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
|
package embedder
import (
"bytes"
"context"
"encoding/json"
"fmt"
"net/http"
"os"
)
// Embedder generates embeddings for text
type Embedder interface {
Embed(ctx context.Context, texts []string) ([][]float32, error)
Dimensions() int
}
// OllamaEmbedder uses Ollama's embedding API
type OllamaEmbedder struct {
baseURL string
model string
dims int
}
// NewOllamaEmbedder creates an Ollama embedder
func NewOllamaEmbedder(model string) *OllamaEmbedder {
baseURL := os.Getenv("CODEVEC_BASE_URL")
if baseURL == "" {
baseURL = "http://localhost:11434"
}
if model == "" {
model = "nomic-embed-text"
}
// Model dimensions
dims := 768 // nomic-embed-text default
switch model {
case "mxbai-embed-large":
dims = 1024
case "all-minilm":
dims = 384
}
return &OllamaEmbedder{
baseURL: baseURL,
model: model,
dims: dims,
}
}
func (e *OllamaEmbedder) Dimensions() int {
return e.dims
}
type ollamaRequest struct {
Model string `json:"model"`
Prompt string `json:"prompt"`
}
type ollamaResponse struct {
Embedding []float32 `json:"embedding"`
}
func (e *OllamaEmbedder) Embed(ctx context.Context, texts []string) ([][]float32, error) {
embeddings := make([][]float32, len(texts))
// Ollama's /api/embeddings takes one prompt at a time
for i, text := range texts {
req := ollamaRequest{
Model: e.model,
Prompt: text,
}
body, err := json.Marshal(req)
if err != nil {
return nil, err
}
httpReq, err := http.NewRequestWithContext(ctx, "POST", e.baseURL+"/api/embeddings", bytes.NewReader(body))
if err != nil {
return nil, err
}
httpReq.Header.Set("Content-Type", "application/json")
resp, err := http.DefaultClient.Do(httpReq)
if err != nil {
return nil, fmt.Errorf("ollama request failed: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("ollama returned status %d", resp.StatusCode)
}
var result ollamaResponse
if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
return nil, err
}
embeddings[i] = result.Embedding
}
return embeddings, nil
}
// OpenAIEmbedder uses OpenAI-compatible embedding API
type OpenAIEmbedder struct {
baseURL string
apiKey string
model string
dims int
}
// NewOpenAIEmbedder creates an OpenAI-compatible embedder
func NewOpenAIEmbedder(model string) *OpenAIEmbedder {
baseURL := os.Getenv("CODEVEC_BASE_URL")
if baseURL == "" {
baseURL = "https://api.openai.com"
}
apiKey := os.Getenv("CODEVEC_API_KEY")
if model == "" {
model = "text-embedding-3-small"
}
dims := 1536 // text-embedding-3-small default
switch model {
case "text-embedding-3-large":
dims = 3072
case "text-embedding-ada-002":
dims = 1536
}
return &OpenAIEmbedder{
baseURL: baseURL,
apiKey: apiKey,
model: model,
dims: dims,
}
}
func (e *OpenAIEmbedder) Dimensions() int {
return e.dims
}
type openaiRequest struct {
Model string `json:"model"`
Input []string `json:"input"`
}
type openaiResponse struct {
Data []struct {
Embedding []float32 `json:"embedding"`
} `json:"data"`
}
func (e *OpenAIEmbedder) Embed(ctx context.Context, texts []string) ([][]float32, error) {
if e.apiKey == "" {
return nil, fmt.Errorf("CODEVEC_API_KEY not set")
}
// Batch in groups of 100
const batchSize = 100
embeddings := make([][]float32, len(texts))
for start := 0; start < len(texts); start += batchSize {
end := start + batchSize
if end > len(texts) {
end = len(texts)
}
batch := texts[start:end]
req := openaiRequest{
Model: e.model,
Input: batch,
}
body, err := json.Marshal(req)
if err != nil {
return nil, err
}
httpReq, err := http.NewRequestWithContext(ctx, "POST", e.baseURL+"/v1/embeddings", bytes.NewReader(body))
if err != nil {
return nil, err
}
httpReq.Header.Set("Content-Type", "application/json")
httpReq.Header.Set("Authorization", "Bearer "+e.apiKey)
resp, err := http.DefaultClient.Do(httpReq)
if err != nil {
return nil, fmt.Errorf("openai request failed: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("openai returned status %d", resp.StatusCode)
}
var result openaiResponse
if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
return nil, err
}
for i, d := range result.Data {
embeddings[start+i] = d.Embedding
}
}
return embeddings, nil
}
// New creates an embedder based on provider name
func New(provider, model string) (Embedder, error) {
switch provider {
case "ollama":
return NewOllamaEmbedder(model), nil
case "openai":
return NewOpenAIEmbedder(model), nil
default:
return nil, fmt.Errorf("unknown provider: %s", provider)
}
}
|
