Initial commit: LlamaLink Go rewrite
Complete rewrite from Python/FastAPI to Go/Gin: - Go backend: auth (API keys + bcrypt), llama.cpp subprocess manager, hot-swap multi-model, rate limiting, quota system, webhooks - Vue 3 SPA admin panel (src/) with Tailwind CSS - Deployment: Docker multi-stage, docker-compose, nginx, systemd - GORM/SQLite models: ApiKey, Model, UsageLog, Quota, Webhook - REST API: /api/v1/admin/* (keys, models, chat, usage, health) - Embedded frontend via go:embed (build output at web/dist/) Removed legacy Python artifacts (app/, tests/, pyproject.toml, etc.)
This commit is contained in:
@@ -0,0 +1,85 @@
|
||||
package middleware
|
||||
|
||||
import (
|
||||
"net/http"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
"golang.org/x/time/rate"
|
||||
)
|
||||
|
||||
type RateLimiter struct {
|
||||
visitors map[string]*visitor
|
||||
mu sync.RWMutex
|
||||
rate rate.Limit
|
||||
burst int
|
||||
}
|
||||
|
||||
type visitor struct {
|
||||
limiter *rate.Limiter
|
||||
lastSeen time.Time
|
||||
}
|
||||
|
||||
func NewRateLimiter(requestsPerMinute int) *RateLimiter {
|
||||
rl := &RateLimiter{
|
||||
visitors: make(map[string]*visitor),
|
||||
rate: rate.Limit(float64(requestsPerMinute) / 60.0),
|
||||
burst: requestsPerMinute / 10,
|
||||
}
|
||||
if requestsPerMinute < 10 {
|
||||
rl.burst = 1
|
||||
}
|
||||
|
||||
// Cleanup old visitors
|
||||
go rl.cleanup()
|
||||
|
||||
return rl
|
||||
}
|
||||
|
||||
func (rl *RateLimiter) cleanup() {
|
||||
ticker := time.NewTicker(5 * time.Minute)
|
||||
for range ticker.C {
|
||||
rl.mu.Lock()
|
||||
for ip, v := range rl.visitors {
|
||||
if time.Since(v.lastSeen) > 10*time.Minute {
|
||||
delete(rl.visitors, ip)
|
||||
}
|
||||
}
|
||||
rl.mu.Unlock()
|
||||
}
|
||||
}
|
||||
|
||||
func (rl *RateLimiter) getVisitor(ip string) *rate.Limiter {
|
||||
rl.mu.Lock()
|
||||
defer rl.mu.Unlock()
|
||||
|
||||
v, exists := rl.visitors[ip]
|
||||
if !exists {
|
||||
v = &visitor{
|
||||
limiter: rate.NewLimiter(rl.rate, rl.burst),
|
||||
lastSeen: time.Now(),
|
||||
}
|
||||
rl.visitors[ip] = v
|
||||
}
|
||||
|
||||
v.lastSeen = time.Now()
|
||||
return v.limiter
|
||||
}
|
||||
|
||||
func RateLimitMiddleware(rl *RateLimiter) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
ip := c.ClientIP()
|
||||
if !rl.getVisitor(ip).Allow() {
|
||||
c.AbortWithStatusJSON(http.StatusTooManyRequests, gin.H{
|
||||
"error": gin.H{
|
||||
"code": "rate_limit_exceeded",
|
||||
"message": "Too many requests",
|
||||
"retry_after_seconds": 60,
|
||||
},
|
||||
})
|
||||
return
|
||||
}
|
||||
c.Next()
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user