优化UserAgent解析

internal/stats/http_request_stat_manager.go

@@ -15,7 +15,6 @@ import (
 	"github.com/iwind/TeaGo/maps"
 	"github.com/iwind/TeaGo/types"
 	timeutil "github.com/iwind/TeaGo/utils/time"
-	"github.com/mssola/user_agent"
 	"strconv"
 	"strings"
 	"time"
@@ -177,7 +176,7 @@ func (this *HTTPRequestStatManager) AddFirewallRuleGroupId(serverId int64, firew
 // Loop 单个循环
 func (this *HTTPRequestStatManager) Loop() error {
 	timeout := time.NewTimer(10 * time.Minute) // 执行的最大时间
-	userAgentParser := &user_agent.UserAgent{}
+	userAgentParser := NewUserAgentParser()
 Loop:
 	for {
 		select {
@@ -223,8 +222,8 @@ Loop:
 			serverId := userAgentString[:atIndex]
 			userAgent := userAgentString[atIndex+1:]
 
-			userAgentParser.Parse(userAgent)
+			var result = userAgentParser.Parse(userAgent)
-			osInfo := userAgentParser.OSInfo()
+			var osInfo = result.os
 			if len(osInfo.Name) > 0 {
 				dotIndex := strings.Index(osInfo.Version, ".")
 				if dotIndex > -1 {
@@ -233,7 +232,7 @@ Loop:
 				this.systemMap[serverId+"@"+osInfo.Name+"@"+osInfo.Version]++
 			}
 
-			browser, browserVersion := userAgentParser.Browser()
+			var browser, browserVersion = result.browserName, result.browserVersion
 			if len(browser) > 0 {
 				dotIndex := strings.Index(browserVersion, ".")
 				if dotIndex > -1 {

internal/stats/user_agent_parser.go

@@ -0,0 +1,82 @@
+// Copyright 2022 Liuxiangchao iwind.liu@gmail.com. All rights reserved.
+
+package stats
+
+import (
+	"github.com/TeaOSLab/EdgeNode/internal/utils"
+	"github.com/mssola/user_agent"
+)
+
+// UserAgentParser UserAgent解析器
+// 只支持单线程
+type UserAgentParser struct {
+	parser *user_agent.UserAgent
+
+	cacheMap1     map[string]UserAgentParserResult
+	cacheMap2     map[string]UserAgentParserResult
+	maxCacheItems int
+
+	cacheCursor int
+}
+
+func NewUserAgentParser() *UserAgentParser {
+	var parser = &UserAgentParser{
+		parser:      &user_agent.UserAgent{},
+		cacheMap1:   map[string]UserAgentParserResult{},
+		cacheMap2:   map[string]UserAgentParserResult{},
+		cacheCursor: 0,
+	}
+
+	parser.init()
+	return parser
+}
+
+func (this *UserAgentParser) init() {
+	var maxCacheItems = 10_000
+	var systemMemory = utils.SystemMemoryGB()
+	if systemMemory >= 16 {
+		maxCacheItems = 40_000
+	} else if systemMemory >= 8 {
+		maxCacheItems = 30_000
+	} else if systemMemory >= 4 {
+		maxCacheItems = 20_000
+	}
+	this.maxCacheItems = maxCacheItems
+}
+
+func (this *UserAgentParser) Parse(userAgent string) (result UserAgentParserResult) {
+	// 限制长度
+	if len(userAgent) == 0 || len(userAgent) > 256 {
+		return
+	}
+
+	cacheResult, ok := this.cacheMap1[userAgent]
+	if ok {
+		return cacheResult
+	}
+
+	cacheResult, ok = this.cacheMap2[userAgent]
+	if ok {
+		return cacheResult
+	}
+
+	this.parser.Parse(userAgent)
+	result.os = this.parser.OSInfo()
+	result.browserName, result.browserVersion = this.parser.Browser()
+
+	if this.cacheCursor == 0 {
+		this.cacheMap1[userAgent] = result
+		if len(this.cacheMap1) >= this.maxCacheItems {
+			this.cacheCursor = 1
+			this.cacheMap2 = map[string]UserAgentParserResult{}
+		}
+	} else {
+		this.cacheMap2[userAgent] = result
+		if len(this.cacheMap2) >= this.maxCacheItems {
+			this.cacheCursor = 0
+			this.cacheMap1 = map[string]UserAgentParserResult{}
+		}
+	}
+
+	return
+}

internal/stats/user_agent_parser_result.go

@@ -0,0 +1,11 @@
+// Copyright 2022 Liuxiangchao iwind.liu@gmail.com. All rights reserved.
+
+package stats
+
+import "github.com/mssola/user_agent"
+
+type UserAgentParserResult struct {
+	os             user_agent.OSInfo
+	browserName    string
+	browserVersion string
+}

internal/stats/user_agent_parser_test.go

@@ -0,0 +1,51 @@
+// Copyright 2022 Liuxiangchao iwind.liu@gmail.com. All rights reserved.
+
+package stats
+
+import (
+	"github.com/iwind/TeaGo/rands"
+	"github.com/iwind/TeaGo/types"
+	"runtime"
+	"runtime/debug"
+	"testing"
+)
+
+func TestUserAgentParser_Parse(t *testing.T) {
+	var parser = NewUserAgentParser()
+	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"))
+}
+
+func TestUserAgentParser_Parse_Unknown(t *testing.T) {
+	var parser = 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(""))
+}
+
+func TestUserAgentParser_Memory(t *testing.T) {
+	var stat1 = &runtime.MemStats{}
+	runtime.ReadMemStats(stat1)
+
+	var parser = 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)))
+	}
+
+	runtime.GC()
+	debug.FreeOSMemory()
+
+	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")
+}
+
+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, 40000)))
+	}
+	b.Log(len(parser.cacheMap1), len(parser.cacheMap2))
+}