mirror of
https://gitcode.com/JianFeeeee/HomeAgent.git
synced 2026-09-21 17:38:10 +00:00
- IO abstraction layer with OutputChannel routing and capability validation - Three-layer memory (Context-Document-Graph) with TF-IDF relevance pruning - OneBot V11 QQ protocol plugin with Reverse WebSocket client - Plugin system with hot-reload (SKILL.md + native factories) - Knowledge system with TF-IDF vector indexing - Personality system (personal.md) - Text memory (JSONL with rotation) - Change tracker (overlayfs) with rollback - Lua adapter VM - Design document (DESIGN.md) Module: gitcode.com/JianFeeeee/HomeAgent
163 lines
2.8 KiB
Go
163 lines
2.8 KiB
Go
package embed
|
|
|
|
import (
|
|
"bytes"
|
|
"encoding/json"
|
|
"fmt"
|
|
"math"
|
|
"net/http"
|
|
"sync"
|
|
"time"
|
|
)
|
|
|
|
type Embedder interface {
|
|
Embed(text string) ([]float64, error)
|
|
Similarity(a, b []float64) float64
|
|
Dimension() int
|
|
}
|
|
|
|
type OllamaEmbedder struct {
|
|
client *http.Client
|
|
baseURL string
|
|
model string
|
|
dimension int
|
|
mu sync.RWMutex
|
|
}
|
|
|
|
func NewOllamaEmbedder(baseURL, model string, dimension int) *OllamaEmbedder {
|
|
if baseURL == "" {
|
|
baseURL = "http://localhost:11434"
|
|
}
|
|
if model == "" {
|
|
model = "nomic-embed-text"
|
|
}
|
|
if dimension <= 0 {
|
|
dimension = 768
|
|
}
|
|
|
|
return &OllamaEmbedder{
|
|
client: &http.Client{
|
|
Timeout: 30 * time.Second,
|
|
},
|
|
baseURL: baseURL,
|
|
model: model,
|
|
dimension: dimension,
|
|
}
|
|
}
|
|
|
|
func (e *OllamaEmbedder) Embed(text string) ([]float64, error) {
|
|
if text == "" {
|
|
return make([]float64, e.dimension), nil
|
|
}
|
|
|
|
reqBody := map[string]interface{}{
|
|
"model": e.model,
|
|
"prompt": text,
|
|
}
|
|
|
|
data, err := json.Marshal(reqBody)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("marshal: %w", err)
|
|
}
|
|
|
|
resp, err := e.client.Post(e.baseURL+"/api/embeddings", "application/json", bytes.NewReader(data))
|
|
if err != nil {
|
|
return nil, fmt.Errorf("ollama api: %w", err)
|
|
}
|
|
defer resp.Body.Close()
|
|
|
|
var result struct {
|
|
Embedding []float64 `json:"embedding"`
|
|
}
|
|
if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
|
|
return nil, fmt.Errorf("decode: %w", err)
|
|
}
|
|
|
|
return result.Embedding, nil
|
|
}
|
|
|
|
func (e *OllamaEmbedder) Similarity(a, b []float64) float64 {
|
|
return cosineSimilarity(a, b)
|
|
}
|
|
|
|
func (e *OllamaEmbedder) Dimension() int {
|
|
return e.dimension
|
|
}
|
|
|
|
type HashEmbedder struct {
|
|
dimension int
|
|
}
|
|
|
|
func NewHashEmbedder(dimension int) *HashEmbedder {
|
|
if dimension <= 0 {
|
|
dimension = 64
|
|
}
|
|
return &HashEmbedder{dimension: dimension}
|
|
}
|
|
|
|
func (e *HashEmbedder) Embed(text string) ([]float64, error) {
|
|
vec := make([]float64, e.dimension)
|
|
runes := []rune(text)
|
|
if len(runes) == 0 {
|
|
return vec, nil
|
|
}
|
|
|
|
// Character-level hash embedding
|
|
for i, r := range runes {
|
|
h := hashRune(r)
|
|
idx := i % e.dimension
|
|
vec[idx] += float64(h) / 65536.0
|
|
}
|
|
|
|
// Normalize
|
|
mag := 0.0
|
|
for _, v := range vec {
|
|
mag += v * v
|
|
}
|
|
if mag > 0 {
|
|
mag = math.Sqrt(mag)
|
|
for i := range vec {
|
|
vec[i] /= mag
|
|
}
|
|
}
|
|
|
|
return vec, nil
|
|
}
|
|
|
|
func (e *HashEmbedder) Similarity(a, b []float64) float64 {
|
|
return cosineSimilarity(a, b)
|
|
}
|
|
|
|
func (e *HashEmbedder) Dimension() int {
|
|
return e.dimension
|
|
}
|
|
|
|
func hashRune(r rune) uint64 {
|
|
h := uint64(r)
|
|
h ^= h >> 33
|
|
h *= 0xff51afd7ed558ccd
|
|
h ^= h >> 33
|
|
h *= 0xc4ceb9fe1a85ec53
|
|
h ^= h >> 33
|
|
return h
|
|
}
|
|
|
|
func cosineSimilarity(a, b []float64) float64 {
|
|
if len(a) != len(b) || len(a) == 0 {
|
|
return 0
|
|
}
|
|
|
|
var dot, na, nb float64
|
|
for i := range a {
|
|
dot += a[i] * b[i]
|
|
na += a[i] * a[i]
|
|
nb += b[i] * b[i]
|
|
}
|
|
|
|
if na == 0 || nb == 0 {
|
|
return 0
|
|
}
|
|
|
|
return dot / (math.Sqrt(na) * math.Sqrt(nb))
|
|
}
|