@qvac/sdk 0.16.0 → 0.17.0

package/dist/schemas/audio-gen.js

@@ -0,0 +1,69 @@
+import { z } from 'zod';
+import { modelSrcInputSchema } from '../schemas/model-src-utils.js';
+const base64Schema = z.string().min(1);
+export const audioGenRuntimeConfigSchema = z
+    .object({
+    useGPU: z.boolean().optional(),
+    inferenceSteps: z.number().int().nonnegative().optional(),
+    shift: z.number().nonnegative().optional(),
+    nGpuLayers: z.number().int().nonnegative().optional(),
+    threads: z.number().int().nonnegative().optional(),
+    backendsDir: z.string().min(1).optional()
+})
+    .strict();
+export const audioGenConfigSchema = audioGenRuntimeConfigSchema
+    .extend({
+    textEncModelSrc: modelSrcInputSchema,
+    lmModelSrc: modelSrcInputSchema,
+    ditModelSrc: modelSrcInputSchema,
+    vaeModelSrc: modelSrcInputSchema
+})
+    .strict();
+const audioGenParamsShape = {
+    modelId: z.string().min(1),
+    caption: z.string().trim().min(1, 'caption must not be empty or whitespace-only'),
+    lyrics: z.string().optional(),
+    seed: z.number().int().optional(),
+    vocalLanguage: z.string().min(1).optional(),
+    bpm: z.number().int().positive().optional(),
+    keyscale: z.string().min(1).optional(),
+    timesignature: z.string().min(1).optional(),
+    duration: z
+        .number()
+        .positive()
+        .optional()
+        .describe('Approximate requested duration in seconds. ACE-Step rounds to its latent frame grid; use output frames or stats.audioDurationMs as authoritative.')
+};
+export const audioGenClientParamsSchema = z.object(audioGenParamsShape).strict();
+export const audioGenStreamRequestSchema = z
+    .object({
+    ...audioGenParamsShape,
+    type: z.literal('audioGenStream'),
+    requestId: z.string().min(1).optional()
+})
+    .strict();
+export const audioGenProgressSchema = z.object({
+    stage: z.string(),
+    step: z.number().int().nonnegative(),
+    total: z.number().int().nonnegative()
+});
+export const audioGenStatsSchema = z.object({
+    audioDurationMs: z.number().optional(),
+    totalTimeMs: z.number().optional(),
+    realTimeFactor: z.number().optional(),
+    backendDevice: z.number().optional(),
+    backendId: z.number().optional()
+});
+export const audioGenStreamResponseSchema = z
+    .object({
+    type: z.literal('audioGenStream'),
+    progress: audioGenProgressSchema.optional(),
+    data: base64Schema.optional(),
+    sampleRate: z.number().int().positive().optional(),
+    channels: z.number().int().positive().optional(),
+    bitsPerSample: z.number().int().positive().optional(),
+    done: z.boolean().default(false),
+    stopReason: z.enum(['completed', 'cancelled']).optional(),
+    stats: audioGenStatsSchema.optional()
+})
+    .strict();

package/dist/schemas/batch-completion-stream.d.ts

@@ -1,6 +1,6 @@
 import { z } from 'zod';
-import { type CompletionEvent, type CompletionFinal, type CompletionStats } from './completion-event';
-import { type McpClientInput } from './mcp-adapter';
+import { type CompletionEvent, type CompletionFinal, type CompletionStats } from './completion-event.js';
+import { type McpClientInput } from './mcp-adapter.js';
 export declare const batchPromptSchema: z.ZodObject<{
     id: z.ZodOptional<z.ZodString>;
     history: z.ZodArray<z.ZodObject<{
@@ -183,6 +183,7 @@ export declare const batchCompletionClientParamsSchema: z.ZodObject<{
         harmony: "harmony";
         qwen35: "qwen35";
         gemma4: "gemma4";
+        dsml: "dsml";
     }>>;
     requestId: z.ZodOptional<z.ZodString>;
 }, z.core.$strict>;
@@ -198,6 +199,7 @@ export declare const batchCompletionStreamRequestSchema: z.ZodObject<{
         harmony: "harmony";
         qwen35: "qwen35";
         gemma4: "gemma4";
+        dsml: "dsml";
     }>>;
     requestId: z.ZodOptional<z.ZodString>;
     prompts: z.ZodArray<z.ZodObject<{
@@ -302,6 +304,7 @@ export declare const batchCompletionEventSchema: z.ZodObject<{
             cacheTokens: z.ZodOptional<z.ZodNumber>;
             promptTokens: z.ZodOptional<z.ZodNumber>;
             generatedTokens: z.ZodOptional<z.ZodNumber>;
+            emittedTokens: z.ZodOptional<z.ZodNumber>;
             avgConcurrentSeq: z.ZodOptional<z.ZodNumber>;
             backendDevice: z.ZodOptional<z.ZodEnum<{
                 gpu: "gpu";
@@ -380,6 +383,7 @@ export declare const batchCompletionStreamResponseSchema: z.ZodObject<{
                 cacheTokens: z.ZodOptional<z.ZodNumber>;
                 promptTokens: z.ZodOptional<z.ZodNumber>;
                 generatedTokens: z.ZodOptional<z.ZodNumber>;
+                emittedTokens: z.ZodOptional<z.ZodNumber>;
                 avgConcurrentSeq: z.ZodOptional<z.ZodNumber>;
                 backendDevice: z.ZodOptional<z.ZodEnum<{
                     gpu: "gpu";
@@ -416,6 +420,7 @@ export declare const batchCompletionStreamResponseSchema: z.ZodObject<{
         cacheTokens: z.ZodOptional<z.ZodNumber>;
         promptTokens: z.ZodOptional<z.ZodNumber>;
         generatedTokens: z.ZodOptional<z.ZodNumber>;
+        emittedTokens: z.ZodOptional<z.ZodNumber>;
         avgConcurrentSeq: z.ZodOptional<z.ZodNumber>;
         backendDevice: z.ZodOptional<z.ZodEnum<{
             gpu: "gpu";

package/dist/schemas/bci-config.d.ts

@@ -45,6 +45,7 @@ export declare const bciConfigSchema: z.ZodObject<{
         addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
             embeddings: "embeddings";
             diffusion: "diffusion";
+            audiogen: "audiogen";
             tts: "tts";
             ocr: "ocr";
             vla: "vla";
@@ -58,6 +59,7 @@ export declare const bciConfigSchema: z.ZodObject<{
             "parakeet-transcription": "parakeet-transcription";
             "ggml-ocr": "ggml-ocr";
             "sdcpp-generation": "sdcpp-generation";
+            "audiogen-ggml": "audiogen-ggml";
             "ggml-vla": "ggml-vla";
             "ggml-classification": "ggml-classification";
             llm: "llm";

package/dist/schemas/bci.d.ts

@@ -1,5 +1,5 @@
 import { z } from 'zod';
-import { type TranscribeSegment } from './transcription';
+import { type TranscribeSegment } from './transcription.js';
 export declare const neuralInputSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
     type: z.ZodLiteral<"base64">;
     value: z.ZodString;

package/dist/schemas/cancel.d.ts

@@ -22,6 +22,7 @@ declare const cancelKindSchema: z.ZodEnum<{
     transcribe: "transcribe";
     translate: "translate";
     diffusion: "diffusion";
+    audiogen: "audiogen";
     tts: "tts";
     ocr: "ocr";
     vla: "vla";
@@ -75,6 +76,7 @@ declare const cancelBroadParamsSchema: z.ZodObject<{
         transcribe: "transcribe";
         translate: "translate";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -98,6 +100,7 @@ declare const cancelParamsSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
         transcribe: "transcribe";
         translate: "translate";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -123,6 +126,7 @@ export declare const cancelRequestSchema: z.ZodIntersection<z.ZodObject<{
         transcribe: "transcribe";
         translate: "translate";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -161,6 +165,7 @@ export declare const cancelBroadSugarSchema: z.ZodObject<{
         transcribe: "transcribe";
         translate: "translate";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";

package/dist/schemas/cancel.js

@@ -26,6 +26,7 @@ const cancelKindSchema = z
     'transcribe',
     'translate',
     'diffusion',
+    'audiogen',
     'tts',
     'ocr',
     'vla',