, 2);
        assert!((metrics.latency_total_ms - 10.0).abs() < 0.001);
    }

    #[test]
    fn test_metrics_avg_latency() {
        let mut metrics = LambdaMetrics::new();

        // Empty metrics returns 0
        assert!((metrics.avg_latency_ms() - 0.0).abs() < 0.001);

        metrics.record_success(4.0, false);
        metrics.record_success(6.0, false);

        // Average of 4.0 and 6.0 = 5.0
        assert!((metrics.avg_latency_ms() - 5.0).abs() < 0.001);
    }

    #[test]
    fn test_metrics_prometheus_format() {
        let mut metrics = LambdaMetrics::new();
        metrics.record_success(5.0, true);
        metrics.record_success(3.0, false);
        metrics.record_failure();
        metrics.record_batch(10, 2, 20.0);

        let prom = metrics.to_prometheus();

        // Verify Prometheus format
        assert!(prom.contains("# HELP lambda_requests_total"));
        assert!(prom.contains("# TYPE lambda_requests_total counter"));
        assert!(prom.contains("lambda_requests_total 15")); // 2 + 1 + 12
        assert!(prom.contains("lambda_requests_success 12")); // 2 + 10
        assert!(prom.contains("lambda_requests_failed 3")); // 1 + 2
        assert!(prom.contains("lambda_cold_starts 1"));
        assert!(prom.contains("lambda_batch_requests 1"));
        assert!(prom.contains("lambda_latency_avg_ms"));
    }

    #[test]
    fn test_metrics_default() {
        let metrics = LambdaMetrics::default();
        assert_eq!(metrics.requests_total, 0);
        assert_eq!(metrics.batch_requests, 0);
    }
include!("lambda_mock_model.rs");
}