aprender-serve 0.64.0

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
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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450

// ============================================================================
// OpenAIModel and OpenAIModelsResponse coverage
// ============================================================================

#[test]
fn test_openai_model_debug() {
    let model = OpenAIModel {
        id: "gpt-4".to_string(),
        object: "model".to_string(),
        created: 1234567890,
        owned_by: "openai".to_string(),
    };
    let debug = format!("{:?}", model);
    assert!(debug.contains("gpt-4"));
}

#[test]
fn test_openai_model_clone() {
    let model = OpenAIModel {
        id: "test".to_string(),
        object: "model".to_string(),
        created: 0,
        owned_by: "realizar".to_string(),
    };
    let cloned = model.clone();
    assert_eq!(cloned.owned_by, "realizar");
}

#[test]
fn test_openai_models_response_debug() {
    let resp = OpenAIModelsResponse {
        object: "list".to_string(),
        data: vec![],
    };
    let debug = format!("{:?}", resp);
    assert!(debug.contains("list"));
}

#[test]
fn test_openai_models_response_clone() {
    let resp = OpenAIModelsResponse {
        object: "list".to_string(),
        data: vec![OpenAIModel {
            id: "m1".to_string(),
            object: "model".to_string(),
            created: 1,
            owned_by: "test".to_string(),
        }],
    };
    let cloned = resp.clone();
    assert_eq!(cloned.data.len(), 1);
}

// ============================================================================
// Request/Response types edge cases
// ============================================================================

#[test]
fn test_tokenize_request_serde() {
    let req = TokenizeRequest {
        text: "Hello world".to_string(),
        model_id: Some("default".to_string()),
    };
    let json = serde_json::to_string(&req).expect("JSON serialization failed");
    assert!(json.contains("Hello world"));
    let parsed: TokenizeRequest = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.model_id.as_deref(), Some("default"));
}

#[test]
fn test_tokenize_request_no_model() {
    let json = r#"{"text": "test"}"#;
    let req: TokenizeRequest = serde_json::from_str(json).expect("JSON deserialization failed");
    assert!(req.model_id.is_none());
}

#[test]
fn test_tokenize_response_serde() {
    let resp = TokenizeResponse {
        token_ids: vec![1, 2, 3],
        num_tokens: 3,
    };
    let json = serde_json::to_string(&resp).expect("JSON serialization failed");
    let parsed: TokenizeResponse = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.num_tokens, 3);
}

#[test]
fn test_generate_request_serde() {
    let req = GenerateRequest {
        prompt: "Hello".to_string(),
        max_tokens: 50,
        temperature: 0.7,
        strategy: "top_k".to_string(),
        top_k: 40,
        top_p: 0.9,
        seed: Some(42),
        model_id: None,
    };
    let json = serde_json::to_string(&req).expect("JSON serialization failed");
    assert!(json.contains("top_k"));
    let parsed: GenerateRequest = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.seed, Some(42));
}

#[test]
fn test_generate_request_minimal() {
    let json = r#"{"prompt": "test"}"#;
    let req: GenerateRequest = serde_json::from_str(json).expect("JSON deserialization failed");
    assert_eq!(req.prompt, "test");
    // Check defaults - default_max_tokens() returns 50, default_temperature() returns 1.0
    assert_eq!(req.max_tokens, 50);
    assert!((req.temperature - 1.0).abs() < f32::EPSILON);
}

#[test]
fn test_generate_response_serde() {
    let resp = GenerateResponse {
        token_ids: vec![1, 2],
        text: "hello".to_string(),
        num_generated: 1,
    };
    let json = serde_json::to_string(&resp).expect("JSON serialization failed");
    let parsed: GenerateResponse = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.text, "hello");
}

#[test]
fn test_batch_tokenize_request_serde() {
    let req = BatchTokenizeRequest {
        texts: vec!["Hello".to_string(), "World".to_string()],
    };
    let json = serde_json::to_string(&req).expect("JSON serialization failed");
    let parsed: BatchTokenizeRequest = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.texts.len(), 2);
}

#[test]
fn test_batch_tokenize_response_serde() {
    let resp = BatchTokenizeResponse {
        results: vec![
            TokenizeResponse {
                token_ids: vec![1],
                num_tokens: 1,
            },
            TokenizeResponse {
                token_ids: vec![2],
                num_tokens: 1,
            },
        ],
    };
    let json = serde_json::to_string(&resp).expect("JSON serialization failed");
    let parsed: BatchTokenizeResponse = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.results.len(), 2);
}

#[test]
fn test_batch_generate_request_serde() {
    let req = BatchGenerateRequest {
        prompts: vec!["A".to_string(), "B".to_string()],
        max_tokens: 10,
        temperature: 0.5,
        strategy: "greedy".to_string(),
        top_k: 1,
        top_p: 1.0,
        seed: None,
    };
    let json = serde_json::to_string(&req).expect("JSON serialization failed");
    let parsed: BatchGenerateRequest = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.prompts.len(), 2);
}

#[test]
fn test_batch_generate_response_serde() {
    let resp = BatchGenerateResponse {
        results: vec![GenerateResponse {
            token_ids: vec![1],
            text: "a".to_string(),
            num_generated: 1,
        }],
    };
    let json = serde_json::to_string(&resp).expect("JSON serialization failed");
    let parsed: BatchGenerateResponse = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.results.len(), 1);
}

#[test]
fn test_stream_token_event_serde() {
    let event = StreamTokenEvent {
        token_id: 42,
        text: "hello".to_string(),
    };
    let json = serde_json::to_string(&event).expect("JSON serialization failed");
    let parsed: StreamTokenEvent = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.token_id, 42);
}

#[test]
fn test_stream_done_event_serde() {
    let event = StreamDoneEvent { num_generated: 10 };
    let json = serde_json::to_string(&event).expect("JSON serialization failed");
    let parsed: StreamDoneEvent = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.num_generated, 10);
}

// ============================================================================
// HTTP endpoints additional edge cases
// ============================================================================

#[tokio::test]
async fn test_empty_prompts_batch_completions() {
    let app = create_test_app_shared();
    let body = serde_json::json!({
        "prompts": [],
        "max_tokens": 5
    });
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/v1/batch/completions")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // Should return BAD_REQUEST for empty prompts
    assert!(
        response.status() == StatusCode::BAD_REQUEST
            || response.status() == StatusCode::NOT_FOUND
            || response.status() == StatusCode::SERVICE_UNAVAILABLE
    );
}

#[tokio::test]
async fn test_empty_texts_batch_tokenize() {
    let app = create_test_app_shared();
    let body = serde_json::json!({
        "texts": []
    });
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/batch/tokenize")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // Should return BAD_REQUEST for empty texts
    assert!(
        response.status() == StatusCode::BAD_REQUEST
            || response.status() == StatusCode::NOT_FOUND
            || response.status() == StatusCode::INTERNAL_SERVER_ERROR
    );
}

#[tokio::test]
async fn test_empty_prompts_batch_generate() {
    let app = create_test_app_shared();
    let body = serde_json::json!({
        "prompts": [],
        "max_tokens": 5
    });
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/batch/generate")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // Should return BAD_REQUEST for empty prompts
    assert!(
        response.status() == StatusCode::BAD_REQUEST
            || response.status() == StatusCode::NOT_FOUND
            || response.status() == StatusCode::INTERNAL_SERVER_ERROR
    );
}

#[tokio::test]
async fn test_invalid_strategy_generate() {
    let app = create_test_app_shared();
    let body = serde_json::json!({
        "prompt": "Hello",
        "max_tokens": 5,
        "strategy": "invalid_strategy"
    });
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/generate")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // Should return BAD_REQUEST for invalid strategy
    // aprender#2376(5): the shared test state has NO model, so this condition is
    // deterministic — a server with no usable model answers 503 on every route.
    // The old assertion accepted a SET of statuses that included the defect, so it
    // could not fail and held the 404-here/500-there split in place.
    assert_eq!(
        response.status(),
        StatusCode::SERVICE_UNAVAILABLE,
        "no model is resident: expected 503"
    );
}

#[tokio::test]
async fn test_models_endpoint_returns_list() {
    let app = create_test_app_shared();
    let response = app
        .oneshot(
            Request::builder()
                .uri("/models")
                .body(Body::empty())
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    assert!(response.status() == StatusCode::OK || response.status() == StatusCode::NOT_FOUND);
}

#[tokio::test]
async fn test_realize_model_handler() {
    let app = create_test_app_shared();
    let response = app
        .oneshot(
            Request::builder()
                .uri("/v1/model")
                .body(Body::empty())
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // May return OK or error depending on model state
    assert!(
        response.status() == StatusCode::OK
            || response.status() == StatusCode::NOT_FOUND
            || response.status() == StatusCode::INTERNAL_SERVER_ERROR
    );
}

#[tokio::test]
async fn test_realize_reload_handler() {
    let app = create_test_app_shared();
    let body = serde_json::json!({});
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/v1/reload")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // May return OK or error depending on model state
    assert!(
        response.status() == StatusCode::OK
            || response.status() == StatusCode::NOT_FOUND
            || response.status() == StatusCode::INTERNAL_SERVER_ERROR
            || response.status() == StatusCode::SERVICE_UNAVAILABLE
    );
}

#[tokio::test]
async fn test_chat_completions_streaming_flag() {
    let app = create_test_app_shared();
    let body = serde_json::json!({
        "model": "default",
        "messages": [{"role": "user", "content": "Hi"}],
        "stream": true,
        "max_tokens": 2
    });
    let response = app
        .oneshot(
            Request::builder()
                .method("POST")
                .uri("/v1/chat/completions")
                .header("content-type", "application/json")
                .body(Body::from(serde_json::to_string(&body).expect("JSON serialization failed")))
                .expect("test value should be present"),
        )
        .await
        .expect("test value should be present");
    // Accept streaming response or error
    // aprender#2609: this was a disjunction over four or five statuses (several
    // listing NOT_FOUND twice), so it excluded nothing and passed against the
    // very behaviour #2609 reports. The shared test app is `demo_mock()` — a
    // server with no model of any kind — so the one correct answer for a
    // MOUNTED route is 503, and that is now what is asserted.
    crate::api::test_helpers::assert_no_model_status(response.status());
}

// ============================================================================
// ChatCompletionRequest coverage
// ============================================================================

#[test]
fn test_chat_completion_request_serde() {
    let req = ChatCompletionRequest {
        model: "gpt-4".to_string(),
        messages: vec![ChatMessage {
            role: "user".to_string(),
            content: "Hello".to_string(),
            name: None,
        
            ..Default::default()
        }],
        max_tokens: Some(100),
        temperature: Some(0.7),
        top_p: Some(0.9),
        top_k: None,
        repeat_penalty: None,
        repeat_last_n: None,
        seed: None,
        n: crate::api::ChoiceCount::ONE,
        stream: false,
        stop: Some(vec!["END".to_string()]),
        user: Some("test-user".to_string()),
    
        ..Default::default()
    };
    let json = serde_json::to_string(&req).expect("JSON serialization failed");
    assert!(json.contains("gpt-4"));
    let parsed: ChatCompletionRequest = serde_json::from_str(&json).expect("JSON deserialization failed");
    assert_eq!(parsed.model, "gpt-4");
    assert_eq!(parsed.user.as_deref(), Some("test-user"));
}

#[test]
fn test_chat_completion_request_minimal() {
    let json = r#"{"model": "test", "messages": []}"#;
    let req: ChatCompletionRequest = serde_json::from_str(json).expect("JSON deserialization failed");
    assert_eq!(req.model, "test");
    assert!(!req.stream);
    assert!(req.max_tokens.is_none());
}