Add SpeechRecognition service source and chat/social voice transcription (#294)

Co-Authored-By: Taromati2 <taromati2@outlook.com>

Author: steve02081504

AGENTS.md

@@ -15,6 +15,7 @@
 - **APIs/Types**: `@src/decl/` (`CharAPI_t` → `charAPI.ts`). Consult for required methods.
 - **Key structs**: `prompt_struct_t` (`@src/decl/prompt_struct.ts`), `chatMetadata_t` (`@src/public/parts/shells/chat/src/chat/session/models.mjs`).
 - **Registries**: `fount.json` → `registries: [{ id, level, path }]`; `GET /api/registries/:name`; helpers: `@src/server/registries.mjs` (backend), `@src/public/pages/scripts/registries.mjs` (frontend).
+- **Service sources**: types under `serviceSources/{AI,search,translate,SpeechRecognition}/` + generators under `serviceGenerators/...`. `SpeechRecognitionSource_t.Recognize({ audio | feed, onResult })` — stream-first feed, buffer convenience; endpoints only on chat (`/speechRecognition/recognize` + `/speechRecognition/session`), social reuses them; config probe via part list / defaults.
 
 ## Dev Guidelines
 

src/decl/SpeechRecognitionSource.ts

@@ -0,0 +1,75 @@
+import { info_t, locale_t } from './basedefs.ts'
+
+/**
+ * 语音识别服务源类型定义。
+ */
+
+/** 增量识别结果。 */
+export class SpeechRecognitionPartial_t {
+	/** 当前累计全文（每次回调都是从头开始的完整文本，非增量片段） */
+	text!: string
+	/** 是否为最终结果片段 */
+	isFinal?: boolean
+}
+
+/** 一次识别会话的最终结果。 */
+export class SpeechRecognitionResult_t {
+	/** 最终文本 */
+	text!: string
+	/** 检测到的语言（若有） */
+	language?: string
+}
+
+/** 实时喂音频控制面。 */
+export class SpeechRecognitionFeedControl_t {
+	/**
+	 * 推送一帧音频（PCM 或源约定格式）。
+	 * @param chunk 音频字节
+	 */
+	send!: (chunk: Uint8Array) => Promise<void>
+	/**
+	 * 声明输入结束并等待收尾。
+	 */
+	end!: () => Promise<void>
+}
+
+/** Recognize 选项。 */
+export class SpeechRecognitionOptions_t {
+	language?: string
+	hotwords?: string[]
+	signal?: AbortSignal
+	/** 实时 / 增量结果回调 */
+	onResult?: (partial: SpeechRecognitionPartial_t) => void
+	/**
+	 * 便捷：一次传入完整音频。与 `feed` 二选一。
+	 */
+	audio?: { buffer: Uint8Array, mime_type: string, name?: string }
+	/**
+	 * 实时输入：在回调里 `await send(chunk)`，结束时 `await end()`。与 `audio` 二选一。
+	 */
+	feed?: (ctl: SpeechRecognitionFeedControl_t) => Promise<void>
+}
+
+/**
+ * 语音识别数据源接口。
+ */
+export class SpeechRecognitionSource_t {
+	filename!: string
+	type!: 'speech-recognition' | string
+	info!: info_t<{
+		provider: string;
+	}>
+	is_paid!: boolean
+	extension!: object
+
+	Unload?: () => Promise<void>
+	/**
+	 * 执行识别。`audio` 与 `feed` 二选一；流式源在 `feed`/`send` 时立即上行并经 `onResult` 出字。
+	 */
+	Recognize!: (options: SpeechRecognitionOptions_t) => Promise<SpeechRecognitionResult_t>
+	interfaces!: {
+		info?: {
+			UpdateInfo: (locales: locale_t[]) => Promise<info_t>,
+		},
+	}
+}

src/decl/SpeechRecognitionSourceGenerator.ts

@@ -0,0 +1,29 @@
+import { locale_t, info_t } from './basedefs.ts'
+import { SpeechRecognitionSource_t } from './SpeechRecognitionSource.ts'
+
+/**
+ * 语音识别数据源生成器接口。
+ */
+export class SpeechRecognitionSourceGenerator_t {
+	info!: info_t
+
+	Init?: () => Promise<void>
+	Load?: () => Promise<void>
+	Unload?: () => Promise<void>
+	Uninstall?: () => Promise<void>
+
+	interfaces!: {
+		info?: {
+			UpdateInfo: (locales: locale_t[]) => Promise<info_t>,
+		},
+		config?: {
+			GetData: () => Promise<any>
+			SetData: (data: any) => Promise<void>
+		},
+		serviceGenerator: {
+			GetConfigDisplayContent: () => Promise<{ html?: string, js?: string }>
+			GetConfigTemplate: () => Promise<any>
+			GetSource: (config: any, args: { username: string, SaveConfig: () => Promise<void> }) => Promise<SpeechRecognitionSource_t>
+		}
+	}
+}

src/decl/locale_data.ts

@@ -929,6 +929,24 @@ export type LocaleData = {
 						}
 					}
 				}
+				'generators/SpeechRecognition': {
+					title: string
+					subtitle: string
+					card: {
+						defaultCheckbox: {
+							title: string
+						}
+					}
+				}
+				'sources/SpeechRecognition': {
+					title: string
+					subtitle: string
+					card: {
+						defaultCheckbox: {
+							title: string
+						}
+					}
+				}
 			}
 		}
 		escapeConfirm: string
@@ -2818,6 +2836,8 @@ export type LocaleData = {
 		}
 		voiceRecording: {
 			errorAccessingMicrophone: string
+			confirmSpeechRecognition: string
+			speechRecognitionFailed: string
 		}
 		attachment: {
 			buttons: {
@@ -2833,6 +2853,12 @@ export type LocaleData = {
 				deleteIcon: {
 					alt: string
 				}
+				more: {
+					title: string
+				}
+				recognize: {
+					title: string
+				}
 			}
 		}
 		charCard: {
@@ -4248,6 +4274,16 @@ export type LocaleData = {
 					label: string
 				}
 			}
+			voiceButton: {
+				title: string
+				'aria-label': string
+			}
+			voiceFailed: string
+			audioTranscript: {
+				placeholder: string
+			}
+			confirmSpeechRecognition: string
+			speechRecognitionFailed: string
 		}
 		feed: {
 			refresh: {
@@ -6029,6 +6065,7 @@ export type LocaleKeyParams = {
 	'chat.stickers.authorLabel': { author: string | number }
 	'chat.stickers.sticker.count': { count: string | number }
 	'chat.typingIndicator.isTyping': { names: string | number }
+	'chat.voiceRecording.speechRecognitionFailed': { error: string | number }
 	'deskpet.toasts.start_failed': { charname: string | number; message: string | number }
 	'deskpet.toasts.started': { charname: string | number }
 	'deskpet.toasts.stop_failed': { charname: string | number; message: string | number }
@@ -6231,6 +6268,8 @@ export type LocaleKeyParams = {
 	'social.actions.repostFailed': { error: string | number }
 	'social.actions.saveFailed': { error: string | number }
 	'social.bootstrapFailed': { error: string | number }
+	'social.composer.speechRecognitionFailed': { error: string | number }
+	'social.composer.voiceFailed': { error: string | number }
 	'social.connectNodeFailed': { error: string | number }
 	'social.drafts.deleteFailed': { error: string | number }
 	'social.drafts.loadFailed': { error: string | number }

src/public/locales/ar-SA.json

@@ -960,6 +960,24 @@
 							"title": "تعيين كمصدر افتراضي لخدمة الترجمة"
 						}
 					}
+				},
+				"generators/SpeechRecognition": {
+					"title": "مولّد خدمة التعرف على الكلام",
+					"subtitle": "حوّل الموجات الصوتية إلى كلمات — ما تلفظه يُكتب أمامك.",
+					"card": {
+						"defaultCheckbox": {
+							"title": "تعيين كمولّد خدمة التعرف على الكلام الافتراضي"
+						}
+					}
+				},
+				"sources/SpeechRecognition": {
+					"title": "مصدر خدمة التعرف على الكلام",
+					"subtitle": "نافذة تسمع الأصوات، حيث يبدأ المعنى.",
+					"card": {
+						"defaultCheckbox": {
+							"title": "تعيين كمصدر خدمة التعرف على الكلام الافتراضي"
+						}
+					}
 				}
 			}
 		},
@@ -2849,7 +2867,9 @@
 			}
 		},
 		"voiceRecording": {
-			"errorAccessingMicrophone": "لا يمكن الوصول إلى الميكروفون"
+			"errorAccessingMicrophone": "لا يمكن الوصول إلى الميكروفون",
+			"confirmSpeechRecognition": "هل تريد ملء النص تلقائياً بالتعرف على الكلام؟ سيُحفظ مرفق التسجيل.",
+			"speechRecognitionFailed": "فشل التعرف على الكلام: ${error}"
 		},
 		"attachment": {
 			"buttons": {
@@ -2864,6 +2884,12 @@
 				},
 				"deleteIcon": {
 					"alt": "أيقونة الحذف"
+				},
+				"more": {
+					"title": "أكثر"
+				},
+				"recognize": {
+					"title": "تحويل الصوت إلى نص"
 				}
 			}
 		},
@@ -4279,7 +4305,17 @@
 					"followers7d": "متابعون لأكثر من 7 أيام",
 					"label": "من يستطيع الرد"
 				}
-			}
+			},
+			"voiceButton": {
+				"title": "تسجيل",
+				"aria-label": "تسجيل"
+			},
+			"voiceFailed": "فشل التسجيل: ${error}",
+			"audioTranscript": {
+				"placeholder": "نص التفريغ أو ملاحظات"
+			},
+			"confirmSpeechRecognition": "هل تريد ملء المحتوى تلقائياً بالتعرف على الكلام؟ سيُحفظ مرفق التسجيل.",
+			"speechRecognitionFailed": "فشل التعرف على الكلام: ${error}"
 		},
 		"feed": {
 			"refresh": {

src/public/locales/de-DE.json

@@ -960,6 +960,24 @@
 							"title": "Als Standard-Übersetzungsdienstquelle festlegen"
 						}
 					}
+				},
+				"generators/SpeechRecognition": {
+					"title": "Generator für Spracherkennungsdienste",
+					"subtitle": "Verwandle Schallwellen in Schrift — gesprochenes Wort, geschriebene Zeile.",
+					"card": {
+						"defaultCheckbox": {
+							"title": "Als Standardgenerator für Spracherkennungsdienste festlegen"
+						}
+					}
+				},
+				"sources/SpeechRecognition": {
+					"title": "Quelle des Spracherkennungsdienstes",
+					"subtitle": "Ein Fenster zum Hören von Stimme und Klang — wo Bedeutung beginnt.",
+					"card": {
+						"defaultCheckbox": {
+							"title": "Als Standardquelle für Spracherkennungsdienste festlegen"
+						}
+					}
 				}
 			}
 		},
@@ -2849,7 +2867,9 @@
 			}
 		},
 		"voiceRecording": {
-			"errorAccessingMicrophone": "Fehler beim Zugriff auf das Mikrofon."
+			"errorAccessingMicrophone": "Fehler beim Zugriff auf das Mikrofon.",
+			"confirmSpeechRecognition": "Text per Spracherkennung automatisch ausfüllen? Die Audioaufnahme bleibt als Anhang erhalten.",
+			"speechRecognitionFailed": "Spracherkennung fehlgeschlagen: ${error}"
 		},
 		"attachment": {
 			"buttons": {
@@ -2864,6 +2884,12 @@
 				},
 				"deleteIcon": {
 					"alt": "Löschen-Symbol"
+				},
+				"more": {
+					"title": "Mehr"
+				},
+				"recognize": {
+					"title": "Sprache in Text"
 				}
 			}
 		},
@@ -4279,7 +4305,17 @@
 					"followers7d": "Follower seit mehr als 7 Tagen",
 					"label": "Wer kann antworten?"
 				}
-			}
+			},
+			"voiceButton": {
+				"title": "Aufnehmen",
+				"aria-label": "Aufnehmen"
+			},
+			"voiceFailed": "Aufnahme fehlgeschlagen: ${error}",
+			"audioTranscript": {
+				"placeholder": "Transkript oder Notizen"
+			},
+			"confirmSpeechRecognition": "Beitrag per Spracherkennung automatisch ausfüllen? Die Audioaufnahme bleibt als Anhang erhalten.",
+			"speechRecognitionFailed": "Spracherkennung fehlgeschlagen: ${error}"
 		},
 		"feed": {
 			"refresh": {

src/public/locales/emoji.json

@@ -960,6 +960,24 @@
 							"title": "📌🌐🚰"
 						}
 					}
+				},
+				"generators/SpeechRecognition": {
+					"title": "🎙️⚙️✨",
+					"subtitle": "🎙️➡️📝",
+					"card": {
+						"defaultCheckbox": {
+							"title": "📌🎙️⚙️✨"
+						}
+					}
+				},
+				"sources/SpeechRecognition": {
+					"title": "🎙️🚰",
+					"subtitle": "👂🎙️➡️💬",
+					"card": {
+						"defaultCheckbox": {
+							"title": "📌🎙️🚰"
+						}
+					}
 				}
 			}
 		},
@@ -2849,7 +2867,9 @@
 			}
 		},
 		"voiceRecording": {
-			"errorAccessingMicrophone": "🚫🎤"
+			"errorAccessingMicrophone": "🚫🎤",
+			"confirmSpeechRecognition": "🎙️➡️📝❓🔊📎✅",
+			"speechRecognitionFailed": "🎙️❌：${error}"
 		},
 		"attachment": {
 			"buttons": {
@@ -2864,6 +2884,12 @@
 				},
 				"deleteIcon": {
 					"alt": "🗑️"
+				},
+				"more": {
+					"title": "⋯"
+				},
+				"recognize": {
+					"title": "🎙️➡️📝"
 				}
 			}
 		},
@@ -4279,7 +4305,17 @@
 					"followers7d": "7️⃣📅➕👀",
 					"label": "💬🔐"
 				}
-			}
+			},
+			"voiceButton": {
+				"title": "🎤",
+				"aria-label": "🎤"
+			},
+			"voiceFailed": "🎤❌：${error}",
+			"audioTranscript": {
+				"placeholder": "📝💬"
+			},
+			"confirmSpeechRecognition": "🎙️➡️📝❓🔊📎✅",
+			"speechRecognitionFailed": "🎙️❌：${error}"
 		},
 		"feed": {
 			"refresh": {