🚧 Optimize
This commit is contained in:
@@ -2,15 +2,14 @@
|
||||
package translator
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"context"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"math/rand"
|
||||
"net/http"
|
||||
"net/http/cookiejar"
|
||||
"net/url"
|
||||
"regexp"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
@@ -21,68 +20,61 @@ import (
|
||||
type BingTranslator struct {
|
||||
httpClient *http.Client // HTTP客户端
|
||||
Timeout time.Duration // 请求超时时间
|
||||
session *BingSession // Bing会话
|
||||
token *tokenInfo // Token信息
|
||||
languages map[string]LanguageInfo // 支持的语言列表
|
||||
}
|
||||
|
||||
// BingSession 保持Bing翻译会话状态
|
||||
type BingSession struct {
|
||||
Cookie map[string]string // 会话Cookie
|
||||
Headers map[string]string // 会话请求头
|
||||
Token string // 翻译Token
|
||||
Key string // 翻译Key
|
||||
IG string // IG参数
|
||||
// tokenInfo 存储token信息
|
||||
type tokenInfo struct {
|
||||
Value string
|
||||
ExpiresAt time.Time
|
||||
}
|
||||
|
||||
// TranslateRequest 翻译请求结构
|
||||
type TranslateRequest struct {
|
||||
Text string `json:"Text"`
|
||||
}
|
||||
|
||||
// TranslateResponse 翻译响应结构
|
||||
type TranslateResponse struct {
|
||||
Translations []struct {
|
||||
Text string `json:"text"`
|
||||
} `json:"translations"`
|
||||
}
|
||||
|
||||
// ErrorResponse 错误响应结构
|
||||
type ErrorResponse struct {
|
||||
Error struct {
|
||||
Code int `json:"code"`
|
||||
Message string `json:"message"`
|
||||
} `json:"error"`
|
||||
}
|
||||
|
||||
// 常量定义
|
||||
const (
|
||||
bingDefaultTimeout = 30 * time.Second
|
||||
bingTranslatorURL = "https://cn.bing.com/translator"
|
||||
bingTranslateAPIURL = "https://cn.bing.com/ttranslatev3"
|
||||
bingTokenURL = "https://edge.microsoft.com/translate/auth"
|
||||
bingTranslateAPIURL = "https://api-edge.cognitive.microsofttranslator.com/translate"
|
||||
tokenValidDuration = 25 * time.Minute // Token有效期,比实际30分钟稍短以确保安全
|
||||
)
|
||||
|
||||
// 错误定义
|
||||
var (
|
||||
ErrBingNetworkError = errors.New("bing translator network error")
|
||||
ErrBingParseError = errors.New("bing translator parse error")
|
||||
ErrBingTokenError = errors.New("failed to get bing translator token")
|
||||
ErrBingEmptyResponse = errors.New("empty response from bing translator")
|
||||
ErrBingRateLimit = errors.New("bing translator rate limit reached")
|
||||
)
|
||||
|
||||
// 用户代理列表
|
||||
var userAgents = []string{
|
||||
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36",
|
||||
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.0 Safari/605.1.15",
|
||||
"Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:124.0) Gecko/20100101 Firefox/124.0",
|
||||
"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36",
|
||||
}
|
||||
|
||||
// NewBingTranslator 创建一个新的Bing翻译器实例
|
||||
func NewBingTranslator() *BingTranslator {
|
||||
// 初始化随机数种子
|
||||
rand.New(rand.NewSource(time.Now().UnixNano()))
|
||||
|
||||
// 创建带Cookie存储的HTTP客户端
|
||||
jar, _ := cookiejar.New(nil)
|
||||
|
||||
translator := &BingTranslator{
|
||||
httpClient: &http.Client{
|
||||
Timeout: bingDefaultTimeout,
|
||||
// 启用Cookie存储
|
||||
Jar: jar,
|
||||
},
|
||||
Timeout: bingDefaultTimeout,
|
||||
session: &BingSession{
|
||||
Headers: make(map[string]string),
|
||||
Cookie: make(map[string]string),
|
||||
},
|
||||
Timeout: bingDefaultTimeout,
|
||||
languages: initBingLanguages(),
|
||||
}
|
||||
|
||||
// 初始化会话
|
||||
translator.refreshSession()
|
||||
|
||||
return translator
|
||||
}
|
||||
|
||||
@@ -91,129 +83,13 @@ func initBingLanguages() map[string]LanguageInfo {
|
||||
// 创建语言映射表
|
||||
languages := make(map[string]LanguageInfo)
|
||||
|
||||
// 添加所有支持的语言
|
||||
// 基于 Microsoft Translator 支持的语言列表
|
||||
// 参考: https://learn.microsoft.com/en-us/azure/ai-services/translator/language-support
|
||||
// 添加支持的语言
|
||||
// 基于 Microsoft Translator API 支持的语言列表
|
||||
// 参考: https://docs.microsoft.com/en-us/azure/cognitive-services/translator/language-support
|
||||
|
||||
// 常用语言
|
||||
languages["en"] = LanguageInfo{Code: "en", Name: "English"}
|
||||
languages["zh-Hans"] = LanguageInfo{Code: "zh-Hans", Name: "Chinese Simplified"}
|
||||
languages["zh-Hant"] = LanguageInfo{Code: "zh-Hant", Name: "Chinese Traditional"}
|
||||
languages["ja"] = LanguageInfo{Code: "ja", Name: "Japanese"}
|
||||
languages["ko"] = LanguageInfo{Code: "ko", Name: "Korean"}
|
||||
languages["fr"] = LanguageInfo{Code: "fr", Name: "French"}
|
||||
languages["fr-ca"] = LanguageInfo{Code: "fr-ca", Name: "French (Canada)"}
|
||||
languages["de"] = LanguageInfo{Code: "de", Name: "German"}
|
||||
languages["es"] = LanguageInfo{Code: "es", Name: "Spanish"}
|
||||
languages["ru"] = LanguageInfo{Code: "ru", Name: "Russian"}
|
||||
languages["pt"] = LanguageInfo{Code: "pt", Name: "Portuguese (Brazil)"}
|
||||
languages["pt-br"] = LanguageInfo{Code: "pt-br", Name: "Portuguese (Brazil)"}
|
||||
languages["pt-pt"] = LanguageInfo{Code: "pt-pt", Name: "Portuguese (Portugal)"}
|
||||
languages["it"] = LanguageInfo{Code: "it", Name: "Italian"}
|
||||
languages["ar"] = LanguageInfo{Code: "ar", Name: "Arabic"}
|
||||
|
||||
// 特殊语言
|
||||
languages["yue"] = LanguageInfo{Code: "yue", Name: "Cantonese (Traditional)"}
|
||||
languages["lzh"] = LanguageInfo{Code: "lzh", Name: "Chinese (Literary)"}
|
||||
|
||||
// 其他语言
|
||||
languages["af"] = LanguageInfo{Code: "af", Name: "Afrikaans"}
|
||||
languages["am"] = LanguageInfo{Code: "am", Name: "Amharic"}
|
||||
languages["as"] = LanguageInfo{Code: "as", Name: "Assamese"}
|
||||
languages["az"] = LanguageInfo{Code: "az", Name: "Azerbaijani (Latin)"}
|
||||
languages["ba"] = LanguageInfo{Code: "ba", Name: "Bashkir"}
|
||||
languages["bg"] = LanguageInfo{Code: "bg", Name: "Bulgarian"}
|
||||
languages["bn"] = LanguageInfo{Code: "bn", Name: "Bangla"}
|
||||
languages["bo"] = LanguageInfo{Code: "bo", Name: "Tibetan"}
|
||||
languages["bs"] = LanguageInfo{Code: "bs", Name: "Bosnian (Latin)"}
|
||||
languages["ca"] = LanguageInfo{Code: "ca", Name: "Catalan"}
|
||||
languages["cs"] = LanguageInfo{Code: "cs", Name: "Czech"}
|
||||
languages["cy"] = LanguageInfo{Code: "cy", Name: "Welsh"}
|
||||
languages["da"] = LanguageInfo{Code: "da", Name: "Danish"}
|
||||
languages["dv"] = LanguageInfo{Code: "dv", Name: "Divehi"}
|
||||
languages["el"] = LanguageInfo{Code: "el", Name: "Greek"}
|
||||
languages["et"] = LanguageInfo{Code: "et", Name: "Estonian"}
|
||||
languages["eu"] = LanguageInfo{Code: "eu", Name: "Basque"}
|
||||
languages["fa"] = LanguageInfo{Code: "fa", Name: "Persian"}
|
||||
languages["fi"] = LanguageInfo{Code: "fi", Name: "Finnish"}
|
||||
languages["fil"] = LanguageInfo{Code: "fil", Name: "Filipino"}
|
||||
languages["fj"] = LanguageInfo{Code: "fj", Name: "Fijian"}
|
||||
languages["fo"] = LanguageInfo{Code: "fo", Name: "Faroese"}
|
||||
languages["ga"] = LanguageInfo{Code: "ga", Name: "Irish"}
|
||||
languages["gl"] = LanguageInfo{Code: "gl", Name: "Galician"}
|
||||
languages["gu"] = LanguageInfo{Code: "gu", Name: "Gujarati"}
|
||||
languages["ha"] = LanguageInfo{Code: "ha", Name: "Hausa"}
|
||||
languages["he"] = LanguageInfo{Code: "he", Name: "Hebrew"}
|
||||
languages["hi"] = LanguageInfo{Code: "hi", Name: "Hindi"}
|
||||
languages["hr"] = LanguageInfo{Code: "hr", Name: "Croatian"}
|
||||
languages["ht"] = LanguageInfo{Code: "ht", Name: "Haitian Creole"}
|
||||
languages["hu"] = LanguageInfo{Code: "hu", Name: "Hungarian"}
|
||||
languages["hy"] = LanguageInfo{Code: "hy", Name: "Armenian"}
|
||||
languages["id"] = LanguageInfo{Code: "id", Name: "Indonesian"}
|
||||
languages["ig"] = LanguageInfo{Code: "ig", Name: "Igbo"}
|
||||
languages["is"] = LanguageInfo{Code: "is", Name: "Icelandic"}
|
||||
languages["ka"] = LanguageInfo{Code: "ka", Name: "Georgian"}
|
||||
languages["kk"] = LanguageInfo{Code: "kk", Name: "Kazakh"}
|
||||
languages["km"] = LanguageInfo{Code: "km", Name: "Khmer"}
|
||||
languages["kn"] = LanguageInfo{Code: "kn", Name: "Kannada"}
|
||||
languages["ku"] = LanguageInfo{Code: "ku", Name: "Kurdish (Arabic) (Central)"}
|
||||
languages["ky"] = LanguageInfo{Code: "ky", Name: "Kyrgyz (Cyrillic)"}
|
||||
languages["lo"] = LanguageInfo{Code: "lo", Name: "Lao"}
|
||||
languages["lt"] = LanguageInfo{Code: "lt", Name: "Lithuanian"}
|
||||
languages["lv"] = LanguageInfo{Code: "lv", Name: "Latvian"}
|
||||
languages["mg"] = LanguageInfo{Code: "mg", Name: "Malagasy"}
|
||||
languages["mi"] = LanguageInfo{Code: "mi", Name: "Maori"}
|
||||
languages["mk"] = LanguageInfo{Code: "mk", Name: "Macedonian"}
|
||||
languages["ml"] = LanguageInfo{Code: "ml", Name: "Malayalam"}
|
||||
languages["mn-Cyrl"] = LanguageInfo{Code: "mn-Cyrl", Name: "Mongolian (Cyrillic)"}
|
||||
languages["mr"] = LanguageInfo{Code: "mr", Name: "Marathi"}
|
||||
languages["ms"] = LanguageInfo{Code: "ms", Name: "Malay (Latin)"}
|
||||
languages["mt"] = LanguageInfo{Code: "mt", Name: "Maltese"}
|
||||
languages["mww"] = LanguageInfo{Code: "mww", Name: "Hmong Daw (Latin)"}
|
||||
languages["my"] = LanguageInfo{Code: "my", Name: "Myanmar (Burmese)"}
|
||||
languages["nb"] = LanguageInfo{Code: "nb", Name: "Norwegian Bokmål"}
|
||||
languages["ne"] = LanguageInfo{Code: "ne", Name: "Nepali"}
|
||||
languages["nl"] = LanguageInfo{Code: "nl", Name: "Dutch"}
|
||||
languages["or"] = LanguageInfo{Code: "or", Name: "Odia"}
|
||||
languages["otq"] = LanguageInfo{Code: "otq", Name: "Queretaro Otomi"}
|
||||
languages["pa"] = LanguageInfo{Code: "pa", Name: "Punjabi"}
|
||||
languages["pl"] = LanguageInfo{Code: "pl", Name: "Polish"}
|
||||
languages["prs"] = LanguageInfo{Code: "prs", Name: "Dari"}
|
||||
languages["ps"] = LanguageInfo{Code: "ps", Name: "Pashto"}
|
||||
languages["ro"] = LanguageInfo{Code: "ro", Name: "Romanian"}
|
||||
languages["rw"] = LanguageInfo{Code: "rw", Name: "Kinyarwanda"}
|
||||
languages["sk"] = LanguageInfo{Code: "sk", Name: "Slovak"}
|
||||
languages["sl"] = LanguageInfo{Code: "sl", Name: "Slovenian"}
|
||||
languages["sm"] = LanguageInfo{Code: "sm", Name: "Samoan (Latin)"}
|
||||
languages["sn"] = LanguageInfo{Code: "sn", Name: "chiShona"}
|
||||
languages["so"] = LanguageInfo{Code: "so", Name: "Somali"}
|
||||
languages["sq"] = LanguageInfo{Code: "sq", Name: "Albanian"}
|
||||
languages["sr-Cyrl"] = LanguageInfo{Code: "sr-Cyrl", Name: "Serbian (Cyrillic)"}
|
||||
languages["sr"] = LanguageInfo{Code: "sr", Name: "Serbian (Latin)"}
|
||||
languages["sr-latn"] = LanguageInfo{Code: "sr-latn", Name: "Serbian (Latin)"}
|
||||
languages["sv"] = LanguageInfo{Code: "sv", Name: "Swedish"}
|
||||
languages["sw"] = LanguageInfo{Code: "sw", Name: "Swahili (Latin)"}
|
||||
languages["ta"] = LanguageInfo{Code: "ta", Name: "Tamil"}
|
||||
languages["te"] = LanguageInfo{Code: "te", Name: "Telugu"}
|
||||
languages["th"] = LanguageInfo{Code: "th", Name: "Thai"}
|
||||
languages["ti"] = LanguageInfo{Code: "ti", Name: "Tigrinya"}
|
||||
languages["tk"] = LanguageInfo{Code: "tk", Name: "Turkmen (Latin)"}
|
||||
languages["tlh-Latn"] = LanguageInfo{Code: "tlh-Latn", Name: "Klingon"}
|
||||
languages["tlh-Piqd"] = LanguageInfo{Code: "tlh-Piqd", Name: "Klingon (plqaD)"}
|
||||
languages["to"] = LanguageInfo{Code: "to", Name: "Tongan"}
|
||||
languages["tr"] = LanguageInfo{Code: "tr", Name: "Turkish"}
|
||||
languages["tt"] = LanguageInfo{Code: "tt", Name: "Tatar (Latin)"}
|
||||
languages["ty"] = LanguageInfo{Code: "ty", Name: "Tahitian"}
|
||||
languages["ug"] = LanguageInfo{Code: "ug", Name: "Uyghur (Arabic)"}
|
||||
languages["uk"] = LanguageInfo{Code: "uk", Name: "Ukrainian"}
|
||||
languages["ur"] = LanguageInfo{Code: "ur", Name: "Urdu"}
|
||||
languages["uz"] = LanguageInfo{Code: "uz", Name: "Uzbek (Latin)"}
|
||||
languages["vi"] = LanguageInfo{Code: "vi", Name: "Vietnamese"}
|
||||
languages["yua"] = LanguageInfo{Code: "yua", Name: "Yucatec Maya"}
|
||||
languages["zu"] = LanguageInfo{Code: "zu", Name: "Zulu"}
|
||||
|
||||
// 添加一些特殊情况的映射
|
||||
languages["zh"] = LanguageInfo{Code: "zh-Hans", Name: "Chinese Simplified"} // 将zh映射到zh-Hans
|
||||
languages["zh-Hans"] = LanguageInfo{Code: "zh-Hans", Name: "Chinese (Simplified)"}
|
||||
languages["zh-Hant"] = LanguageInfo{Code: "zh-Hant", Name: "Chinese (Traditional)"}
|
||||
|
||||
return languages
|
||||
}
|
||||
@@ -224,91 +100,59 @@ func (t *BingTranslator) SetTimeout(timeout time.Duration) {
|
||||
t.httpClient.Timeout = timeout
|
||||
}
|
||||
|
||||
// getRandomUserAgent 获取随机用户代理
|
||||
func getRandomUserAgent() string {
|
||||
return userAgents[rand.Intn(len(userAgents))]
|
||||
}
|
||||
|
||||
// refreshSession 刷新翻译会话
|
||||
func (t *BingTranslator) refreshSession() error {
|
||||
// 设置随机用户代理
|
||||
userAgent := getRandomUserAgent()
|
||||
t.session.Headers["User-Agent"] = userAgent
|
||||
t.session.Headers["Referer"] = bingTranslatorURL
|
||||
t.session.Headers["Accept"] = "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8"
|
||||
t.session.Headers["Accept-Language"] = "en-US,en;q=0.5"
|
||||
t.session.Headers["Connection"] = "keep-alive"
|
||||
t.session.Headers["Upgrade-Insecure-Requests"] = "1"
|
||||
t.session.Headers["Cache-Control"] = "max-age=0"
|
||||
// getToken 获取访问token
|
||||
func (t *BingTranslator) getToken(ctx context.Context) (string, error) {
|
||||
// 检查token是否有效
|
||||
if t.token != nil && time.Now().Before(t.token.ExpiresAt) {
|
||||
return t.token.Value, nil
|
||||
}
|
||||
|
||||
// 创建请求
|
||||
req, err := http.NewRequest("GET", bingTranslatorURL, nil)
|
||||
req, err := http.NewRequestWithContext(ctx, "GET", bingTokenURL, nil)
|
||||
if err != nil {
|
||||
return fmt.Errorf("the creation request failed: %w", err)
|
||||
return "", fmt.Errorf("failed to create token request: %w", err)
|
||||
}
|
||||
|
||||
// 设置请求头
|
||||
for k, v := range t.session.Headers {
|
||||
req.Header.Set(k, v)
|
||||
}
|
||||
req.Header.Set("Accept", "*/*")
|
||||
req.Header.Set("Accept-Language", "zh-TW,zh;q=0.9,ja;q=0.8,zh-CN;q=0.7,en-US;q=0.6,en;q=0.5")
|
||||
req.Header.Set("Cache-Control", "no-cache")
|
||||
req.Header.Set("Pragma", "no-cache")
|
||||
req.Header.Set("Referrer-Policy", "strict-origin-when-cross-origin")
|
||||
req.Header.Set("Sec-Fetch-Dest", "empty")
|
||||
req.Header.Set("Sec-Fetch-Mode", "cors")
|
||||
req.Header.Set("Sec-Fetch-Site", "none")
|
||||
|
||||
// 发送请求
|
||||
resp, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
return fmt.Errorf("%w: %v", ErrBingNetworkError, err)
|
||||
return "", fmt.Errorf("failed to get token: %w", err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
// 保存Cookie
|
||||
for _, cookie := range resp.Cookies() {
|
||||
t.session.Cookie[cookie.Name] = cookie.Value
|
||||
// 检查状态码
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return "", fmt.Errorf("token request failed with status %d", resp.StatusCode)
|
||||
}
|
||||
|
||||
// 读取响应内容
|
||||
// 读取响应
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read response failed: %w", err)
|
||||
return "", fmt.Errorf("failed to read token response: %w", err)
|
||||
}
|
||||
|
||||
content := string(body)
|
||||
|
||||
// 提取参数
|
||||
// 1. 提取key和token
|
||||
paramsPattern := regexp.MustCompile(`params_AbusePreventionHelper\s*=\s*(\[.*?\]);`)
|
||||
paramsMatch := paramsPattern.FindStringSubmatch(content)
|
||||
if paramsMatch == nil || len(paramsMatch) < 2 {
|
||||
return fmt.Errorf("%w: params_AbusePreventionHelper could not be extracted", ErrBingTokenError)
|
||||
token := strings.TrimSpace(string(body))
|
||||
if token == "" {
|
||||
return "", fmt.Errorf("empty token received")
|
||||
}
|
||||
|
||||
// 解析参数数组
|
||||
paramsStr := paramsMatch[1]
|
||||
paramsStr = strings.ReplaceAll(paramsStr, "[", "")
|
||||
paramsStr = strings.ReplaceAll(paramsStr, "]", "")
|
||||
paramsParts := strings.Split(paramsStr, ",")
|
||||
|
||||
if len(paramsParts) < 2 {
|
||||
return fmt.Errorf("%w: params_AbusePreventionHelper format is incorrect", ErrBingTokenError)
|
||||
// 缓存token
|
||||
t.token = &tokenInfo{
|
||||
Value: token,
|
||||
ExpiresAt: time.Now().Add(tokenValidDuration),
|
||||
}
|
||||
|
||||
// 提取key和token
|
||||
t.session.Key = strings.Trim(paramsParts[0], `"' `)
|
||||
t.session.Token = strings.Trim(paramsParts[1], `"' `)
|
||||
|
||||
// 2. 提取IG值
|
||||
igPattern := regexp.MustCompile(`IG:"(\w+)"`)
|
||||
igMatch := igPattern.FindStringSubmatch(content)
|
||||
if igMatch == nil || len(igMatch) < 2 {
|
||||
return fmt.Errorf("%w: Unable to extract IG values", ErrBingTokenError)
|
||||
}
|
||||
|
||||
t.session.IG = igMatch[1]
|
||||
|
||||
// 更新会话头部
|
||||
t.session.Headers["IG"] = t.session.IG
|
||||
t.session.Headers["key"] = t.session.Key
|
||||
t.session.Headers["token"] = t.session.Token
|
||||
|
||||
return nil
|
||||
return token, nil
|
||||
}
|
||||
|
||||
// Translate 使用标准语言标签进行文本翻译
|
||||
@@ -328,139 +172,94 @@ func (t *BingTranslator) TranslateWithParams(text string, params TranslationPara
|
||||
|
||||
// translate 执行实际翻译操作
|
||||
func (t *BingTranslator) translate(text, from, to string) (string, error) {
|
||||
// 如果没有会话或关键参数缺失,刷新会话
|
||||
if t.session == nil || t.session.Token == "" || t.session.Key == "" || t.session.IG == "" {
|
||||
if err := t.refreshSession(); err != nil {
|
||||
return "", fmt.Errorf("the refresh session failed: %w", err)
|
||||
}
|
||||
if text == "" {
|
||||
return "", fmt.Errorf("text cannot be empty")
|
||||
}
|
||||
|
||||
// 生成随机IID
|
||||
randNum := rand.Intn(10) // 0-9的随机数
|
||||
iid := fmt.Sprintf("translator.5019.%d", 1+randNum%3) // 生成随机IID
|
||||
// 创建带超时的context
|
||||
ctx, cancel := context.WithTimeout(context.Background(), t.Timeout)
|
||||
defer cancel()
|
||||
|
||||
// 构建URL - 确保使用双&符号
|
||||
reqURL := fmt.Sprintf("%s?isVertical=1&&IG=%s&IID=%s",
|
||||
bingTranslateAPIURL, t.session.IG, iid)
|
||||
|
||||
// 标准化语言代码
|
||||
fromLang := t.GetStandardLanguageCode(from)
|
||||
toLang := t.GetStandardLanguageCode(to)
|
||||
|
||||
// 构建表单数据
|
||||
formData := url.Values{}
|
||||
formData.Set("fromLang", fromLang)
|
||||
formData.Set("text", text)
|
||||
formData.Set("to", toLang)
|
||||
formData.Set("token", t.session.Token)
|
||||
formData.Set("key", t.session.Key)
|
||||
|
||||
formDataStr := formData.Encode()
|
||||
|
||||
// 创建请求
|
||||
req, err := http.NewRequest("POST", reqURL, strings.NewReader(formDataStr))
|
||||
// 获取token
|
||||
token, err := t.getToken(ctx)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("The creation request failed: %w", err)
|
||||
return "", fmt.Errorf("failed to get token: %w", err)
|
||||
}
|
||||
|
||||
// 构建请求URL
|
||||
params := url.Values{}
|
||||
if from != "auto" {
|
||||
params.Set("from", from)
|
||||
}
|
||||
params.Set("to", to)
|
||||
params.Set("api-version", "3.0")
|
||||
params.Set("includeSentenceLength", "true")
|
||||
|
||||
fullURL := fmt.Sprintf("%s?%s", bingTranslateAPIURL, params.Encode())
|
||||
|
||||
// 构建请求体
|
||||
requestBody := []TranslateRequest{{Text: text}}
|
||||
jsonBody, err := json.Marshal(requestBody)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to marshal request: %w", err)
|
||||
}
|
||||
|
||||
// 创建HTTP请求
|
||||
req, err := http.NewRequestWithContext(ctx, "POST", fullURL, bytes.NewBuffer(jsonBody))
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to create request: %w", err)
|
||||
}
|
||||
|
||||
// 设置请求头
|
||||
req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
|
||||
req.Header.Set("User-Agent", t.session.Headers["User-Agent"])
|
||||
req.Header.Set("Referer", bingTranslatorURL)
|
||||
req.Header.Set("Accept", "*/*")
|
||||
req.Header.Set("Accept-Language", "en-US,en;q=0.5")
|
||||
req.Header.Set("Origin", "https://cn.bing.com")
|
||||
req.Header.Set("Connection", "keep-alive")
|
||||
req.Header.Set("X-Requested-With", "XMLHttpRequest")
|
||||
|
||||
// 添加Cookie
|
||||
for name, value := range t.session.Cookie {
|
||||
req.AddCookie(&http.Cookie{
|
||||
Name: name,
|
||||
Value: value,
|
||||
})
|
||||
}
|
||||
req.Header.Set("Accept", "application/json")
|
||||
req.Header.Set("Authorization", "Bearer "+token)
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
req.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36")
|
||||
|
||||
// 发送请求
|
||||
resp, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("%w: %v", ErrBingNetworkError, err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
// 读取响应
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
resp.Body.Close()
|
||||
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("read response failed: %w", err)
|
||||
return "", fmt.Errorf("failed to read response: %w", err)
|
||||
}
|
||||
|
||||
if len(body) == 0 {
|
||||
return "", ErrBingEmptyResponse
|
||||
// 检查HTTP状态码
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return "", fmt.Errorf("HTTP error %d: %s", resp.StatusCode, string(body))
|
||||
}
|
||||
|
||||
// 尝试解析响应
|
||||
var result interface{}
|
||||
if err := json.Unmarshal(body, &result); err != nil {
|
||||
// 首先尝试解析为错误响应
|
||||
var errorResp ErrorResponse
|
||||
if err := json.Unmarshal(body, &errorResp); err == nil && errorResp.Error.Code != 0 {
|
||||
return "", fmt.Errorf("translation error %d: %s", errorResp.Error.Code, errorResp.Error.Message)
|
||||
}
|
||||
|
||||
// 解析翻译响应
|
||||
var translateResp []TranslateResponse
|
||||
if err := json.Unmarshal(body, &translateResp); err != nil {
|
||||
return "", fmt.Errorf("%w: %v", ErrBingParseError, err)
|
||||
}
|
||||
|
||||
// 检查是否是字典类型
|
||||
if resultDict, ok := result.(map[string]interface{}); ok {
|
||||
// 检查是否需要验证码
|
||||
if _, hasCaptcha := resultDict["ShowCaptcha"]; hasCaptcha {
|
||||
return "", ErrBingRateLimit
|
||||
}
|
||||
|
||||
// 检查状态码
|
||||
if statusCode, hasStatus := resultDict["statusCode"]; hasStatus {
|
||||
if statusCode.(float64) == 400 {
|
||||
// 检查是否有错误消息
|
||||
if errorMsg, hasError := resultDict["errorMessage"]; hasError && errorMsg.(string) != "" {
|
||||
return "", fmt.Errorf("translation failed: %s", errorMsg)
|
||||
}
|
||||
// 如果没有明确的错误消息,可能是API变更或其他问题
|
||||
return "", fmt.Errorf("translation request failed (status code: 400)")
|
||||
} else if statusCode.(float64) == 429 {
|
||||
return "", ErrBingRateLimit
|
||||
}
|
||||
}
|
||||
|
||||
// 尝试从错误响应中提取详细信息
|
||||
if message, hasMessage := resultDict["message"]; hasMessage {
|
||||
return "", fmt.Errorf("translation failed: %v", message)
|
||||
}
|
||||
|
||||
// 尝试从响应中获取翻译结果
|
||||
if translations, hasTranslations := resultDict["translations"]; hasTranslations {
|
||||
if translationsArray, ok := translations.([]interface{}); ok && len(translationsArray) > 0 {
|
||||
if translation, ok := translationsArray[0].(map[string]interface{}); ok {
|
||||
if text, ok := translation["text"].(string); ok {
|
||||
return text, nil
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// 其他错误
|
||||
return "", fmt.Errorf("translation failed: %v", resultDict)
|
||||
if len(translateResp) == 0 || len(translateResp[0].Translations) == 0 {
|
||||
return "", ErrBingEmptyResponse
|
||||
}
|
||||
|
||||
// 应该是数组类型
|
||||
if resultArray, ok := result.([]interface{}); ok && len(resultArray) > 0 {
|
||||
firstItem := resultArray[0]
|
||||
if itemDict, ok := firstItem.(map[string]interface{}); ok {
|
||||
if translations, ok := itemDict["translations"].([]interface{}); ok && len(translations) > 0 {
|
||||
if translation, ok := translations[0].(map[string]interface{}); ok {
|
||||
if text, ok := translation["text"].(string); ok {
|
||||
return text, nil
|
||||
}
|
||||
}
|
||||
}
|
||||
// 合并所有翻译片段
|
||||
var result strings.Builder
|
||||
for i, translation := range translateResp[0].Translations {
|
||||
if i > 0 {
|
||||
result.WriteString(" ")
|
||||
}
|
||||
result.WriteString(translation.Text)
|
||||
}
|
||||
|
||||
return "", fmt.Errorf("%w: The response format is not as expected", ErrBingParseError)
|
||||
return result.String(), nil
|
||||
}
|
||||
|
||||
// GetSupportedLanguages 获取翻译器支持的语言列表
|
||||
@@ -473,11 +272,3 @@ func (t *BingTranslator) IsLanguageSupported(languageCode string) bool {
|
||||
_, exists := t.languages[languageCode]
|
||||
return exists
|
||||
}
|
||||
|
||||
// GetStandardLanguageCode 获取标准化的语言代码
|
||||
func (t *BingTranslator) GetStandardLanguageCode(languageCode string) string {
|
||||
if info, exists := t.languages[languageCode]; exists {
|
||||
return info.Code
|
||||
}
|
||||
return languageCode // 如果没有找到映射,返回原始代码
|
||||
}
|
||||
|
||||
@@ -112,40 +112,8 @@ func initDeeplLanguages() map[string]LanguageInfo {
|
||||
// 基于 DeepL API 支持的语言列表
|
||||
// 参考: https://developers.deepl.com/docs/resources/supported-languages
|
||||
|
||||
// 源语言和目标语言
|
||||
languages["ar"] = LanguageInfo{Code: "AR", Name: "Arabic"}
|
||||
languages["bg"] = LanguageInfo{Code: "BG", Name: "Bulgarian"}
|
||||
languages["cs"] = LanguageInfo{Code: "CS", Name: "Czech"}
|
||||
languages["da"] = LanguageInfo{Code: "DA", Name: "Danish"}
|
||||
languages["de"] = LanguageInfo{Code: "DE", Name: "German"}
|
||||
languages["el"] = LanguageInfo{Code: "EL", Name: "Greek"}
|
||||
// 源语言和目标语言 - 精简为中英互译
|
||||
languages["en"] = LanguageInfo{Code: "EN", Name: "English"}
|
||||
languages["en-gb"] = LanguageInfo{Code: "EN-GB", Name: "English (British)"}
|
||||
languages["en-us"] = LanguageInfo{Code: "EN-US", Name: "English (American)"}
|
||||
languages["es"] = LanguageInfo{Code: "ES", Name: "Spanish"}
|
||||
languages["et"] = LanguageInfo{Code: "ET", Name: "Estonian"}
|
||||
languages["fi"] = LanguageInfo{Code: "FI", Name: "Finnish"}
|
||||
languages["fr"] = LanguageInfo{Code: "FR", Name: "French"}
|
||||
languages["hu"] = LanguageInfo{Code: "HU", Name: "Hungarian"}
|
||||
languages["id"] = LanguageInfo{Code: "ID", Name: "Indonesian"}
|
||||
languages["it"] = LanguageInfo{Code: "IT", Name: "Italian"}
|
||||
languages["ja"] = LanguageInfo{Code: "JA", Name: "Japanese"}
|
||||
languages["ko"] = LanguageInfo{Code: "KO", Name: "Korean"}
|
||||
languages["lt"] = LanguageInfo{Code: "LT", Name: "Lithuanian"}
|
||||
languages["lv"] = LanguageInfo{Code: "LV", Name: "Latvian"}
|
||||
languages["nb"] = LanguageInfo{Code: "NB", Name: "Norwegian Bokmål"}
|
||||
languages["nl"] = LanguageInfo{Code: "NL", Name: "Dutch"}
|
||||
languages["pl"] = LanguageInfo{Code: "PL", Name: "Polish"}
|
||||
languages["pt"] = LanguageInfo{Code: "PT", Name: "Portuguese"}
|
||||
languages["pt-br"] = LanguageInfo{Code: "PT-BR", Name: "Portuguese (Brazilian)"}
|
||||
languages["pt-pt"] = LanguageInfo{Code: "PT-PT", Name: "Portuguese (Portugal)"}
|
||||
languages["ro"] = LanguageInfo{Code: "RO", Name: "Romanian"}
|
||||
languages["ru"] = LanguageInfo{Code: "RU", Name: "Russian"}
|
||||
languages["sk"] = LanguageInfo{Code: "SK", Name: "Slovak"}
|
||||
languages["sl"] = LanguageInfo{Code: "SL", Name: "Slovenian"}
|
||||
languages["sv"] = LanguageInfo{Code: "SV", Name: "Swedish"}
|
||||
languages["tr"] = LanguageInfo{Code: "TR", Name: "Turkish"}
|
||||
languages["uk"] = LanguageInfo{Code: "UK", Name: "Ukrainian"}
|
||||
languages["zh"] = LanguageInfo{Code: "ZH", Name: "Chinese"}
|
||||
|
||||
return languages
|
||||
@@ -157,11 +125,6 @@ func (t *DeeplTranslator) SetTimeout(timeout time.Duration) {
|
||||
t.httpClient.Timeout = timeout
|
||||
}
|
||||
|
||||
// SetDeeplHost 设置DeepL主机
|
||||
func (t *DeeplTranslator) SetDeeplHost(host string) {
|
||||
t.DeeplHost = host
|
||||
}
|
||||
|
||||
// Translate 使用标准语言标签进行文本翻译
|
||||
func (t *DeeplTranslator) Translate(text string, from language.Tag, to language.Tag) (string, error) {
|
||||
return t.translate(text, from.String(), to.String())
|
||||
@@ -319,9 +282,3 @@ func (t *DeeplTranslator) IsLanguageSupported(languageCode string) bool {
|
||||
_, ok := t.languages[strings.ToLower(languageCode)]
|
||||
return ok
|
||||
}
|
||||
|
||||
// GetStandardLanguageCode 获取标准化的语言代码
|
||||
func (t *DeeplTranslator) GetStandardLanguageCode(languageCode string) string {
|
||||
// 简单返回小写版本作为标准代码
|
||||
return strings.ToLower(languageCode)
|
||||
}
|
||||
|
||||
@@ -2,20 +2,17 @@
|
||||
package translator
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"log"
|
||||
"math"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"regexp"
|
||||
"strconv"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"github.com/robertkrimen/otto"
|
||||
"golang.org/x/text/language"
|
||||
)
|
||||
|
||||
@@ -24,405 +21,228 @@ var (
|
||||
ErrBadNetwork = errors.New("bad network, please check your internet connection")
|
||||
)
|
||||
|
||||
// GoogleTranslator Google翻译器结构体,统一管理翻译功能
|
||||
// 常量定义
|
||||
const (
|
||||
googleTranslateTKK = "448487.932609646" // 固定TKK值
|
||||
)
|
||||
|
||||
// GoogleTranslator 带token的Google翻译器(使用translate.google.com)
|
||||
type GoogleTranslator struct {
|
||||
GoogleHost string // Google服务主机
|
||||
vm *otto.Otto // JavaScript虚拟机
|
||||
ttk otto.Value // 翻译token缓存
|
||||
httpClient *http.Client // HTTP客户端
|
||||
Timeout time.Duration // 请求超时时间
|
||||
languages map[string]LanguageInfo // 支持的语言列表
|
||||
}
|
||||
|
||||
// NewGoogleTranslator 创建一个新的Google翻译器实例
|
||||
// GoogleTranslatorTokenFree 无token的Google翻译器(使用translate.googleapis.com)
|
||||
type GoogleTranslatorTokenFree struct {
|
||||
httpClient *http.Client // HTTP客户端
|
||||
Timeout time.Duration // 请求超时时间
|
||||
languages map[string]LanguageInfo // 支持的语言列表
|
||||
}
|
||||
|
||||
// NewGoogleTranslator 创建一个新的带token的Google翻译器实例
|
||||
func NewGoogleTranslator() *GoogleTranslator {
|
||||
translator := &GoogleTranslator{
|
||||
GoogleHost: "google.com",
|
||||
vm: otto.New(),
|
||||
Timeout: defaultTimeout,
|
||||
return &GoogleTranslator{
|
||||
httpClient: &http.Client{Timeout: defaultTimeout},
|
||||
Timeout: defaultTimeout,
|
||||
languages: initGoogleLanguages(),
|
||||
}
|
||||
|
||||
// 初始化ttk
|
||||
translator.ttk, _ = otto.ToValue("0")
|
||||
|
||||
return translator
|
||||
}
|
||||
|
||||
// initGoogleLanguages 初始化Google翻译器支持的语言列表
|
||||
func initGoogleLanguages() map[string]LanguageInfo {
|
||||
// 创建语言映射表
|
||||
languages := make(map[string]LanguageInfo)
|
||||
|
||||
// 添加所有支持的语言
|
||||
// 参考: https://cloud.google.com/translate/docs/languages
|
||||
|
||||
// 添加自动检测
|
||||
languages["auto"] = LanguageInfo{Code: "auto", Name: "Auto Detect"}
|
||||
|
||||
// 主要语言
|
||||
// 只支持三种语言
|
||||
languages["en"] = LanguageInfo{Code: "en", Name: "English"}
|
||||
languages["zh-cn"] = LanguageInfo{Code: "zh-CN", Name: "Chinese (Simplified)"}
|
||||
languages["zh-tw"] = LanguageInfo{Code: "zh-TW", Name: "Chinese (Traditional)"}
|
||||
languages["ja"] = LanguageInfo{Code: "ja", Name: "Japanese"}
|
||||
languages["ko"] = LanguageInfo{Code: "ko", Name: "Korean"}
|
||||
languages["fr"] = LanguageInfo{Code: "fr", Name: "French"}
|
||||
languages["de"] = LanguageInfo{Code: "de", Name: "German"}
|
||||
languages["es"] = LanguageInfo{Code: "es", Name: "Spanish"}
|
||||
languages["ru"] = LanguageInfo{Code: "ru", Name: "Russian"}
|
||||
languages["it"] = LanguageInfo{Code: "it", Name: "Italian"}
|
||||
languages["pt"] = LanguageInfo{Code: "pt", Name: "Portuguese"}
|
||||
|
||||
// 其他语言
|
||||
languages["af"] = LanguageInfo{Code: "af", Name: "Afrikaans"}
|
||||
languages["sq"] = LanguageInfo{Code: "sq", Name: "Albanian"}
|
||||
languages["am"] = LanguageInfo{Code: "am", Name: "Amharic"}
|
||||
languages["ar"] = LanguageInfo{Code: "ar", Name: "Arabic"}
|
||||
languages["hy"] = LanguageInfo{Code: "hy", Name: "Armenian"}
|
||||
languages["az"] = LanguageInfo{Code: "az", Name: "Azerbaijani"}
|
||||
languages["eu"] = LanguageInfo{Code: "eu", Name: "Basque"}
|
||||
languages["be"] = LanguageInfo{Code: "be", Name: "Belarusian"}
|
||||
languages["bn"] = LanguageInfo{Code: "bn", Name: "Bengali"}
|
||||
languages["bs"] = LanguageInfo{Code: "bs", Name: "Bosnian"}
|
||||
languages["bg"] = LanguageInfo{Code: "bg", Name: "Bulgarian"}
|
||||
languages["ca"] = LanguageInfo{Code: "ca", Name: "Catalan"}
|
||||
languages["ceb"] = LanguageInfo{Code: "ceb", Name: "Cebuano"}
|
||||
languages["zh"] = LanguageInfo{Code: "zh", Name: "Chinese"}
|
||||
languages["co"] = LanguageInfo{Code: "co", Name: "Corsican"}
|
||||
languages["hr"] = LanguageInfo{Code: "hr", Name: "Croatian"}
|
||||
languages["cs"] = LanguageInfo{Code: "cs", Name: "Czech"}
|
||||
languages["da"] = LanguageInfo{Code: "da", Name: "Danish"}
|
||||
languages["nl"] = LanguageInfo{Code: "nl", Name: "Dutch"}
|
||||
languages["eo"] = LanguageInfo{Code: "eo", Name: "Esperanto"}
|
||||
languages["et"] = LanguageInfo{Code: "et", Name: "Estonian"}
|
||||
languages["fi"] = LanguageInfo{Code: "fi", Name: "Finnish"}
|
||||
languages["fy"] = LanguageInfo{Code: "fy", Name: "Frisian"}
|
||||
languages["gl"] = LanguageInfo{Code: "gl", Name: "Galician"}
|
||||
languages["ka"] = LanguageInfo{Code: "ka", Name: "Georgian"}
|
||||
languages["el"] = LanguageInfo{Code: "el", Name: "Greek"}
|
||||
languages["gu"] = LanguageInfo{Code: "gu", Name: "Gujarati"}
|
||||
languages["ht"] = LanguageInfo{Code: "ht", Name: "Haitian Creole"}
|
||||
languages["ha"] = LanguageInfo{Code: "ha", Name: "Hausa"}
|
||||
languages["haw"] = LanguageInfo{Code: "haw", Name: "Hawaiian"}
|
||||
languages["he"] = LanguageInfo{Code: "he", Name: "Hebrew"}
|
||||
languages["hi"] = LanguageInfo{Code: "hi", Name: "Hindi"}
|
||||
languages["hmn"] = LanguageInfo{Code: "hmn", Name: "Hmong"}
|
||||
languages["hu"] = LanguageInfo{Code: "hu", Name: "Hungarian"}
|
||||
languages["is"] = LanguageInfo{Code: "is", Name: "Icelandic"}
|
||||
languages["ig"] = LanguageInfo{Code: "ig", Name: "Igbo"}
|
||||
languages["id"] = LanguageInfo{Code: "id", Name: "Indonesian"}
|
||||
languages["ga"] = LanguageInfo{Code: "ga", Name: "Irish"}
|
||||
languages["jw"] = LanguageInfo{Code: "jw", Name: "Javanese"}
|
||||
languages["kn"] = LanguageInfo{Code: "kn", Name: "Kannada"}
|
||||
languages["kk"] = LanguageInfo{Code: "kk", Name: "Kazakh"}
|
||||
languages["km"] = LanguageInfo{Code: "km", Name: "Khmer"}
|
||||
languages["ku"] = LanguageInfo{Code: "ku", Name: "Kurdish"}
|
||||
languages["ky"] = LanguageInfo{Code: "ky", Name: "Kyrgyz"}
|
||||
languages["lo"] = LanguageInfo{Code: "lo", Name: "Lao"}
|
||||
languages["la"] = LanguageInfo{Code: "la", Name: "Latin"}
|
||||
languages["lv"] = LanguageInfo{Code: "lv", Name: "Latvian"}
|
||||
languages["lt"] = LanguageInfo{Code: "lt", Name: "Lithuanian"}
|
||||
languages["lb"] = LanguageInfo{Code: "lb", Name: "Luxembourgish"}
|
||||
languages["mk"] = LanguageInfo{Code: "mk", Name: "Macedonian"}
|
||||
languages["mg"] = LanguageInfo{Code: "mg", Name: "Malagasy"}
|
||||
languages["ms"] = LanguageInfo{Code: "ms", Name: "Malay"}
|
||||
languages["ml"] = LanguageInfo{Code: "ml", Name: "Malayalam"}
|
||||
languages["mt"] = LanguageInfo{Code: "mt", Name: "Maltese"}
|
||||
languages["mi"] = LanguageInfo{Code: "mi", Name: "Maori"}
|
||||
languages["mr"] = LanguageInfo{Code: "mr", Name: "Marathi"}
|
||||
languages["mn"] = LanguageInfo{Code: "mn", Name: "Mongolian"}
|
||||
languages["my"] = LanguageInfo{Code: "my", Name: "Myanmar (Burmese)"}
|
||||
languages["ne"] = LanguageInfo{Code: "ne", Name: "Nepali"}
|
||||
languages["no"] = LanguageInfo{Code: "no", Name: "Norwegian"}
|
||||
languages["ny"] = LanguageInfo{Code: "ny", Name: "Nyanja (Chichewa)"}
|
||||
languages["ps"] = LanguageInfo{Code: "ps", Name: "Pashto"}
|
||||
languages["fa"] = LanguageInfo{Code: "fa", Name: "Persian"}
|
||||
languages["pl"] = LanguageInfo{Code: "pl", Name: "Polish"}
|
||||
languages["pt-br"] = LanguageInfo{Code: "pt-BR", Name: "Portuguese (Brazil)"}
|
||||
languages["pt-pt"] = LanguageInfo{Code: "pt-PT", Name: "Portuguese (Portugal)"}
|
||||
languages["pa"] = LanguageInfo{Code: "pa", Name: "Punjabi"}
|
||||
languages["ro"] = LanguageInfo{Code: "ro", Name: "Romanian"}
|
||||
languages["sm"] = LanguageInfo{Code: "sm", Name: "Samoan"}
|
||||
languages["gd"] = LanguageInfo{Code: "gd", Name: "Scots Gaelic"}
|
||||
languages["sr"] = LanguageInfo{Code: "sr", Name: "Serbian"}
|
||||
languages["st"] = LanguageInfo{Code: "st", Name: "Sesotho"}
|
||||
languages["sn"] = LanguageInfo{Code: "sn", Name: "Shona"}
|
||||
languages["sd"] = LanguageInfo{Code: "sd", Name: "Sindhi"}
|
||||
languages["si"] = LanguageInfo{Code: "si", Name: "Sinhala (Sinhalese)"}
|
||||
languages["sk"] = LanguageInfo{Code: "sk", Name: "Slovak"}
|
||||
languages["sl"] = LanguageInfo{Code: "sl", Name: "Slovenian"}
|
||||
languages["so"] = LanguageInfo{Code: "so", Name: "Somali"}
|
||||
languages["su"] = LanguageInfo{Code: "su", Name: "Sundanese"}
|
||||
languages["sw"] = LanguageInfo{Code: "sw", Name: "Swahili"}
|
||||
languages["sv"] = LanguageInfo{Code: "sv", Name: "Swedish"}
|
||||
languages["tl"] = LanguageInfo{Code: "tl", Name: "Tagalog (Filipino)"}
|
||||
languages["tg"] = LanguageInfo{Code: "tg", Name: "Tajik"}
|
||||
languages["ta"] = LanguageInfo{Code: "ta", Name: "Tamil"}
|
||||
languages["te"] = LanguageInfo{Code: "te", Name: "Telugu"}
|
||||
languages["th"] = LanguageInfo{Code: "th", Name: "Thai"}
|
||||
languages["tr"] = LanguageInfo{Code: "tr", Name: "Turkish"}
|
||||
languages["uk"] = LanguageInfo{Code: "uk", Name: "Ukrainian"}
|
||||
languages["ur"] = LanguageInfo{Code: "ur", Name: "Urdu"}
|
||||
languages["uz"] = LanguageInfo{Code: "uz", Name: "Uzbek"}
|
||||
languages["vi"] = LanguageInfo{Code: "vi", Name: "Vietnamese"}
|
||||
languages["cy"] = LanguageInfo{Code: "cy", Name: "Welsh"}
|
||||
languages["xh"] = LanguageInfo{Code: "xh", Name: "Xhosa"}
|
||||
languages["yi"] = LanguageInfo{Code: "yi", Name: "Yiddish"}
|
||||
languages["yo"] = LanguageInfo{Code: "yo", Name: "Yoruba"}
|
||||
languages["zu"] = LanguageInfo{Code: "zu", Name: "Zulu"}
|
||||
|
||||
return languages
|
||||
}
|
||||
|
||||
// generateToken 生成翻译token
|
||||
func generateToken(query string) string {
|
||||
// 实现TypeScript中的token生成逻辑
|
||||
tkkSplited := strings.Split(googleTranslateTKK, ".")
|
||||
tkkIndex, _ := strconv.Atoi(tkkSplited[0])
|
||||
tkkKey, _ := strconv.Atoi(tkkSplited[1])
|
||||
|
||||
// 转换查询字符串为字节数组
|
||||
bytesArray := transformQuery(query)
|
||||
|
||||
// 计算hash
|
||||
encodingRound := tkkIndex
|
||||
for _, b := range bytesArray {
|
||||
encodingRound += int(b)
|
||||
encodingRound = shiftLeftOrRightThenSumOrXor(encodingRound, "+-a^+6")
|
||||
}
|
||||
encodingRound = shiftLeftOrRightThenSumOrXor(encodingRound, "+-3^+b+-f")
|
||||
|
||||
encodingRound ^= tkkKey
|
||||
if encodingRound <= 0 {
|
||||
encodingRound = (encodingRound & 2147483647) + 2147483648
|
||||
}
|
||||
|
||||
normalizedResult := encodingRound % 1000000
|
||||
return fmt.Sprintf("%d.%d", normalizedResult, normalizedResult^tkkIndex)
|
||||
}
|
||||
|
||||
// transformQuery 转换查询字符串
|
||||
func transformQuery(query string) []byte {
|
||||
var bytesArray []byte
|
||||
runes := []rune(query)
|
||||
|
||||
for i := 0; i < len(runes); i++ {
|
||||
charCode := int(runes[i])
|
||||
|
||||
if charCode < 128 {
|
||||
bytesArray = append(bytesArray, byte(charCode))
|
||||
} else if charCode < 2048 {
|
||||
bytesArray = append(bytesArray, byte((charCode>>6)|192))
|
||||
bytesArray = append(bytesArray, byte((charCode&63)|128))
|
||||
} else {
|
||||
if (charCode&64512) == 55296 && i+1 < len(runes) && (int(runes[i+1])&64512) == 56320 {
|
||||
charCode = 65536 + ((charCode & 1023) << 10) + (int(runes[i+1]) & 1023)
|
||||
i++
|
||||
bytesArray = append(bytesArray, byte((charCode>>18)|240))
|
||||
bytesArray = append(bytesArray, byte(((charCode>>12)&63)|128))
|
||||
} else {
|
||||
bytesArray = append(bytesArray, byte((charCode>>12)|224))
|
||||
}
|
||||
bytesArray = append(bytesArray, byte(((charCode>>6)&63)|128))
|
||||
bytesArray = append(bytesArray, byte((charCode&63)|128))
|
||||
}
|
||||
}
|
||||
|
||||
return bytesArray
|
||||
}
|
||||
|
||||
// shiftLeftOrRightThenSumOrXor 位运算操作
|
||||
func shiftLeftOrRightThenSumOrXor(num int, optString string) int {
|
||||
for i := 0; i < len(optString)-2; i += 3 {
|
||||
acc := int(optString[i+2])
|
||||
if acc >= 'a' {
|
||||
acc = acc - 87
|
||||
} else {
|
||||
acc = acc - '0'
|
||||
}
|
||||
|
||||
if optString[i+1] == '+' {
|
||||
acc = num >> uint(acc)
|
||||
} else {
|
||||
acc = num << uint(acc)
|
||||
}
|
||||
|
||||
if optString[i] == '+' {
|
||||
num = (num + acc) & 4294967295
|
||||
} else {
|
||||
num ^= acc
|
||||
}
|
||||
}
|
||||
return num
|
||||
}
|
||||
|
||||
// SetTimeout 设置请求超时时间
|
||||
func (t *GoogleTranslator) SetTimeout(timeout time.Duration) {
|
||||
t.Timeout = timeout
|
||||
t.httpClient.Timeout = timeout
|
||||
}
|
||||
|
||||
// SetGoogleHost 设置Google主机
|
||||
func (t *GoogleTranslator) SetGoogleHost(host string) {
|
||||
t.GoogleHost = host
|
||||
}
|
||||
|
||||
// Translate 使用Go语言提供的标准语言标签进行文本翻译
|
||||
func (t *GoogleTranslator) Translate(text string, from language.Tag, to language.Tag) (string, error) {
|
||||
return t.translate(text, from.String(), to.String(), false)
|
||||
return t.translate(text, from.String(), to.String())
|
||||
}
|
||||
|
||||
// TranslateWithParams 使用简单字符串参数进行文本翻译
|
||||
func (t *GoogleTranslator) TranslateWithParams(text string, params TranslationParams) (string, error) {
|
||||
// 设置超时时间(如果有指定)
|
||||
if params.Timeout > 0 {
|
||||
t.SetTimeout(params.Timeout)
|
||||
}
|
||||
|
||||
return t.translate(text, params.From, params.To, true)
|
||||
return t.translate(text, params.From, params.To)
|
||||
}
|
||||
|
||||
// translate 执行实际翻译操作
|
||||
func (t *GoogleTranslator) translate(text, from, to string, withVerification bool) (string, error) {
|
||||
if withVerification {
|
||||
if _, err := language.Parse(from); err != nil && from != "auto" {
|
||||
log.Println("[WARNING], '" + from + "' is a invalid language, switching to 'auto'")
|
||||
from = "auto"
|
||||
}
|
||||
if _, err := language.Parse(to); err != nil {
|
||||
log.Println("[WARNING], '" + to + "' is a invalid language, switching to 'en'")
|
||||
to = "en"
|
||||
}
|
||||
// translate 执行实际翻译操作(带token版本)
|
||||
func (t *GoogleTranslator) translate(text, from, to string) (string, error) {
|
||||
ctx, cancel := context.WithTimeout(context.Background(), t.Timeout)
|
||||
defer cancel()
|
||||
|
||||
// 生成token
|
||||
token := generateToken(text)
|
||||
|
||||
// 构建请求URL
|
||||
apiURL := "https://translate.google.com/translate_a/single"
|
||||
params := url.Values{}
|
||||
params.Set("client", "t")
|
||||
params.Set("sl", from)
|
||||
params.Set("tl", to)
|
||||
params.Set("hl", to)
|
||||
params.Set("ie", "UTF-8")
|
||||
params.Set("oe", "UTF-8")
|
||||
params.Set("otf", "1")
|
||||
params.Set("ssel", "0")
|
||||
params.Set("tsel", "0")
|
||||
params.Set("kc", "7")
|
||||
params.Set("q", text)
|
||||
params.Set("tk", token)
|
||||
|
||||
// 添加dt参数
|
||||
dtParams := []string{"at", "bd", "ex", "ld", "md", "qca", "rw", "rm", "ss", "t"}
|
||||
for _, dt := range dtParams {
|
||||
params.Add("dt", dt)
|
||||
}
|
||||
|
||||
textValue, _ := otto.ToValue(text)
|
||||
urlStr := fmt.Sprintf("https://translate.%s/translate_a/single", t.GoogleHost)
|
||||
token := t.getToken(textValue)
|
||||
fullURL := apiURL + "?" + params.Encode()
|
||||
|
||||
data := map[string]string{
|
||||
"client": "gtx",
|
||||
"sl": from,
|
||||
"tl": to,
|
||||
"hl": to,
|
||||
"ie": "UTF-8",
|
||||
"oe": "UTF-8",
|
||||
"otf": "1",
|
||||
"ssel": "0",
|
||||
"tsel": "0",
|
||||
"kc": "7",
|
||||
"q": text,
|
||||
}
|
||||
|
||||
u, err := url.Parse(urlStr)
|
||||
// 创建请求
|
||||
req, err := http.NewRequestWithContext(ctx, "GET", fullURL, nil)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
parameters := url.Values{}
|
||||
for k, v := range data {
|
||||
parameters.Add(k, v)
|
||||
}
|
||||
for _, v := range []string{"at", "bd", "ex", "ld", "md", "qca", "rw", "rm", "ss", "t"} {
|
||||
parameters.Add("dt", v)
|
||||
}
|
||||
|
||||
parameters.Add("tk", token)
|
||||
u.RawQuery = parameters.Encode()
|
||||
|
||||
req, err := http.NewRequest("GET", u.String(), nil)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
r, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
if errors.Is(err, http.ErrHandlerTimeout) {
|
||||
return "", ErrBadNetwork
|
||||
}
|
||||
return "", err
|
||||
}
|
||||
|
||||
if r.StatusCode != http.StatusOK {
|
||||
return "", fmt.Errorf("API error: status code %d", r.StatusCode)
|
||||
}
|
||||
|
||||
raw, err := io.ReadAll(r.Body)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
defer r.Body.Close()
|
||||
|
||||
var resp []interface{}
|
||||
err = json.Unmarshal(raw, &resp)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
responseText := ""
|
||||
for _, obj := range resp[0].([]interface{}) {
|
||||
if len(obj.([]interface{})) == 0 {
|
||||
break
|
||||
}
|
||||
|
||||
t, ok := obj.([]interface{})[0].(string)
|
||||
if ok {
|
||||
responseText += t
|
||||
}
|
||||
}
|
||||
|
||||
return responseText, nil
|
||||
}
|
||||
|
||||
// getToken 获取翻译API所需的token
|
||||
func (t *GoogleTranslator) getToken(text otto.Value) string {
|
||||
ttk, err := t.updateTTK()
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
|
||||
tk, err := t.generateToken(text, ttk)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
|
||||
return strings.Replace(tk.String(), "&tk=", "", -1)
|
||||
}
|
||||
|
||||
// updateTTK 更新TTK值
|
||||
func (t *GoogleTranslator) updateTTK() (otto.Value, error) {
|
||||
timestamp := time.Now().UnixNano() / 3600000
|
||||
now := math.Floor(float64(timestamp))
|
||||
ttk, err := strconv.ParseFloat(t.ttk.String(), 64)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
}
|
||||
|
||||
if ttk == now {
|
||||
return t.ttk, nil
|
||||
}
|
||||
|
||||
req, err := http.NewRequest("GET", fmt.Sprintf("https://translate.%s", t.GoogleHost), nil)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
}
|
||||
|
||||
// 发送请求
|
||||
resp, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
return "", ErrBadNetwork
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return "", fmt.Errorf("API error: status code %d", resp.StatusCode)
|
||||
}
|
||||
|
||||
// 读取响应
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
return "", err
|
||||
}
|
||||
|
||||
matches := regexp.MustCompile(`tkk:\s?'(.+?)'`).FindStringSubmatch(string(body))
|
||||
if len(matches) > 0 {
|
||||
v, err := otto.ToValue(matches[0])
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
}
|
||||
t.ttk = v
|
||||
return v, nil
|
||||
// 解析JSON响应
|
||||
var result []interface{}
|
||||
if err := json.Unmarshal(body, &result); err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
return t.ttk, nil
|
||||
}
|
||||
|
||||
// generateToken 生成翻译API所需的token
|
||||
func (t *GoogleTranslator) generateToken(a otto.Value, TTK otto.Value) (otto.Value, error) {
|
||||
err := t.vm.Set("x", a)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
if len(result) == 0 {
|
||||
return "", errors.New("unexpected response format")
|
||||
}
|
||||
|
||||
_ = t.vm.Set("internalTTK", TTK)
|
||||
// 提取翻译文本
|
||||
translations, ok := result[0].([]interface{})
|
||||
if !ok {
|
||||
return "", errors.New("unexpected response format")
|
||||
}
|
||||
|
||||
result, err := t.vm.Run(`
|
||||
function sM(a) {
|
||||
var b;
|
||||
if (null !== yr)
|
||||
b = yr;
|
||||
else {
|
||||
b = wr(String.fromCharCode(84));
|
||||
var c = wr(String.fromCharCode(75));
|
||||
b = [b(), b()];
|
||||
b[1] = c();
|
||||
b = (yr = window[b.join(c())] || "") || ""
|
||||
}
|
||||
var d = wr(String.fromCharCode(116))
|
||||
, c = wr(String.fromCharCode(107))
|
||||
, d = [d(), d()];
|
||||
d[1] = c();
|
||||
c = "&" + d.join("") + "=";
|
||||
d = b.split(".");
|
||||
b = Number(d[0]) || 0;
|
||||
for (var e = [], f = 0, g = 0; g < a.length; g++) {
|
||||
var l = a.charCodeAt(g);
|
||||
128 > l ? e[f++] = l : (2048 > l ? e[f++] = l >> 6 | 192 : (55296 == (l & 64512) && g + 1 < a.length && 56320 == (a.charCodeAt(g + 1) & 64512) ? (l = 65536 + ((l & 1023) << 10) + (a.charCodeAt(++g) & 1023),
|
||||
e[f++] = l >> 18 | 240,
|
||||
e[f++] = l >> 12 & 63 | 128) : e[f++] = l >> 12 | 224,
|
||||
e[f++] = l >> 6 & 63 | 128),
|
||||
e[f++] = l & 63 | 128)
|
||||
}
|
||||
a = b;
|
||||
for (f = 0; f < e.length; f++)
|
||||
a += e[f],
|
||||
a = xr(a, "+-a^+6");
|
||||
a = xr(a, "+-3^+b+-f");
|
||||
a ^= Number(d[1]) || 0;
|
||||
0 > a && (a = (a & 2147483647) + 2147483648);
|
||||
a %= 1E6;
|
||||
return c + (a.toString() + "." + (a ^ b))
|
||||
}
|
||||
|
||||
var yr = null;
|
||||
var wr = function(a) {
|
||||
return function() {
|
||||
return a
|
||||
var translatedText strings.Builder
|
||||
for _, translation := range translations {
|
||||
if chunk, ok := translation.([]interface{}); ok && len(chunk) > 0 {
|
||||
if text, ok := chunk[0].(string); ok {
|
||||
translatedText.WriteString(text)
|
||||
}
|
||||
}
|
||||
, xr = function(a, b) {
|
||||
for (var c = 0; c < b.length - 2; c += 3) {
|
||||
var d = b.charAt(c + 2)
|
||||
, d = "a" <= d ? d.charCodeAt(0) - 87 : Number(d)
|
||||
, d = "+" == b.charAt(c + 1) ? a >>> d : a << d;
|
||||
a = "+" == b.charAt(c) ? a + d & 4294967295 : a ^ d
|
||||
}
|
||||
return a
|
||||
};
|
||||
|
||||
var window = {
|
||||
TKK: internalTTK
|
||||
};
|
||||
|
||||
sM(x)
|
||||
`)
|
||||
if err != nil {
|
||||
return otto.UndefinedValue(), err
|
||||
}
|
||||
|
||||
return result, nil
|
||||
return translatedText.String(), nil
|
||||
}
|
||||
|
||||
// GetSupportedLanguages 获取翻译器支持的语言列表
|
||||
@@ -436,8 +256,24 @@ func (t *GoogleTranslator) IsLanguageSupported(languageCode string) bool {
|
||||
return ok
|
||||
}
|
||||
|
||||
// GetStandardLanguageCode 获取标准化的语言代码
|
||||
func (t *GoogleTranslator) GetStandardLanguageCode(languageCode string) string {
|
||||
// 简单返回小写版本作为标准代码
|
||||
return strings.ToLower(languageCode)
|
||||
// visitArrayItems 递归访问数组项
|
||||
func visitArrayItems(arr []interface{}, visitor func(interface{})) {
|
||||
for _, obj := range arr {
|
||||
if subArr, ok := obj.([]interface{}); ok {
|
||||
visitArrayItems(subArr, visitor)
|
||||
} else {
|
||||
visitor(obj)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// GetSupportedLanguages 获取翻译器支持的语言列表
|
||||
func (t *GoogleTranslatorTokenFree) GetSupportedLanguages() map[string]LanguageInfo {
|
||||
return t.languages
|
||||
}
|
||||
|
||||
// IsLanguageSupported 检查指定的语言代码是否受支持
|
||||
func (t *GoogleTranslatorTokenFree) IsLanguageSupported(languageCode string) bool {
|
||||
_, ok := t.languages[strings.ToLower(languageCode)]
|
||||
return ok
|
||||
}
|
||||
|
||||
103
internal/common/translator/google_translator_free.go
Normal file
103
internal/common/translator/google_translator_free.go
Normal file
@@ -0,0 +1,103 @@
|
||||
package translator
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"golang.org/x/text/language"
|
||||
"io"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"time"
|
||||
)
|
||||
|
||||
// NewGoogleTranslatorTokenFree 创建一个新的无token的Google翻译器实例
|
||||
func NewGoogleTranslatorTokenFree() *GoogleTranslatorTokenFree {
|
||||
return &GoogleTranslatorTokenFree{
|
||||
httpClient: &http.Client{Timeout: defaultTimeout},
|
||||
Timeout: defaultTimeout,
|
||||
languages: initGoogleLanguages(),
|
||||
}
|
||||
}
|
||||
|
||||
// SetTimeout 设置请求超时时间
|
||||
func (t *GoogleTranslatorTokenFree) SetTimeout(timeout time.Duration) {
|
||||
t.Timeout = timeout
|
||||
t.httpClient.Timeout = timeout
|
||||
}
|
||||
|
||||
// Translate 使用Go语言提供的标准语言标签进行文本翻译
|
||||
func (t *GoogleTranslatorTokenFree) Translate(text string, from language.Tag, to language.Tag) (string, error) {
|
||||
return t.translate(text, from.String(), to.String())
|
||||
}
|
||||
|
||||
// TranslateWithParams 使用简单字符串参数进行文本翻译
|
||||
func (t *GoogleTranslatorTokenFree) TranslateWithParams(text string, params TranslationParams) (string, error) {
|
||||
if params.Timeout > 0 {
|
||||
t.SetTimeout(params.Timeout)
|
||||
}
|
||||
return t.translate(text, params.From, params.To)
|
||||
}
|
||||
|
||||
// translate 执行实际翻译操作(无token版本)
|
||||
func (t *GoogleTranslatorTokenFree) translate(text, from, to string) (string, error) {
|
||||
ctx, cancel := context.WithTimeout(context.Background(), t.Timeout)
|
||||
defer cancel()
|
||||
|
||||
// 构建请求URL(无token版本)
|
||||
apiURL := "https://translate.googleapis.com/translate_a/t"
|
||||
params := url.Values{}
|
||||
params.Set("client", "dict-chrome-ex")
|
||||
params.Set("sl", from)
|
||||
params.Set("tl", to)
|
||||
params.Set("q", text)
|
||||
|
||||
fullURL := apiURL + "?" + params.Encode()
|
||||
|
||||
// 创建请求
|
||||
req, err := http.NewRequestWithContext(ctx, "GET", fullURL, nil)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
|
||||
|
||||
// 发送请求
|
||||
resp, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
return "", ErrBadNetwork
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return "", fmt.Errorf("API error: status code %d", resp.StatusCode)
|
||||
}
|
||||
|
||||
// 读取响应
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
// 解析JSON响应
|
||||
var result []interface{}
|
||||
if err := json.Unmarshal(body, &result); err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
// 提取翻译文本
|
||||
var translatedTexts []string
|
||||
visitArrayItems(result, func(obj interface{}) {
|
||||
if text, ok := obj.(string); ok {
|
||||
translatedTexts = append(translatedTexts, text)
|
||||
}
|
||||
})
|
||||
|
||||
if len(translatedTexts) == 0 {
|
||||
return "", errors.New("no translation found")
|
||||
}
|
||||
|
||||
// 返回第一个翻译结果
|
||||
return translatedTexts[0], nil
|
||||
}
|
||||
210
internal/common/translator/tartunlp_translator.go
Normal file
210
internal/common/translator/tartunlp_translator.go
Normal file
@@ -0,0 +1,210 @@
|
||||
// Package translator 提供文本翻译功能
|
||||
package translator
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"context"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"golang.org/x/text/language"
|
||||
)
|
||||
|
||||
// TartuNLPTranslator TartuNLP翻译器结构体
|
||||
type TartuNLPTranslator struct {
|
||||
httpClient *http.Client // HTTP客户端
|
||||
Timeout time.Duration // 请求超时时间
|
||||
languages map[string]LanguageInfo // 支持的语言列表
|
||||
}
|
||||
|
||||
// 常量定义
|
||||
const (
|
||||
tartuNLPDefaultTimeout = 300 * time.Second // 默认超时时间300秒
|
||||
tartuNLPAPIURL = "https://api.tartunlp.ai/translation/v2" // TartuNLP API地址
|
||||
tartuNLPLengthLimit = 5000 // 长度限制5000字符
|
||||
)
|
||||
|
||||
// 错误定义
|
||||
var (
|
||||
ErrTartuNLPNetworkError = errors.New("tartunlp translator network error")
|
||||
ErrTartuNLPUnsupportedLang = errors.New("tartunlp translator unsupported language")
|
||||
ErrTartuNLPResponseError = errors.New("tartunlp translator response error")
|
||||
ErrTartuNLPLengthExceeded = errors.New("tartunlp translator text length exceeded")
|
||||
)
|
||||
|
||||
// TartuNLPRequest TartuNLP请求结构体
|
||||
type TartuNLPRequest struct {
|
||||
Text []string `json:"text"` // 要翻译的文本数组
|
||||
Src string `json:"src"` // 源语言
|
||||
Tgt string `json:"tgt"` // 目标语言
|
||||
}
|
||||
|
||||
// TartuNLPResponse TartuNLP响应结构体
|
||||
type TartuNLPResponse struct {
|
||||
Result []string `json:"result"` // 翻译结果数组
|
||||
}
|
||||
|
||||
// NewTartuNLPTranslator 创建一个新的TartuNLP翻译器实例
|
||||
func NewTartuNLPTranslator() *TartuNLPTranslator {
|
||||
translator := &TartuNLPTranslator{
|
||||
httpClient: &http.Client{
|
||||
Timeout: tartuNLPDefaultTimeout,
|
||||
},
|
||||
Timeout: tartuNLPDefaultTimeout,
|
||||
languages: initTartuNLPLanguages(),
|
||||
}
|
||||
|
||||
return translator
|
||||
}
|
||||
|
||||
// initTartuNLPLanguages 初始化TartuNLP翻译器支持的语言列表
|
||||
func initTartuNLPLanguages() map[string]LanguageInfo {
|
||||
// 创建语言映射表
|
||||
languages := make(map[string]LanguageInfo)
|
||||
|
||||
// 添加支持的语言
|
||||
// 基于 TartuNLP API 支持的语言列表
|
||||
languages["en"] = LanguageInfo{Code: "en", Name: "English"}
|
||||
languages["et"] = LanguageInfo{Code: "et", Name: "Estonian"}
|
||||
languages["de"] = LanguageInfo{Code: "de", Name: "German"}
|
||||
languages["lt"] = LanguageInfo{Code: "lt", Name: "Lithuanian"}
|
||||
languages["lv"] = LanguageInfo{Code: "lv", Name: "Latvian"}
|
||||
languages["fi"] = LanguageInfo{Code: "fi", Name: "Finnish"}
|
||||
languages["ru"] = LanguageInfo{Code: "ru", Name: "Russian"}
|
||||
languages["no"] = LanguageInfo{Code: "no", Name: "Norwegian"}
|
||||
languages["hu"] = LanguageInfo{Code: "hu", Name: "Hungarian"}
|
||||
languages["se"] = LanguageInfo{Code: "se", Name: "Swedish"}
|
||||
|
||||
return languages
|
||||
}
|
||||
|
||||
// SetTimeout 设置请求超时时间
|
||||
func (t *TartuNLPTranslator) SetTimeout(timeout time.Duration) {
|
||||
t.Timeout = timeout
|
||||
t.httpClient.Timeout = timeout
|
||||
}
|
||||
|
||||
// Translate 使用标准语言标签进行文本翻译
|
||||
func (t *TartuNLPTranslator) Translate(text string, from language.Tag, to language.Tag) (string, error) {
|
||||
return t.translate(text, from.String(), to.String())
|
||||
}
|
||||
|
||||
// TranslateWithParams 使用简单字符串参数进行文本翻译
|
||||
func (t *TartuNLPTranslator) TranslateWithParams(text string, params TranslationParams) (string, error) {
|
||||
// 设置超时时间(如果有指定)
|
||||
if params.Timeout > 0 {
|
||||
t.SetTimeout(params.Timeout)
|
||||
}
|
||||
|
||||
return t.translate(text, params.From, params.To)
|
||||
}
|
||||
|
||||
// checkLengthLimit 检查文本长度是否超出限制
|
||||
func (t *TartuNLPTranslator) checkLengthLimit(text string) error {
|
||||
if len(text) > tartuNLPLengthLimit {
|
||||
return fmt.Errorf("%w: text length %d exceeds limit %d", ErrTartuNLPLengthExceeded, len(text), tartuNLPLengthLimit)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// translate 执行实际翻译操作
|
||||
func (t *TartuNLPTranslator) translate(text, from, to string) (string, error) {
|
||||
if text == "" {
|
||||
return "", fmt.Errorf("text cannot be empty")
|
||||
}
|
||||
|
||||
// 检查文本长度限制
|
||||
if err := t.checkLengthLimit(text); err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
// 转换语言代码为TartuNLP格式
|
||||
fromLower := strings.ToLower(from)
|
||||
toLower := strings.ToLower(to)
|
||||
|
||||
// 验证源语言支持
|
||||
if _, ok := t.languages[fromLower]; !ok {
|
||||
return "", fmt.Errorf("%w: source language '%s' not supported by TartuNLP", ErrTartuNLPUnsupportedLang, from)
|
||||
}
|
||||
|
||||
// 验证目标语言支持
|
||||
if _, ok := t.languages[toLower]; !ok {
|
||||
return "", fmt.Errorf("%w: target language '%s' not supported by TartuNLP", ErrTartuNLPUnsupportedLang, to)
|
||||
}
|
||||
|
||||
// 创建带超时的context
|
||||
ctx, cancel := context.WithTimeout(context.Background(), t.Timeout)
|
||||
defer cancel()
|
||||
|
||||
// 构建请求体
|
||||
request := TartuNLPRequest{
|
||||
Text: []string{text},
|
||||
Src: fromLower,
|
||||
Tgt: toLower,
|
||||
}
|
||||
|
||||
// 序列化请求
|
||||
jsonData, err := json.Marshal(request)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to marshal request: %w", err)
|
||||
}
|
||||
|
||||
// 创建HTTP请求
|
||||
req, err := http.NewRequestWithContext(ctx, "POST", tartuNLPAPIURL, bytes.NewBuffer(jsonData))
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to create request: %w", err)
|
||||
}
|
||||
|
||||
// 设置请求头
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
req.Header.Set("Accept", "application/json")
|
||||
req.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36")
|
||||
|
||||
// 发送请求
|
||||
resp, err := t.httpClient.Do(req)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("%w: %v", ErrTartuNLPNetworkError, err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
// 检查HTTP状态码
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
body, _ := io.ReadAll(resp.Body)
|
||||
return "", fmt.Errorf("%w: HTTP %d - %s", ErrTartuNLPResponseError, resp.StatusCode, string(body))
|
||||
}
|
||||
|
||||
// 读取响应体
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to read response body: %w", err)
|
||||
}
|
||||
|
||||
// 解析响应
|
||||
var response TartuNLPResponse
|
||||
if err := json.Unmarshal(body, &response); err != nil {
|
||||
return "", fmt.Errorf("%w: failed to parse response: %v", ErrTartuNLPResponseError, err)
|
||||
}
|
||||
|
||||
// 检查响应结果
|
||||
if len(response.Result) == 0 {
|
||||
return "", fmt.Errorf("%w: empty translation result", ErrTartuNLPResponseError)
|
||||
}
|
||||
|
||||
return response.Result[0], nil
|
||||
}
|
||||
|
||||
// GetSupportedLanguages 获取翻译器支持的语言列表
|
||||
func (t *TartuNLPTranslator) GetSupportedLanguages() map[string]LanguageInfo {
|
||||
return t.languages
|
||||
}
|
||||
|
||||
// IsLanguageSupported 检查指定的语言代码是否受支持
|
||||
func (t *TartuNLPTranslator) IsLanguageSupported(languageCode string) bool {
|
||||
_, exists := t.languages[strings.ToLower(languageCode)]
|
||||
return exists
|
||||
}
|
||||
@@ -32,6 +32,8 @@ const (
|
||||
YoudaoTranslatorType TranslatorType = "youdao"
|
||||
// DeeplTranslatorType DeepL翻译器
|
||||
DeeplTranslatorType TranslatorType = "deepl"
|
||||
// TartuNLPTranslatorType TartuNLP翻译器
|
||||
TartuNLPTranslatorType TranslatorType = "tartunlp"
|
||||
)
|
||||
|
||||
// LanguageInfo 语言信息结构体
|
||||
@@ -56,9 +58,6 @@ type Translator interface {
|
||||
|
||||
// IsLanguageSupported 检查指定的语言代码是否受支持
|
||||
IsLanguageSupported(languageCode string) bool
|
||||
|
||||
// GetStandardLanguageCode 获取标准化的语言代码
|
||||
GetStandardLanguageCode(languageCode string) string
|
||||
}
|
||||
|
||||
// TranslatorFactory 翻译器工厂,用于创建不同类型的翻译器
|
||||
@@ -80,6 +79,8 @@ func (f *TranslatorFactory) Create(translatorType TranslatorType) (Translator, e
|
||||
return NewYoudaoTranslator(), nil
|
||||
case DeeplTranslatorType:
|
||||
return NewDeeplTranslator(), nil
|
||||
case TartuNLPTranslatorType:
|
||||
return NewTartuNLPTranslator(), nil
|
||||
default:
|
||||
return nil, fmt.Errorf("unsupported translator type: %s", translatorType)
|
||||
}
|
||||
|
||||
@@ -50,8 +50,7 @@ func initYoudaoLanguages() map[string]LanguageInfo {
|
||||
// 创建语言映射表
|
||||
languages := make(map[string]LanguageInfo)
|
||||
|
||||
// 自动检测
|
||||
languages["auto"] = LanguageInfo{Code: "auto", Name: "Auto"}
|
||||
languages["auto"] = LanguageInfo{Code: "AUTO", Name: "Auto"}
|
||||
|
||||
return languages
|
||||
}
|
||||
@@ -64,8 +63,7 @@ func (t *YoudaoTranslator) SetTimeout(timeout time.Duration) {
|
||||
|
||||
// Translate 使用标准语言标签进行文本翻译
|
||||
func (t *YoudaoTranslator) Translate(text string, from language.Tag, to language.Tag) (string, error) {
|
||||
// 有道翻译不需要指定源语言和目标语言,它会自动检测
|
||||
return t.translate(text)
|
||||
return t.translate(text, to.String())
|
||||
}
|
||||
|
||||
// TranslateWithParams 使用简单字符串参数进行文本翻译
|
||||
@@ -75,16 +73,15 @@ func (t *YoudaoTranslator) TranslateWithParams(text string, params TranslationPa
|
||||
t.SetTimeout(params.Timeout)
|
||||
}
|
||||
|
||||
// 有道翻译不需要指定源语言和目标语言,它会自动检测
|
||||
return t.translate(text)
|
||||
return t.translate(text, params.To)
|
||||
}
|
||||
|
||||
// translate 执行实际翻译操作
|
||||
func (t *YoudaoTranslator) translate(text string) (string, error) {
|
||||
func (t *YoudaoTranslator) translate(text string, typeName string) (string, error) {
|
||||
// 构建表单数据
|
||||
form := url.Values{}
|
||||
form.Add("inputtext", text)
|
||||
form.Add("type", "AUTO")
|
||||
form.Add("type", typeName)
|
||||
|
||||
// 创建请求
|
||||
req, err := http.NewRequest("POST", youdaoTranslateURL, strings.NewReader(form.Encode()))
|
||||
|
||||
@@ -89,14 +89,15 @@ func (s *TranslationService) TranslateWith(text string, from string, to string,
|
||||
return trans.TranslateWithParams(text, params)
|
||||
}
|
||||
|
||||
// GetAvailableTranslators 获取所有可用翻译器类型
|
||||
// GetTranslators 获取所有可用翻译器类型
|
||||
// @returns {[]string} 翻译器类型列表
|
||||
func (s *TranslationService) GetAvailableTranslators() []string {
|
||||
func (s *TranslationService) GetTranslators() []string {
|
||||
return []string{
|
||||
string(translator.GoogleTranslatorType),
|
||||
string(translator.BingTranslatorType),
|
||||
string(translator.GoogleTranslatorType),
|
||||
string(translator.YoudaoTranslatorType),
|
||||
string(translator.DeeplTranslatorType),
|
||||
string(translator.TartuNLPTranslatorType),
|
||||
}
|
||||
}
|
||||
|
||||
@@ -122,12 +123,3 @@ func (s *TranslationService) IsLanguageSupported(translatorType translator.Trans
|
||||
}
|
||||
return translator.IsLanguageSupported(languageCode)
|
||||
}
|
||||
|
||||
// GetStandardLanguageCode 获取标准化的语言代码
|
||||
func (s *TranslationService) GetStandardLanguageCode(translatorType translator.TranslatorType, languageCode string) string {
|
||||
translator, err := s.getTranslator(translatorType)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
return translator.GetStandardLanguageCode(languageCode)
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user