提升UA解析性能（2-4倍）

internal/stats/user_agent_parser.go

@@ -3,38 +3,51 @@
 package stats
 
 import (
+	"github.com/TeaOSLab/EdgeNode/internal/goman"
 	"github.com/TeaOSLab/EdgeNode/internal/utils"
 	"github.com/TeaOSLab/EdgeNode/internal/utils/fnv"
+	syncutils "github.com/TeaOSLab/EdgeNode/internal/utils/sync"
 	"github.com/mssola/useragent"
 	"sync"
+	"time"
 )
 
 var SharedUserAgentParser = NewUserAgentParser()
 
+const userAgentShardingCount = 8
+
 // UserAgentParser UserAgent解析器
 type UserAgentParser struct {
-	parser *useragent.UserAgent
+	cacheMaps [userAgentShardingCount]map[uint64]UserAgentParserResult
+	pool      *sync.Pool
+	mu        *syncutils.RWMutex
 
-	cacheMap1     map[uint64]UserAgentParserResult
-	cacheMap2     map[uint64]UserAgentParserResult
 	maxCacheItems int
-
-	cacheCursor int
-	locker      sync.RWMutex
+	gcTicker      *time.Ticker
+	gcIndex       int
 }
 
+// NewUserAgentParser 获取新解析器
 func NewUserAgentParser() *UserAgentParser {
 	var parser = &UserAgentParser{
-		parser:      &useragent.UserAgent{},
-		cacheMap1:   map[uint64]UserAgentParserResult{},
-		cacheMap2:   map[uint64]UserAgentParserResult{},
-		cacheCursor: 0,
+		pool: &sync.Pool{
+			New: func() any {
+				return &useragent.UserAgent{}
+			},
+		},
+		cacheMaps: [userAgentShardingCount]map[uint64]UserAgentParserResult{},
+		mu:        syncutils.NewRWMutex(userAgentShardingCount),
+	}
+
+	for i := 0; i < userAgentShardingCount; i++ {
+		parser.cacheMaps[i] = map[uint64]UserAgentParserResult{}
 	}
 
 	parser.init()
 	return parser
 }
 
+// 初始化
 func (this *UserAgentParser) init() {
 	var maxCacheItems = 10_000
 	var systemMemory = utils.SystemMemoryGB()
@@ -46,8 +59,16 @@ func (this *UserAgentParser) init() {
 		maxCacheItems = 20_000
 	}
 	this.maxCacheItems = maxCacheItems
+
+	this.gcTicker = time.NewTicker(5 * time.Second)
+	goman.New(func() {
+		for range this.gcTicker.C {
+			this.GC()
+		}
+	})
 }
 
+// Parse 解析UserAgent
 func (this *UserAgentParser) Parse(userAgent string) (result UserAgentParserResult) {
 	// 限制长度
 	if len(userAgent) == 0 || len(userAgent) > 256 {
@@ -55,28 +76,22 @@ func (this *UserAgentParser) Parse(userAgent string) (result UserAgentParserResu
 	}
 
 	var userAgentKey = fnv.HashString(userAgent)
+	var shardingIndex = int(userAgentKey % userAgentShardingCount)
 
-	this.locker.RLock()
-	cacheResult, ok := this.cacheMap1[userAgentKey]
+	this.mu.RLock(shardingIndex)
+	cacheResult, ok := this.cacheMaps[shardingIndex][userAgentKey]
 	if ok {
-		this.locker.RUnlock()
+		this.mu.RUnlock(shardingIndex)
 		return cacheResult
 	}
+	this.mu.RUnlock(shardingIndex)
 
-	cacheResult, ok = this.cacheMap2[userAgentKey]
-	if ok {
-		this.locker.RUnlock()
-		return cacheResult
-	}
-	this.locker.RUnlock()
-
-	this.locker.Lock()
-	defer this.locker.Unlock()
-
-	this.parser.Parse(userAgent)
-	result.OS = this.parser.OSInfo()
-	result.BrowserName, result.BrowserVersion = this.parser.Browser()
-	result.IsMobile = this.parser.Mobile()
+	var parser = this.pool.Get().(*useragent.UserAgent)
+	parser.Parse(userAgent)
+	result.OS = parser.OSInfo()
+	result.BrowserName, result.BrowserVersion = parser.Browser()
+	result.IsMobile = parser.Mobile()
+	this.pool.Put(parser)
 
 	// 忽略特殊字符
 	if len(result.BrowserName) > 0 {
@@ -87,19 +102,45 @@ func (this *UserAgentParser) Parse(userAgent string) (result UserAgentParserResu
 		}
 	}
 
-	if this.cacheCursor == 0 {
-		this.cacheMap1[userAgentKey] = result
-		if len(this.cacheMap1) >= this.maxCacheItems {
-			this.cacheCursor = 1
-			this.cacheMap2 = map[uint64]UserAgentParserResult{}
-		}
-	} else {
-		this.cacheMap2[userAgentKey] = result
-		if len(this.cacheMap2) >= this.maxCacheItems {
-			this.cacheCursor = 0
-			this.cacheMap1 = map[uint64]UserAgentParserResult{}
-		}
-	}
+	this.mu.Lock(shardingIndex)
+	this.cacheMaps[shardingIndex][userAgentKey] = result
+	this.mu.Unlock(shardingIndex)
 
 	return
 }
+
+// MaxCacheItems 读取能容纳的缓存最大数量
+func (this *UserAgentParser) MaxCacheItems() int {
+	return this.maxCacheItems
+}
+
+// Len 读取当前缓存数量
+func (this *UserAgentParser) Len() int {
+	var total = 0
+	for i := 0; i < userAgentShardingCount; i++ {
+		this.mu.RLock(i)
+		total += len(this.cacheMaps[i])
+		this.mu.RUnlock(i)
+	}
+	return total
+}
+
+// GC 回收多余的缓存
+func (this *UserAgentParser) GC() {
+	var total = this.Len()
+	if total > this.maxCacheItems {
+		for {
+			var shardingIndex = this.gcIndex
+
+			this.mu.Lock(shardingIndex)
+			total -= len(this.cacheMaps[shardingIndex])
+			this.cacheMaps[shardingIndex] = map[uint64]UserAgentParserResult{}
+			this.gcIndex = (this.gcIndex + 1) % userAgentShardingCount
+			this.mu.Unlock(shardingIndex)
+
+			if total <= this.maxCacheItems {
+				break
+			}
+		}
+	}
+}

internal/stats/user_agent_parser_test.go

@@ -1,17 +1,21 @@
 // Copyright 2022 Liuxiangchao iwind.liu@gmail.com. All rights reserved.
 
-package stats
+package stats_test
 
 import (
+	"github.com/TeaOSLab/EdgeNode/internal/stats"
+	"github.com/TeaOSLab/EdgeNode/internal/utils/testutils"
+	"github.com/iwind/TeaGo/assert"
 	"github.com/iwind/TeaGo/rands"
 	"github.com/iwind/TeaGo/types"
 	"runtime"
 	"runtime/debug"
 	"testing"
+	"time"
 )
 
 func TestUserAgentParser_Parse(t *testing.T) {
-	var parser = NewUserAgentParser()
+	var parser = stats.NewUserAgentParser()
 	for i := 0; i < 4; i++ {
 		t.Log(parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/1"))
 		t.Log(parser.Parse("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/96.0.4664.110 Safari/537.36"))
@@ -19,7 +23,7 @@ func TestUserAgentParser_Parse(t *testing.T) {
 }
 
 func TestUserAgentParser_Parse_Unknown(t *testing.T) {
-	var parser = NewUserAgentParser()
+	var parser = stats.NewUserAgentParser()
 	t.Log(parser.Parse("Mozilla/5.0 (Wind 10.0; WOW64; rv:49.0) Apple/537.36 (KHTML, like Gecko) Chr/88.0.4324.96 Sa/537.36 Test/1"))
 	t.Log(parser.Parse(""))
 }
@@ -28,10 +32,10 @@ func TestUserAgentParser_Memory(t *testing.T) {
 	var stat1 = &runtime.MemStats{}
 	runtime.ReadMemStats(stat1)
 
-	var parser = NewUserAgentParser()
+	var parser = stats.NewUserAgentParser()
 
 	for i := 0; i < 1_000_000; i++ {
-		parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 100_000)))
+		parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 1_000_000)))
 	}
 
 	runtime.GC()
@@ -40,32 +44,76 @@ func TestUserAgentParser_Memory(t *testing.T) {
 	var stat2 = &runtime.MemStats{}
 	runtime.ReadMemStats(stat2)
 
-	t.Log("max cache items:", parser.maxCacheItems)
-	t.Log("cache1:", len(parser.cacheMap1), "cache2:", len(parser.cacheMap2), "cache3:", (stat2.HeapInuse-stat1.HeapInuse)/1024/1024, "MB")
+	t.Log("max cache items:", parser.MaxCacheItems())
+	t.Log("cache:", parser.Len(), "usage:", (stat2.HeapInuse-stat1.HeapInuse)>>20, "MB")
 }
 
-func BenchmarkUserAgentParser_Parse(b *testing.B) {
-	var parser = NewUserAgentParser()
-	for i := 0; i < b.N; i++ {
-		parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 1_000_000)))
+func TestNewUserAgentParser_GC(t *testing.T) {
+	if !testutils.IsSingleTesting() {
+		return
 	}
-	b.Log(len(parser.cacheMap1), len(parser.cacheMap2))
-}
 
-func BenchmarkUserAgentParser_Parse2(b *testing.B) {
-	var parser = NewUserAgentParser()
-	for i := 0; i < b.N; i++ {
+	var parser = stats.NewUserAgentParser()
+
+	for i := 0; i < 1_000_000; i++ {
 		parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 100_000)))
 	}
-	b.Log(len(parser.cacheMap1), len(parser.cacheMap2))
+
+	time.Sleep(60 * time.Second) // wait to gc
+	t.Log(parser.Len(), "cache items")
 }
 
-func BenchmarkUserAgentParser_Parse3(b *testing.B) {
-	var parser = NewUserAgentParser()
+func TestNewUserAgentParser_Mobile(t *testing.T) {
+	var a = assert.NewAssertion(t)
+	var parser = stats.NewUserAgentParser()
+	for _, userAgent := range []string{
+		"Mozilla/5.0 (iPhone; CPU iPhone OS 12_2 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148",
+		"Mozilla/5.0 (Linux; U; Android 2.2.1; en-us; Nexus One Build/FRG83) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1",
+	} {
+		a.IsTrue(parser.Parse(userAgent).IsMobile)
+	}
+}
+
+func BenchmarkUserAgentParser_Parse_Many_LimitCPU(b *testing.B) {
+	runtime.GOMAXPROCS(4)
+
+	var parser = stats.NewUserAgentParser()
+	b.RunParallel(func(pb *testing.PB) {
+		for pb.Next() {
+			parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 1_000_000)))
+		}
+	})
+	b.Log(parser.Len())
+}
+
+func BenchmarkUserAgentParser_Parse_Many(b *testing.B) {
+	var parser = stats.NewUserAgentParser()
+	b.RunParallel(func(pb *testing.PB) {
+		for pb.Next() {
+			parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 1_000_000)))
+		}
+	})
+	b.Log(parser.Len())
+}
+
+func BenchmarkUserAgentParser_Parse_Few_LimitCPU(b *testing.B) {
+	runtime.GOMAXPROCS(4)
+
+	var parser = stats.NewUserAgentParser()
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
 			parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 100_000)))
 		}
 	})
-	b.Log(len(parser.cacheMap1), len(parser.cacheMap2))
+	b.Log(parser.Len())
+}
+
+func BenchmarkUserAgentParser_Parse_Few(b *testing.B) {
+	var parser = stats.NewUserAgentParser()
+	b.RunParallel(func(pb *testing.PB) {
+		for pb.Next() {
+			parser.Parse("Mozilla/5.0 (Windows NT 10.0; WOW64; rv:49.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.96 Safari/537.36 Test/" + types.String(rands.Int(0, 100_000)))
+		}
+	})
+	b.Log(parser.Len())
 }