@huggingface/inference 2.1.3 → 2.2.0

package/dist/index.js

@@ -42,6 +42,7 @@ __export(src_exports, {
   textGeneration: () => textGeneration,
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
+  textToSpeech: () => textToSpeech,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -72,6 +73,7 @@ __export(tasks_exports, {
   textGeneration: () => textGeneration,
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
+  textToSpeech: () => textToSpeech,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -286,7 +288,11 @@ async function* streamingRequest(args, options) {
       onChunk(value);
       for (const event of events) {
         if (event.data.length > 0) {
-          yield JSON.parse(event.data);
+          const data = JSON.parse(event.data);
+          if (typeof data === "object" && data !== null && "error" in data) {
+            throw new Error(data.error);
+          }
+          yield data;
         }
       }
       events = [];
@@ -326,6 +332,16 @@ async function automaticSpeechRecognition(args, options) {
   return res;
 }
 
+// src/tasks/audio/textToSpeech.ts
+async function textToSpeech(args, options) {
+  const res = await request(args, options);
+  const isValidOutput = res && res instanceof Blob;
+  if (!isValidOutput) {
+    throw new InferenceOutputError("Expected Blob");
+  }
+  return res;
+}
+
 // src/tasks/cv/imageClassification.ts
 async function imageClassification(args, options) {
   const res = await request(args, options);
@@ -657,6 +673,7 @@ var HfInferenceEndpoint = class {
   textGeneration,
   textGenerationStream,
   textToImage,
+  textToSpeech,
   tokenClassification,
   translation,
   visualQuestionAnswering,

package/dist/index.mjs

@@ -27,6 +27,7 @@ __export(tasks_exports, {
   textGeneration: () => textGeneration,
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
+  textToSpeech: () => textToSpeech,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -241,7 +242,11 @@ async function* streamingRequest(args, options) {
       onChunk(value);
       for (const event of events) {
         if (event.data.length > 0) {
-          yield JSON.parse(event.data);
+          const data = JSON.parse(event.data);
+          if (typeof data === "object" && data !== null && "error" in data) {
+            throw new Error(data.error);
+          }
+          yield data;
         }
       }
       events = [];
@@ -281,6 +286,16 @@ async function automaticSpeechRecognition(args, options) {
   return res;
 }
 
+// src/tasks/audio/textToSpeech.ts
+async function textToSpeech(args, options) {
+  const res = await request(args, options);
+  const isValidOutput = res && res instanceof Blob;
+  if (!isValidOutput) {
+    throw new InferenceOutputError("Expected Blob");
+  }
+  return res;
+}
+
 // src/tasks/cv/imageClassification.ts
 async function imageClassification(args, options) {
   const res = await request(args, options);
@@ -611,6 +626,7 @@ export {
   textGeneration,
   textGenerationStream,
   textToImage,
+  textToSpeech,
   tokenClassification,
   translation,
   visualQuestionAnswering,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@huggingface/inference",
-  "version": "2.1.3",
+  "version": "2.2.0",
   "license": "MIT",
   "author": "Tim Mikeladze <tim.mikeladze@gmail.com>",
   "description": "Typescript wrapper for the Hugging Face Inference API",

package/src/tasks/audio/textToSpeech.ts

@@ -0,0 +1,25 @@
+import { InferenceOutputError } from "../../lib/InferenceOutputError";
+import type { BaseArgs, Options } from "../../types";
+import { request } from "../custom/request";
+
+export type TextToSpeechArgs = BaseArgs & {
+	/**
+	 * The text to generate an audio from
+	 */
+	inputs: string;
+};
+
+export type TextToSpeechOutput = Blob;
+
+/**
+ * This task synthesize an audio of a voice pronouncing a given text.
+ * Recommended model: espnet/kan-bayashi_ljspeech_vits
+ */
+export async function textToSpeech(args: TextToSpeechArgs, options?: Options): Promise<TextToSpeechOutput> {
+	const res = await request<TextToSpeechOutput>(args, options);
+	const isValidOutput = res && res instanceof Blob;
+	if (!isValidOutput) {
+		throw new InferenceOutputError("Expected Blob");
+	}
+	return res;
+}

package/src/tasks/custom/streamingRequest.ts

@@ -65,7 +65,11 @@ export async function* streamingRequest<T>(
 			onChunk(value);
 			for (const event of events) {
 				if (event.data.length > 0) {
-					yield JSON.parse(event.data) as T;
+					const data = JSON.parse(event.data);
+					if (typeof data === "object" && data !== null && "error" in data) {
+						throw new Error(data.error);
+					}
+					yield data as T;
 				}
 			}
 			events = [];

package/src/tasks/index.ts

@@ -5,6 +5,7 @@ export * from "./custom/streamingRequest";
 // Audio tasks
 export * from "./audio/audioClassification";
 export * from "./audio/automaticSpeechRecognition";
+export * from "./audio/textToSpeech";
 
 // Commputer Vision tasks
 export * from "./cv/imageClassification";

package/src/tasks/nlp/textGeneration.ts

@@ -44,6 +44,10 @@ export type TextGenerationArgs = BaseArgs & {
 		 * (Default: None). Float to define the tokens that are within the sample operation of text generation. Add tokens in the sample for more probable to least probable until the sum of the probabilities is greater than top_p.
 		 */
 		top_p?: number;
+		/**
+		 * (Default: None). Integer. The maximum number of tokens from the input.
+		 */
+		truncate?: number;
 	};
 };