dsh-audiogen 0.4.3 → 0.4.5

package/lib/index.js

@@ -24,6 +24,8 @@ const SETTINGS_API = {
 const GENERATE_API = "/api/dsh-audiogen/generate";
 /** Loopback-only task cancellation route (aborts the host-side upstream call). */
 const TASK_API = { cancel: "/api/dsh-audiogen/task/cancel" };
+/** Loopback-only prompt enhancement route (uses the agent's default model). */
+const ENHANCE_API = "/api/dsh-audiogen/prompt/enhance";
 /** Host-mediated built-in provider catalog (channels the user can instantiate). */
 const PRESETS_API = "/api/dsh-audiogen/presets";
 /** Host-mediated model/voice discovery endpoint. */
@@ -864,6 +866,55 @@ async function generateAudio(channel, request, signal) {
 	return genericAudio(channel, request, signal);
 }
 //#endregion
+//#region src/prompt-enhance.ts
+/** 按生成模式给出增强指令（系统提示）。 */
+function instructionsFor(mode) {
+	return [
+		"你是一个音频提示词增强助手。用户给出一个粗略的音频生成需求，",
+		"请将其扩写为一段可直接提交给音频生成模型的中文或英文描述。",
+		"只输出增强后的描述本身，不要输出任何解释、前后缀、引号或代码块。",
+		"保持用户原始意图，不要改变其核心内容；为最终生成的音频服务。",
+		"描述控制在 200-600 字左右。"
+	].join("") + {
+		tts: "这是文本转语音（TTS）任务：让文本更适合朗读——口语化、自然、带合适的情感标签（如 (laughs)、(whisper)），避免生僻多音字和超长句，可适当补足上下文使语句完整，但不要改写原意。",
+		music: "这是音乐生成任务：扩写音乐风格/情绪/乐器/结构/节奏变化/氛围，使用音频模型熟悉的描述词汇（如 cinematic orchestral、lo-fi、bpm、弦乐进出、旋律动机、前中后段结构），如用户未指定可补充风格建议，但保持原方向。",
+		sfx: "这是音效生成任务：扩写声音材质、动作过程、空间感、节奏（先轻后重、清脆短促等）、环境氛围，用具体拟声与材质词，避免抽象概括。",
+		voice_design: "这是音色设计任务：扩写人声/音色特征——性别年龄、音域、音质（低沉/清亮/沙哑）、语速、情绪性格、适用场景，用可感知的描述，方便语音模型合成。"
+	}[mode];
+}
+/** 读取 Agent 默认模型并调用 LLM 增强，返回增强后的文本。 */
+async function enhancePromptText(deps, prompt, mode) {
+	const text = prompt.trim();
+	if (text === "") throw new AudioGenError("提示词为空，无法增强", "enhance-empty-prompt");
+	const value = (deps.settings.describe({ redactSecrets: true }) ?? []).find((candidate) => String(candidate.ns) === "agent-default-model")?.value ?? {};
+	const provider = typeof value.provider === "string" && value.provider.trim() !== "" ? value.provider.trim() : "";
+	const model = typeof value.model === "string" && value.model.trim() !== "" ? value.model.trim() : "";
+	if (provider === "" || model === "") throw new AudioGenError("未找到 Agent 默认模型（agent-default-model）：请先在「设置 → 模型」中配置默认模型", "no-default-model");
+	const runtime = deps.llm?.();
+	if (runtime === void 0 || runtime.stream === void 0) throw new AudioGenError("宿主 LLM 服务不可用（ctx.llm 未注册）", "llm-unavailable");
+	let output = "";
+	for await (const chunk of runtime.stream({
+		provider,
+		model,
+		messages: [{
+			role: "user",
+			content: text
+		}],
+		system: instructionsFor(mode),
+		temperature: .7,
+		maxTokens: 1200
+	})) {
+		const record = chunk;
+		if (record.type === "text-delta" && typeof record.text === "string") output += record.text;
+	}
+	return stripFences(output.trim());
+}
+/** 去掉模型可能包裹的 ``` 代码围栏。 */
+function stripFences(value) {
+	if (value === "") return value;
+	return value.replace(/^```[a-zA-Z]*\s*\n?/, "").replace(/\n?```\s*$/, "").trim();
+}
+//#endregion
 //#region src/audio-presets.ts
 const AUDIO_PRESETS = [
 	{
@@ -2046,6 +2097,36 @@ function makeRoutes(deps) {
 				});
 			}
 		},
+		{
+			kind: "exact",
+			path: ENHANCE_API,
+			handler: async (req, res) => {
+				if (!guard(req, res, "POST")) return;
+				const body = await readJsonBody(req);
+				const prompt = typeof body?.prompt === "string" ? body.prompt.trim() : "";
+				if (prompt === "") {
+					writeJson(res, 200, {
+						ok: false,
+						code: "bad-request",
+						message: "prompt is required"
+					});
+					return;
+				}
+				const mode = body?.mode === "music" ? "music" : body?.mode === "sfx" ? "sfx" : body?.mode === "voice_design" ? "voice_design" : "tts";
+				try {
+					writeJson(res, 200, {
+						ok: true,
+						enhanced: await deps.enhance(prompt, mode)
+					});
+				} catch (error) {
+					writeJson(res, 200, {
+						ok: false,
+						code: "enhance-failed",
+						message: messageOf(error)
+					});
+				}
+			}
+		},
 		{
 			kind: "prefix",
 			path: AUDIO_API.file,
@@ -2483,6 +2564,10 @@ function registerAgentAudioTools(ctx, resolve) {
 				type: "string",
 				description: "Optional preview text for voice_design."
 			},
+			enhance_prompt: {
+				type: "boolean",
+				description: "Enhance the prompt with the agent default model before generating (uses the configured model settings, no extra key). Best-effort: on failure the original prompt is used."
+			},
 			speed: {
 				type: "number",
 				description: "Optional speaking rate / speed multiplier where supported. MiniMax range 0.5-2.0 (default 1)."
@@ -2725,6 +2810,9 @@ function registerAgentAudioTools(ctx, resolve) {
 			/** 单模型执行：生成 + 保存文件 + 历史 + 可选资源库；错误收敛为分组结果。 */
 			const runOne = async (picked) => {
 				const request = buildRequest(picked);
+				if (args.enhance_prompt === true && config.enhance !== void 0) try {
+					request.prompt = await config.enhance(request.prompt, request.mode);
+				} catch {}
 				try {
 					const release = await (config.budget?.acquire(exec.signal) ?? Promise.resolve(() => {}));
 					let outputs;
@@ -3137,6 +3225,14 @@ function apply(ctx, config) {
 		};
 	};
 	const budget = createGenerationBudget(() => resolve().maxConcurrentGenerations);
+	const enhance = async (prompt, mode) => {
+		const seam = ctx.get("settings");
+		if (seam?.describe === void 0) throw new AudioGenError("设置服务不可用，无法增强提示词", "settings-unavailable");
+		return enhancePromptText({
+			settings: seam,
+			llm: () => ctx.get("llm")
+		}, prompt, mode);
+	};
 	const channelsView = () => {
 		const value = resolve();
 		return {
@@ -3151,7 +3247,8 @@ function apply(ctx, config) {
 				settings: seam,
 				resolveChannels: channelsView,
 				autoSave: () => resolve().autoSaveToLibrary,
-				budget
+				budget,
+				enhance
 			}).map((route) => ctx.webServer.register(route));
 			return () => {
 				for (const dispose of disposers) dispose();
@@ -3167,7 +3264,8 @@ function apply(ctx, config) {
 				channels: value.channels,
 				defaultChannelId: value.defaultChannelId,
 				autoSaveToLibrary: value.autoSaveToLibrary,
-				budget
+				budget,
+				enhance
 			};
 		}), "dsh-audiogen: agent audio tools");
 	});

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dsh-audiogen",
   "description": "AI audio generation plugin for the dsh web GUI: multi-vendor TTS/music/sound-effect channels (OpenAI-compatible, ElevenLabs, MiniMax, Stability AI and custom), per-channel model/voice catalogs, Agent tool and a sidebar AI 音频 panel.",
-  "version": "0.4.3",
+  "version": "0.4.5",
   "type": "module",
   "main": "lib/index.js",
   "exports": {

package/src/agent-audio-tools.ts

@@ -21,6 +21,8 @@ export interface AgentAudioToolConfig {
   autoSaveToLibrary: boolean
   /** 全局并发闸门（与面板路由共享「最大并发生成数」）。 */
   budget?: GenerationBudget
+  /** 提示词增强（复用 Agent 默认模型）；enhance_prompt=true 时在生成前调用。 */
+  enhance?: (prompt: string, mode: AudioMode) => Promise<string>
 }
 
 interface AgentAudioRef {
@@ -157,6 +159,7 @@ export function registerAgentAudioTools(ctx: Context, resolve: () => AgentAudioT
       },
       voice: { type: 'string', description: 'Optional voice id/name for TTS providers. Required for MiniMax TTS (e.g. male-qn-qingse, female-shaonv); fetch the account voices in Settings > Plugins > AI Audio.' },
       preview_text: { type: 'string', description: 'Optional preview text for voice_design.' },
+      enhance_prompt: { type: 'boolean', description: 'Enhance the prompt with the agent default model before generating (uses the configured model settings, no extra key). Best-effort: on failure the original prompt is used.' },
       speed: { type: 'number', description: 'Optional speaking rate / speed multiplier where supported. MiniMax range 0.5-2.0 (default 1).' },
       duration: { type: 'number', description: 'Requested duration in seconds for music/sfx.' },
       lyrics: { type: 'string', description: 'Lyrics for music generation (MiniMax music-3.0/music-cover). Required unless is_instrumental is true. Split verses with an empty line.' },
@@ -302,6 +305,14 @@ export function registerAgentAudioTools(ctx: Context, resolve: () => AgentAudioT
       /** 单模型执行：生成 + 保存文件 + 历史 + 可选资源库；错误收敛为分组结果。 */
       const runOne = async (picked: { channel: AudioChannel; alias: string; upstream: string }): Promise<AgentAudioGroup> => {
         const request = buildRequest(picked)
+        // 可选：生成前用 Agent 默认模型增强 prompt（失败则沿用原文）
+        if (args.enhance_prompt === true && config.enhance !== undefined) {
+          try {
+            request.prompt = await config.enhance(request.prompt, request.mode)
+          } catch {
+            // 增强失败不阻断生成
+          }
+        }
         try {
           // 与面板路由共享全局并发闸门（限流时排队；取消时立即出队）。
           const release = await (config.budget?.acquire(exec.signal) ?? Promise.resolve(() => { /* 默认不限制 */ }))

package/src/client/api.ts

@@ -4,7 +4,7 @@
  */
 
 import {
-  GENERATE_API, HISTORY_API, LIBRARY_API, TASK_API,
+  ENHANCE_API, GENERATE_API, HISTORY_API, LIBRARY_API, TASK_API,
   type GenerateAudioRequest, type GeneratedAudio, type HistoryEntry,
   type LibraryEntry, type LibrarySaveRequest, type LibraryUpdateRequest,
 } from '../protocol.ts'
@@ -42,12 +42,26 @@ export class AudiogenApi {
     await postJson(TASK_API.cancel, { taskId }).catch(() => { /* best-effort */ })
   }
 
+  /** 提示词增强（复用 Agent 默认模型）。 */
+  async enhancePrompt(prompt: string, mode: string): Promise<{ ok: boolean; enhanced?: string; code?: string; message?: string }> {
+    const response = await postJson(ENHANCE_API, { prompt, mode })
+    const body = await response.json() as { ok?: boolean; enhanced?: string; code?: string; message?: string }
+    return { ok: body.ok === true, ...(body.enhanced === undefined ? {} : { enhanced: body.enhanced }), ...(body.code === undefined ? {} : { code: body.code }), ...(body.message === undefined ? {} : { message: body.message }) }
+  }
+
   async history(): Promise<HistoryEntry[]> {
     const response = await postJson(HISTORY_API.list, {})
     const body = await response.json() as { ok?: boolean; history?: HistoryEntry[] }
     return body.ok === true ? (body.history ?? []) : []
   }
 
+  /** 删除一条历史记录（返回删后的列表）。 */
+  async removeHistory(id: string): Promise<HistoryEntry[]> {
+    const response = await postJson(HISTORY_API.remove, { id })
+    const body = await response.json() as { ok?: boolean; history?: HistoryEntry[] }
+    return body.ok === true ? (body.history ?? []) : []
+  }
+
   async clearHistory(): Promise<void> {
     await postJson(HISTORY_API.clear, {})
   }

package/src/client/audio-panel.module.css

@@ -111,8 +111,8 @@
 .formCol::-webkit-scrollbar-thumb { background: var(--dsw-alias-border-l2); border-radius: 999px; }
 
 .modeRow {
-  display: grid;
-  grid-template-columns: repeat(4, minmax(0, 1fr));
+  display: flex;
+  flex-wrap: wrap;
   gap: 6px;
 }
 
@@ -120,8 +120,10 @@
   display: flex;
   align-items: center;
   justify-content: center;
+  flex: 1 1 auto;
+  min-width: 0;
   min-height: 34px;
-  padding: 6px 4px;
+  padding: 6px 10px;
   border: 1px solid var(--dsw-alias-border-l2, #d1d5db);
   border-radius: 9px;
   background: var(--dsw-alias-bg-layer-1, #fff);
@@ -953,14 +955,14 @@
   font-weight: 600;
   padding: 3px 10px;
   border-radius: 999px;
-  background: var(--dsw-alias-bg-hover, #f3f4f6);
+  background: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.14));
   border: 1px solid var(--dsw-alias-border-l2, #d1d5db);
   color: var(--dsw-alias-label-primary, #1f2328);
 }
 
 .resultGroupError {
   font-size: 12px;
-  color: #dc2626;
+  color: var(--dsw-alias-state-error-primary, #dc2626);
 }
 
 .resultGroupCount {
@@ -1021,7 +1023,7 @@
 }
 
 .taskCard[data-state='failed'] {
-  border-color: #dc2626;
+  border-color: var(--dsw-alias-state-error-primary, #dc2626);
 }
 
 .taskCard[data-state='cancelled'] {
@@ -1051,11 +1053,11 @@
 }
 
 .taskStatus[data-state='done'] {
-  color: #16a34a;
+  color: var(--dsw-alias-state-success-primary, #16a34a);
 }
 
 .taskStatus[data-state='failed'] {
-  color: #dc2626;
+  color: var(--dsw-alias-state-error-primary, #dc2626);
 }
 
 .taskActions {
@@ -1083,7 +1085,8 @@
 }
 
 .historyTab[data-active='true'] {
-  background: var(--dsw-alias-bg-hover, #f3f4f6);
+  /* 半透明交互强调色（亮色=深色 tint，暗色=白色 tint），避免选中整块白底 */
+  background: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.14));
   color: var(--dsw-alias-label-primary, #1f2328);
   font-weight: 600;
 }
@@ -1106,11 +1109,11 @@
 .historyCompareBadge {
   font-size: 11px;
   font-weight: 600;
-  color: #7c3aed;
-  border: 1px solid #ddd6fe;
+  color: var(--dsw-alias-label-secondary, #6b7280);
+  border: 1px solid var(--dsw-alias-border-l2, #d1d5db);
   border-radius: 999px;
   padding: 1px 8px;
-  background: #f5f3ff;
+  background: var(--dsw-alias-bg-layer-3, #fff);
   white-space: nowrap;
 }
 
@@ -1133,3 +1136,181 @@
   color: var(--dsw-alias-label-tertiary, #9ca3af);
   padding: 0 6px;
 }
+
+.historyTime {
+  font-size: 11px;
+  color: var(--dsw-alias-label-tertiary, #9ca3af);
+  margin: 2px 0;
+  white-space: nowrap;
+}
+
+/* 表单分区与两列排布 */
+.formSection {
+  font-size: 11px;
+  font-weight: 700;
+  letter-spacing: 0.04em;
+  color: var(--dsw-alias-label-tertiary, #9ca3af);
+  border-bottom: 1px solid var(--dsw-alias-border-l1, #e5e7eb);
+  padding-bottom: 4px;
+  margin: 10px 0 2px;
+}
+
+.formFields {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: 10px;
+  align-items: end;
+}
+
+.fieldCell {
+  min-width: 0;
+}
+
+.fieldFull {
+  grid-column: 1 / -1;
+}
+
+/* 模式胶囊 */
+.modeIcon {
+  font-size: 13px;
+  margin-right: 4px;
+}
+
+.modeButton[data-active='true'] {
+  border-color: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.14));
+  background: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.14));
+  color: var(--dsw-alias-label-primary, #1f2328);
+  font-weight: 600;
+}
+
+/* 任务细进度条 */
+.taskBar {
+  display: block;
+  flex: 1;
+  min-width: 60px;
+  height: 4px;
+  border-radius: 999px;
+  background: var(--dsw-alias-border-l1, #e5e7eb);
+  overflow: hidden;
+}
+
+.taskBar i {
+  display: block;
+  height: 100%;
+  border-radius: 999px;
+  background: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.35));
+  transition: width 0.3s;
+}
+
+/* 历史行内图标操作 */
+.historyIcon {
+  width: 26px;
+  height: 26px;
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  border-radius: 8px;
+  border: 1px solid transparent;
+  background: none;
+  color: var(--dsw-alias-label-secondary, #6b7280);
+  font-size: 14px;
+  cursor: pointer;
+}
+
+.historyIcon:hover {
+  background: var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.14));
+  color: var(--dsw-alias-label-primary, #1f2328);
+}
+
+.historyIcon[title^='删除']:hover {
+  color: var(--dsw-alias-state-error-primary, #dc2626);
+}
+
+/* 历史 prompt 两行截断 */
+.historyPrompt {
+  display: -webkit-box;
+  -webkit-line-clamp: 2;
+  -webkit-box-orient: vertical;
+  overflow: hidden;
+}
+
+/* ------------------------------------------------------- P3 打磨：响应式与统一质感 */
+
+/* 焦点态统一（键盘可达性） */
+.input:focus-visible,
+.select:focus-visible,
+.textarea:focus-visible,
+.modeButton:focus-visible,
+.ghostButton:focus-visible,
+.historyAction:focus-visible,
+.historyIcon:focus-visible {
+  outline: 2px solid var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.35));
+  outline-offset: 1px;
+}
+
+.generate:focus-visible {
+  outline: 2px solid var(--dsw-alias-interactive-bg-hover-accent, rgba(38, 49, 72, 0.35));
+  outline-offset: 2px;
+}
+
+/* 中窄屏：历史移到下方整行 */
+@media (max-width: 1100px) {
+  .studio {
+    flex-wrap: wrap;
+  }
+
+  .historyCol {
+    width: 100%;
+    min-width: 0;
+    max-width: none;
+  }
+}
+
+/* 窄屏：单列堆叠 */
+@media (max-width: 720px) {
+  .formCol {
+    width: 100%;
+    min-width: 0;
+    max-width: none;
+  }
+
+  .resultCol {
+    min-height: 320px;
+  }
+}
+
+/* 提示词增强 */
+.formSectionRow {
+  display: flex;
+  align-items: flex-end;
+  justify-content: space-between;
+  gap: 8px;
+}
+
+.formSectionRow .formSection {
+  flex: 1;
+}
+
+.enhanceCard {
+  border: 1px solid var(--dsw-alias-border-l1, #e5e7eb);
+  border-radius: 10px;
+  padding: 10px;
+  background: var(--dsw-alias-bg-layer-2, #fafafa);
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+}
+
+.enhanceCardHead {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: 8px;
+  font-size: 12px;
+  color: var(--dsw-alias-label-secondary, #6b7280);
+}
+
+.enhanceActions {
+  display: flex;
+  gap: 6px;
+}