You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

175 lines
5.2 KiB

package echomeet
const (
MaxShortAudioProcess = 10
MaxLongAudioProcess = 10
MaxAIAwaitProcess = 10
TaskTypeShortAudioAwait = "Echomeet:short_audio_await"
TaskTypeShortAudioProcess = "Echomeet:short_audio_process"
TaskTypeAIAwait = "Echomeet:ai_await"
TaskTypeAIAwaitProcess = "Echomeet:ai_process"
// TranscribeQueryMinInterval 主动查询第三方转写服务的最小间隔(秒)
TranscribeQueryMinInterval int64 = 10
// TranscribePollConcurrency 轮询第三方查询的并发度,避免 for 串行被单条慢请求拖住整轮
TranscribePollConcurrency = 10
)
// toByteDanceLang 将内部 BCP-47 语言码转换为字节跳动 audomodel 支持的格式。
// 不在支持列表内返回空字符串,调用方应路由到其他引擎。
var bcp47ToByteDance = map[string]string{
"": "",
"zh-CN": "zh", "zh-TW": "zh", "zh-HK": "zh",
"en-US": "en-US", "en-GB": "en-US",
"ja-JP": "ja-JP",
"ko-KR": "ko-KR",
"id-ID": "id-ID",
"es-MX": "es-MX", "es-ES": "es-MX",
"pt-BR": "pt-BR", "pt-PT": "pt-BR",
"de-DE": "de-DE",
"fr-FR": "fr-FR",
"fil-PH": "fil-PH",
"ms-MY": "ms-MY",
"th-TH": "th-TH",
"ar-SA": "ar-SA",
}
func toByteDanceLang(bcp47 string) (lang string, supported bool) {
lang, supported = bcp47ToByteDance[bcp47]
return
}
// toDashScopeLang 将内部 BCP-47 语言码转换为百炼 language_hints 的短码格式。
// 返回空字符串表示该语言不传 hints,由引擎自动检测。
//
// ⚠️ 这张表只做「格式转换」,不代表模型支持该语种——阿里两套模型支持范围不同
// (qwen-audio-3.0-asr-flash-filetrans 27 种 / paraformer-v2 仅 8 种),
// 真正的支持性校验在 sys/aliyun/filetrans 的 modelLangs,选路过滤则由后台 languages 白名单负责。
var bcp47ToDashScope = map[string]string{
"zh-CN": "zh", "zh-TW": "zh", "zh-HK": "zh",
"yue-HK": "yue",
"en-US": "en", "en-GB": "en",
"ja-JP": "ja",
"ko-KR": "ko",
"ru-RU": "ru", "ru-UA": "ru",
"vi-VN": "vi",
"tr-TR": "tr",
"pl-PL": "pl",
"nl-NL": "nl",
"it-IT": "it",
"uk-UA": "uk",
"de-DE": "de",
"fr-FR": "fr",
"es-ES": "es", "es-MX": "es",
"pt-BR": "pt", "pt-PT": "pt",
"ar-SA": "ar",
"id-ID": "id",
"ms-MY": "ms",
"th-TH": "th",
"fil-PH": "fil",
}
// toDashScopeLang 返回 DashScope language_hints 值,空字符串表示自动检测
func toDashScopeLang(bcp47 string) string {
return bcp47ToDashScope[bcp47] // 不在 map 中自动返回 ""
}
// ============== 翻译平台语言码转换 ==============
// 输入统一是客户端的 BCP-47 码(如 zh-CN / en-US / ja-JP);
// 每个平台有自己的接受格式,逐家做映射。未识别一律返回原值,让平台自己决定是否报错,
// 绝对不要 fallback 成 "zh"——那会让"未知语言→中文"出现在调用栈深处难以发现。
// 火山翻译 / 阿里 MT:短码 zh / en / ja / ...
var bcp47ToShortLang = map[string]string{
"zh-CN": "zh", "zh-TW": "zh", "zh-HK": "zh",
"en-US": "en", "en-GB": "en",
"ja-JP": "ja",
"ko-KR": "ko",
"id-ID": "id",
"es-MX": "es", "es-ES": "es",
"pt-BR": "pt", "pt-PT": "pt",
"de-DE": "de",
"fr-FR": "fr",
"fil-PH": "fil",
"ms-MY": "ms",
"th-TH": "th",
"ar-SA": "ar",
"ru-RU": "ru", "ru-UA": "ru",
"vi-VN": "vi",
"tr-TR": "tr",
"pl-PL": "pl",
"nl-NL": "nl",
"it-IT": "it",
"uk-UA": "uk",
}
// 火山翻译:BCP-47 → 短码(zh / en / ja ...)
func toByteDanceTranslateLang(bcp47 string) string {
if v, ok := bcp47ToShortLang[bcp47]; ok {
return v
}
return bcp47
}
// 阿里 MT:与火山相同的短码体系
func toAliTranslateLang(bcp47 string) string {
if v, ok := bcp47ToShortLang[bcp47]; ok {
return v
}
return bcp47
}
// 说明:实时翻译接口(modules/user/api_translate.go)用的是 comm.BCP47ToShortLang,
// 与本表同源。本函数保持原样不动——echomeet 在跑生产,不为共用去改它的行为。
// Google Cloud Translation:直接接受 BCP-47(en / en-US / zh-CN 都支持),原样透传
func toGoogleTranslateLang(bcp47 string) string {
return bcp47
}
// Azure Translator:中文要 zh-Hans / zh-Hant,其他用短 ISO(en / ja / ko ...)
var bcp47ToMicrosoft = map[string]string{
"zh-CN": "zh-Hans", "zh-TW": "zh-Hant", "zh-HK": "zh-Hant",
"en-US": "en", "en-GB": "en",
"ja-JP": "ja",
"ko-KR": "ko",
"id-ID": "id",
"es-MX": "es", "es-ES": "es",
"pt-BR": "pt", "pt-PT": "pt",
"de-DE": "de",
"fr-FR": "fr",
"fil-PH": "fil",
"ms-MY": "ms",
"th-TH": "th",
"ar-SA": "ar",
"ru-RU": "ru", "ru-UA": "ru",
"vi-VN": "vi",
"tr-TR": "tr",
"pl-PL": "pl",
"nl-NL": "nl",
"it-IT": "it",
"uk-UA": "uk",
}
func toMicrosoftTranslateLang(bcp47 string) string {
if v, ok := bcp47ToMicrosoft[bcp47]; ok {
return v
}
return bcp47
}
// isByteDanceSupported 判断 BCP-47 语言是否被字节跳动支持
func isByteDanceSupported(bcp47 string) bool {
_, ok := bcp47ToByteDance[bcp47]
return ok
}
// baseLang 取 BCP-47 的基础语言子标签(首个 '-' 之前),如 es-MX→es、zh-TW→zh。
// 无区域后缀(如 "es")或空串则原样返回。仅用于模板查找的区域回退,不参与各平台语言码映射。
func baseLang(bcp47 string) string {
for i := 0; i < len(bcp47); i++ {
if bcp47[i] == '-' {
return bcp47[:i]
}
}
return bcp47
}