types.go 13.4 KB
Newer Older
Jeffrey Morgan's avatar
Jeffrey Morgan committed
1
2
package api

3
import (
Michael Yang's avatar
Michael Yang committed
4
	"encoding/json"
5
	"fmt"
Michael Yang's avatar
Michael Yang committed
6
	"math"
7
	"os"
8
	"reflect"
9
	"strconv"
10
	"strings"
11
12
	"time"
)
Michael Yang's avatar
Michael Yang committed
13

Patrick Devine's avatar
Patrick Devine committed
14
type StatusError struct {
15
16
17
	StatusCode   int
	Status       string
	ErrorMessage string `json:"error"`
Patrick Devine's avatar
Patrick Devine committed
18
19
20
}

func (e StatusError) Error() string {
21
22
23
24
25
26
27
28
29
30
	switch {
	case e.Status != "" && e.ErrorMessage != "":
		return fmt.Sprintf("%s: %s", e.Status, e.ErrorMessage)
	case e.Status != "":
		return e.Status
	case e.ErrorMessage != "":
		return e.ErrorMessage
	default:
		// this should not happen
		return "something went wrong, please see the ollama server logs for details"
Patrick Devine's avatar
Patrick Devine committed
31
32
33
	}
}

Patrick Devine's avatar
Patrick Devine committed
34
35
type ImageData []byte

36
type GenerateRequest struct {
37
38
39
40
41
42
43
44
45
46
	Model     string      `json:"model"`
	Prompt    string      `json:"prompt"`
	System    string      `json:"system"`
	Template  string      `json:"template"`
	Context   []int       `json:"context,omitempty"`
	Stream    *bool       `json:"stream,omitempty"`
	Raw       bool        `json:"raw,omitempty"`
	Format    string      `json:"format"`
	KeepAlive *Duration   `json:"keep_alive,omitempty"`
	Images    []ImageData `json:"images,omitempty"`
47

48
	Options map[string]interface{} `json:"options"`
49
50
}

Bruce MacDonald's avatar
Bruce MacDonald committed
51
type ChatRequest struct {
52
53
54
55
56
	Model     string    `json:"model"`
	Messages  []Message `json:"messages"`
	Stream    *bool     `json:"stream,omitempty"`
	Format    string    `json:"format"`
	KeepAlive *Duration `json:"keep_alive,omitempty"`
Bruce MacDonald's avatar
Bruce MacDonald committed
57
58
59
60
61

	Options map[string]interface{} `json:"options"`
}

type Message struct {
62
63
	Role    string      `json:"role"` // one of ["system", "user", "assistant"]
	Content string      `json:"content"`
64
	Images  []ImageData `json:"images,omitempty"`
Bruce MacDonald's avatar
Bruce MacDonald committed
65
66
67
68
69
}

type ChatResponse struct {
	Model     string    `json:"model"`
	CreatedAt time.Time `json:"created_at"`
70
	Message   Message   `json:"message"`
Bruce MacDonald's avatar
Bruce MacDonald committed
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85

	Done bool `json:"done"`

	Metrics
}

type Metrics struct {
	TotalDuration      time.Duration `json:"total_duration,omitempty"`
	LoadDuration       time.Duration `json:"load_duration,omitempty"`
	PromptEvalCount    int           `json:"prompt_eval_count,omitempty"`
	PromptEvalDuration time.Duration `json:"prompt_eval_duration,omitempty"`
	EvalCount          int           `json:"eval_count,omitempty"`
	EvalDuration       time.Duration `json:"eval_duration,omitempty"`
}

86
// Options specified in GenerateRequest, if you add a new option here add it to the API docs also
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
type Options struct {
	Runner

	// Predict options used at runtime
	NumKeep          int      `json:"num_keep,omitempty"`
	Seed             int      `json:"seed,omitempty"`
	NumPredict       int      `json:"num_predict,omitempty"`
	TopK             int      `json:"top_k,omitempty"`
	TopP             float32  `json:"top_p,omitempty"`
	TFSZ             float32  `json:"tfs_z,omitempty"`
	TypicalP         float32  `json:"typical_p,omitempty"`
	RepeatLastN      int      `json:"repeat_last_n,omitempty"`
	Temperature      float32  `json:"temperature,omitempty"`
	RepeatPenalty    float32  `json:"repeat_penalty,omitempty"`
	PresencePenalty  float32  `json:"presence_penalty,omitempty"`
	FrequencyPenalty float32  `json:"frequency_penalty,omitempty"`
	Mirostat         int      `json:"mirostat,omitempty"`
	MirostatTau      float32  `json:"mirostat_tau,omitempty"`
	MirostatEta      float32  `json:"mirostat_eta,omitempty"`
	PenalizeNewline  bool     `json:"penalize_newline,omitempty"`
	Stop             []string `json:"stop,omitempty"`
}

// Runner options which must be set when the model is loaded into memory
type Runner struct {
Michael Yang's avatar
Michael Yang committed
112
113
114
115
116
117
118
119
120
121
122
123
124
	UseNUMA   bool `json:"numa,omitempty"`
	NumCtx    int  `json:"num_ctx,omitempty"`
	NumBatch  int  `json:"num_batch,omitempty"`
	NumGQA    int  `json:"num_gqa,omitempty"`
	NumGPU    int  `json:"num_gpu,omitempty"`
	MainGPU   int  `json:"main_gpu,omitempty"`
	LowVRAM   bool `json:"low_vram,omitempty"`
	F16KV     bool `json:"f16_kv,omitempty"`
	LogitsAll bool `json:"logits_all,omitempty"`
	VocabOnly bool `json:"vocab_only,omitempty"`
	UseMMap   bool `json:"use_mmap,omitempty"`
	UseMLock  bool `json:"use_mlock,omitempty"`
	NumThread int  `json:"num_thread,omitempty"`
125
126
}

Bruce MacDonald's avatar
Bruce MacDonald committed
127
type EmbeddingRequest struct {
128
129
130
	Model     string    `json:"model"`
	Prompt    string    `json:"prompt"`
	KeepAlive *Duration `json:"keep_alive,omitempty"`
Bruce MacDonald's avatar
Bruce MacDonald committed
131
132
133
134
135
136
137
138

	Options map[string]interface{} `json:"options"`
}

type EmbeddingResponse struct {
	Embedding []float64 `json:"embedding"`
}

139
type CreateRequest struct {
Michael Yang's avatar
Michael Yang committed
140
141
142
143
144
	Model        string `json:"model"`
	Path         string `json:"path"`
	Modelfile    string `json:"modelfile"`
	Stream       *bool  `json:"stream,omitempty"`
	Quantization string `json:"quantization,omitempty"`
Michael Yang's avatar
Michael Yang committed
145
146
147

	// Name is deprecated, see Model
	Name string `json:"name"`
148
149
}

150
type DeleteRequest struct {
Michael Yang's avatar
Michael Yang committed
151
152
153
	Model string `json:"model"`

	// Name is deprecated, see Model
154
155
156
	Name string `json:"name"`
}

Patrick Devine's avatar
Patrick Devine committed
157
type ShowRequest struct {
158
159
160
161
162
	Model    string `json:"model"`
	System   string `json:"system"`
	Template string `json:"template"`

	Options map[string]interface{} `json:"options"`
Michael Yang's avatar
Michael Yang committed
163
164
165

	// Name is deprecated, see Model
	Name string `json:"name"`
Patrick Devine's avatar
Patrick Devine committed
166
167
168
}

type ShowResponse struct {
Patrick Devine's avatar
Patrick Devine committed
169
170
171
172
173
174
	License    string       `json:"license,omitempty"`
	Modelfile  string       `json:"modelfile,omitempty"`
	Parameters string       `json:"parameters,omitempty"`
	Template   string       `json:"template,omitempty"`
	System     string       `json:"system,omitempty"`
	Details    ModelDetails `json:"details,omitempty"`
175
	Messages   []Message    `json:"messages,omitempty"`
Patrick Devine's avatar
Patrick Devine committed
176
177
}

Patrick Devine's avatar
Patrick Devine committed
178
179
180
181
182
type CopyRequest struct {
	Source      string `json:"source"`
	Destination string `json:"destination"`
}

Bruce MacDonald's avatar
Bruce MacDonald committed
183
type PullRequest struct {
Michael Yang's avatar
Michael Yang committed
184
	Model    string `json:"model"`
185
	Insecure bool   `json:"insecure,omitempty"`
186
187
188
	Username string `json:"username"`
	Password string `json:"password"`
	Stream   *bool  `json:"stream,omitempty"`
Michael Yang's avatar
Michael Yang committed
189
190
191

	// Name is deprecated, see Model
	Name string `json:"name"`
Bruce MacDonald's avatar
Bruce MacDonald committed
192
193
}

194
type ProgressResponse struct {
195
196
	Status    string `json:"status"`
	Digest    string `json:"digest,omitempty"`
Michael Yang's avatar
Michael Yang committed
197
198
	Total     int64  `json:"total,omitempty"`
	Completed int64  `json:"completed,omitempty"`
Bruce MacDonald's avatar
Bruce MacDonald committed
199
200
}

201
type PushRequest struct {
Michael Yang's avatar
Michael Yang committed
202
	Model    string `json:"model"`
203
	Insecure bool   `json:"insecure,omitempty"`
204
205
206
	Username string `json:"username"`
	Password string `json:"password"`
	Stream   *bool  `json:"stream,omitempty"`
Michael Yang's avatar
Michael Yang committed
207
208
209

	// Name is deprecated, see Model
	Name string `json:"name"`
210
}
211

Patrick Devine's avatar
Patrick Devine committed
212
type ListResponse struct {
213
	Models []ModelResponse `json:"models"`
Patrick Devine's avatar
Patrick Devine committed
214
215
}

216
type ModelResponse struct {
Patrick Devine's avatar
Patrick Devine committed
217
	Name       string       `json:"name"`
Michael Yang's avatar
Michael Yang committed
218
	Model      string       `json:"model"`
Patrick Devine's avatar
Patrick Devine committed
219
220
221
222
	ModifiedAt time.Time    `json:"modified_at"`
	Size       int64        `json:"size"`
	Digest     string       `json:"digest"`
	Details    ModelDetails `json:"details,omitempty"`
Patrick Devine's avatar
Patrick Devine committed
223
224
}

Patrick Devine's avatar
Patrick Devine committed
225
226
227
228
type TokenResponse struct {
	Token string `json:"token"`
}

Michael Yang's avatar
Michael Yang committed
229
type GenerateResponse struct {
230
231
	Model     string    `json:"model"`
	CreatedAt time.Time `json:"created_at"`
232
	Response  string    `json:"response"`
233

Michael Yang's avatar
Michael Yang committed
234
235
	Done    bool  `json:"done"`
	Context []int `json:"context,omitempty"`
236

Bruce MacDonald's avatar
Bruce MacDonald committed
237
	Metrics
238
239
}

Patrick Devine's avatar
Patrick Devine committed
240
type ModelDetails struct {
241
	ParentModel       string   `json:"parent_model"`
Patrick Devine's avatar
Patrick Devine committed
242
243
244
245
246
247
248
	Format            string   `json:"format"`
	Family            string   `json:"family"`
	Families          []string `json:"families"`
	ParameterSize     string   `json:"parameter_size"`
	QuantizationLevel string   `json:"quantization_level"`
}

Bruce MacDonald's avatar
Bruce MacDonald committed
249
250
251
func (m *Metrics) Summary() {
	if m.TotalDuration > 0 {
		fmt.Fprintf(os.Stderr, "total duration:       %v\n", m.TotalDuration)
252
253
	}

Bruce MacDonald's avatar
Bruce MacDonald committed
254
255
	if m.LoadDuration > 0 {
		fmt.Fprintf(os.Stderr, "load duration:        %v\n", m.LoadDuration)
Michael Yang's avatar
Michael Yang committed
256
257
	}

Bruce MacDonald's avatar
Bruce MacDonald committed
258
259
	if m.PromptEvalCount > 0 {
		fmt.Fprintf(os.Stderr, "prompt eval count:    %d token(s)\n", m.PromptEvalCount)
260
261
	}

Bruce MacDonald's avatar
Bruce MacDonald committed
262
263
264
	if m.PromptEvalDuration > 0 {
		fmt.Fprintf(os.Stderr, "prompt eval duration: %s\n", m.PromptEvalDuration)
		fmt.Fprintf(os.Stderr, "prompt eval rate:     %.2f tokens/s\n", float64(m.PromptEvalCount)/m.PromptEvalDuration.Seconds())
265
266
	}

Bruce MacDonald's avatar
Bruce MacDonald committed
267
268
	if m.EvalCount > 0 {
		fmt.Fprintf(os.Stderr, "eval count:           %d token(s)\n", m.EvalCount)
269
270
	}

Bruce MacDonald's avatar
Bruce MacDonald committed
271
272
273
	if m.EvalDuration > 0 {
		fmt.Fprintf(os.Stderr, "eval duration:        %s\n", m.EvalDuration)
		fmt.Fprintf(os.Stderr, "eval rate:            %.2f tokens/s\n", float64(m.EvalCount)/m.EvalDuration.Seconds())
274
	}
275
276
}

277
278
var ErrInvalidOpts = fmt.Errorf("invalid options")

279
280
281
282
283
284
285
286
287
288
289
290
291
func (opts *Options) FromMap(m map[string]interface{}) error {
	valueOpts := reflect.ValueOf(opts).Elem() // names of the fields in the options struct
	typeOpts := reflect.TypeOf(opts).Elem()   // types of the fields in the options struct

	// build map of json struct tags to their types
	jsonOpts := make(map[string]reflect.StructField)
	for _, field := range reflect.VisibleFields(typeOpts) {
		jsonTag := strings.Split(field.Tag.Get("json"), ",")[0]
		if jsonTag != "" {
			jsonOpts[jsonTag] = field
		}
	}

292
	invalidOpts := []string{}
293
294
295
296
	for key, val := range m {
		if opt, ok := jsonOpts[key]; ok {
			field := valueOpts.FieldByName(opt.Name)
			if field.IsValid() && field.CanSet() {
Michael Yang's avatar
Michael Yang committed
297
298
299
300
				if val == nil {
					continue
				}

301
302
				switch field.Kind() {
				case reflect.Int:
Michael Yang's avatar
Michael Yang committed
303
304
305
306
307
308
309
					switch t := val.(type) {
					case int64:
						field.SetInt(t)
					case float64:
						// when JSON unmarshals numbers, it uses float64, not int
						field.SetInt(int64(t))
					default:
310
						return fmt.Errorf("option %q must be of type integer", key)
311
312
313
314
					}
				case reflect.Bool:
					val, ok := val.(bool)
					if !ok {
315
						return fmt.Errorf("option %q must be of type boolean", key)
316
317
318
319
320
321
					}
					field.SetBool(val)
				case reflect.Float32:
					// JSON unmarshals to float64
					val, ok := val.(float64)
					if !ok {
322
						return fmt.Errorf("option %q must be of type float32", key)
323
324
325
326
327
					}
					field.SetFloat(val)
				case reflect.String:
					val, ok := val.(string)
					if !ok {
328
						return fmt.Errorf("option %q must be of type string", key)
329
330
331
332
333
334
					}
					field.SetString(val)
				case reflect.Slice:
					// JSON unmarshals to []interface{}, not []string
					val, ok := val.([]interface{})
					if !ok {
335
						return fmt.Errorf("option %q must be of type array", key)
336
337
338
339
340
341
					}
					// convert []interface{} to []string
					slice := make([]string, len(val))
					for i, item := range val {
						str, ok := item.(string)
						if !ok {
342
							return fmt.Errorf("option %q must be of an array of strings", key)
343
344
345
346
347
348
349
350
						}
						slice[i] = str
					}
					field.Set(reflect.ValueOf(slice))
				default:
					return fmt.Errorf("unknown type loading config params: %v", field.Kind())
				}
			}
351
352
		} else {
			invalidOpts = append(invalidOpts, key)
353
354
		}
	}
355
356
357
358

	if len(invalidOpts) > 0 {
		return fmt.Errorf("%w: %v", ErrInvalidOpts, strings.Join(invalidOpts, ", "))
	}
359
360
361
	return nil
}

Michael Yang's avatar
Michael Yang committed
362
363
func DefaultOptions() Options {
	return Options{
364
365
		// options set on request to runner
		NumPredict:       -1,
366
		NumKeep:          0,
Michael Yang's avatar
Michael Yang committed
367
368
369
370
371
		Temperature:      0.8,
		TopK:             40,
		TopP:             0.9,
		TFSZ:             1.0,
		TypicalP:         1.0,
372
373
374
375
		RepeatLastN:      64,
		RepeatPenalty:    1.1,
		PresencePenalty:  0.0,
		FrequencyPenalty: 0.0,
Michael Yang's avatar
Michael Yang committed
376
377
378
		Mirostat:         0,
		MirostatTau:      5.0,
		MirostatEta:      0.1,
Michael Yang's avatar
Michael Yang committed
379
		PenalizeNewline:  true,
380
		Seed:             -1,
Michael Yang's avatar
Michael Yang committed
381

382
383
		Runner: Runner{
			// options set when the model is loaded
Michael Yang's avatar
Michael Yang committed
384
385
386
387
388
389
390
391
392
393
			NumCtx:    2048,
			NumBatch:  512,
			NumGPU:    -1, // -1 here indicates that NumGPU should be set dynamically
			NumGQA:    1,
			NumThread: 0, // let the runtime decide
			LowVRAM:   false,
			F16KV:     true,
			UseMLock:  false,
			UseMMap:   true,
			UseNUMA:   false,
394
		},
Michael Yang's avatar
Michael Yang committed
395
	}
Jeffrey Morgan's avatar
Jeffrey Morgan committed
396
}
Michael Yang's avatar
Michael Yang committed
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412

type Duration struct {
	time.Duration
}

func (d *Duration) UnmarshalJSON(b []byte) (err error) {
	var v any
	if err := json.Unmarshal(b, &v); err != nil {
		return err
	}

	d.Duration = 5 * time.Minute

	switch t := v.(type) {
	case float64:
		if t < 0 {
bnorick's avatar
bnorick committed
413
			d.Duration = time.Duration(math.MaxInt64)
414
415
		} else {
			d.Duration = time.Duration(t * float64(time.Second))
Michael Yang's avatar
Michael Yang committed
416
417
418
419
420
421
		}
	case string:
		d.Duration, err = time.ParseDuration(t)
		if err != nil {
			return err
		}
422
		if d.Duration < 0 {
bnorick's avatar
bnorick committed
423
			d.Duration = time.Duration(math.MaxInt64)
424
		}
Michael Yang's avatar
Michael Yang committed
425
426
427
428
	}

	return nil
}
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488

// FormatParams converts specified parameter options to their correct types
func FormatParams(params map[string][]string) (map[string]interface{}, error) {
	opts := Options{}
	valueOpts := reflect.ValueOf(&opts).Elem() // names of the fields in the options struct
	typeOpts := reflect.TypeOf(opts)           // types of the fields in the options struct

	// build map of json struct tags to their types
	jsonOpts := make(map[string]reflect.StructField)
	for _, field := range reflect.VisibleFields(typeOpts) {
		jsonTag := strings.Split(field.Tag.Get("json"), ",")[0]
		if jsonTag != "" {
			jsonOpts[jsonTag] = field
		}
	}

	out := make(map[string]interface{})
	// iterate params and set values based on json struct tags
	for key, vals := range params {
		if opt, ok := jsonOpts[key]; !ok {
			return nil, fmt.Errorf("unknown parameter '%s'", key)
		} else {
			field := valueOpts.FieldByName(opt.Name)
			if field.IsValid() && field.CanSet() {
				switch field.Kind() {
				case reflect.Float32:
					floatVal, err := strconv.ParseFloat(vals[0], 32)
					if err != nil {
						return nil, fmt.Errorf("invalid float value %s", vals)
					}

					out[key] = float32(floatVal)
				case reflect.Int:
					intVal, err := strconv.ParseInt(vals[0], 10, 64)
					if err != nil {
						return nil, fmt.Errorf("invalid int value %s", vals)
					}

					out[key] = intVal
				case reflect.Bool:
					boolVal, err := strconv.ParseBool(vals[0])
					if err != nil {
						return nil, fmt.Errorf("invalid bool value %s", vals)
					}

					out[key] = boolVal
				case reflect.String:
					out[key] = vals[0]
				case reflect.Slice:
					// TODO: only string slices are supported right now
					out[key] = vals
				default:
					return nil, fmt.Errorf("unknown type %s for %s", field.Kind(), key)
				}
			}
		}
	}

	return out, nil
}