SDK regeneration

cartesia-ai · Nov 29, 2024 · 818d16c · 818d16c
1 parent 4293760
commit 818d16c
Show file tree

Hide file tree

Showing 187 changed files with 384 additions and 4,717 deletions.
diff --git a/jest.config.js b/jest.config.js
@@ -1,12 +1,5 @@
 /** @type {import('jest').Config} */
 module.exports = {
-    preset: "ts-jest/presets/default-esm",
+    preset: "ts-jest",
     testEnvironment: "node",
-    extensionsToTreatAsEsm: [".ts"],
-    globals: {
-        "ts-jest": {
-            useESM: true,
-        },
-    },
-    transformIgnorePatterns: ["/node_modules/(?!(emittery)/)"],
 };
diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
     "name": "@cartesia/cartesia-js",
-    "version": "1.3.1",
+    "version": "2.0.0-alpha0",
     "private": false,
     "repository": "https://github.com/cartesia-ai/cartesia-js",
     "main": "./index.js",
@@ -12,33 +12,29 @@
         "test": "jest"
     },
     "dependencies": {
-        "emittery": "^1.0.3",
+        "url-join": "4.0.1",
         "form-data": "^4.0.0",
-        "form-data-encoder": "^4.0.2",
         "formdata-node": "^6.0.3",
-        "human-id": "^4.1.1",
         "node-fetch": "2.7.0",
         "qs": "6.11.2",
         "readable-stream": "^4.5.2",
-        "url-join": "4.0.1",
-        "ws": "^8.14.2"
+        "form-data-encoder": "^4.0.2"
     },
     "devDependencies": {
-        "@types/jest": "29.5.5",
-        "@types/node": "17.0.33",
-        "@types/node-fetch": "2.6.9",
+        "@types/url-join": "4.0.1",
         "@types/qs": "6.9.8",
+        "@types/node-fetch": "2.6.9",
         "@types/readable-stream": "^4.0.15",
-        "@types/url-join": "4.0.1",
-        "@types/ws": "^8.5.13",
         "fetch-mock-jest": "^1.5.1",
+        "webpack": "^5.94.0",
+        "ts-loader": "^9.3.1",
         "jest": "29.7.0",
+        "@types/jest": "29.5.5",
+        "ts-jest": "29.1.1",
         "jest-environment-jsdom": "29.7.0",
+        "@types/node": "17.0.33",
         "prettier": "2.7.1",
-        "ts-jest": "29.1.1",
-        "ts-loader": "^9.3.1",
-        "typescript": "4.6.4",
-        "webpack": "^5.94.0"
+        "typescript": "4.6.4"
     },
     "browser": {
         "fs": false,

diff --git a/reference.md b/reference.md
@@ -44,7 +44,7 @@ await client.apiStatus.get();
 
 ## Tts
 
-<details><summary><code>client.tts.<a href="/src/api/resources/tts/client/Client.ts">bytes</a>({ ...params }) -> stream.Readable</code></summary>
+<details><summary><code>client.tts.<a href="/src/api/resources/tts/client/Client.ts">bytes</a>({ ...params }) -> void</code></summary>
 <dl>
 <dd>
 
@@ -58,17 +58,17 @@ await client.apiStatus.get();
 
 ```typescript
 await client.tts.bytes({
-    modelId: "sonic-english",
+    model_id: "sonic-english",
     transcript: "Hello, world!",
     voice: {
         mode: "id",
         id: "694f9389-aac1-45b6-b726-9d9369183238",
     },
     language: "en",
-    outputFormat: {
+    output_format: {
         container: "mp3",
-        sampleRate: 44100,
-        bitRate: 128000,
+        sample_rate: 44100,
+        bit_rate: 128000,
     },
 });
 ```
@@ -119,18 +119,13 @@ await client.tts.bytes({
 
 ```typescript
 const response = await client.tts.sse({
-    modelId: "string",
+    model_id: "string",
     transcript: "string",
     voice: {
         mode: "id",
-        id: "string",
-        experimentalControls: {
-            speed: 1.1,
-            emotion: "anger:lowest",
-        },
     },
     language: "en",
-    outputFormat: {
+    output_format: {
         container: "raw",
     },
     duration: 1.1,
@@ -205,10 +200,10 @@ This endpoint is priced at 15 characters per second of input audio.
 
 ```typescript
 await client.voiceChanger.bytes(fs.createReadStream("/path/to/your/file"), {
-    voiceId: "694f9389-aac1-45b6-b726-9d9369183238",
-    outputFormatContainer: "mp3",
-    outputFormatSampleRate: 44100,
-    outputFormatBitRate: 128000,
+    "voice[id]": "694f9389-aac1-45b6-b726-9d9369183238",
+    "output_format[container]": "mp3",
+    "output_format[sample_rate]": 44100,
+    "output_format[bit_rate]": 128000,
 });
 ```
 
@@ -266,10 +261,10 @@ await client.voiceChanger.bytes(fs.createReadStream("/path/to/your/file"), {
 
 ```typescript
 const response = await client.voiceChanger.sse(fs.createReadStream("/path/to/your/file"), {
-    voiceId: "694f9389-aac1-45b6-b726-9d9369183238",
-    outputFormatContainer: "mp3",
-    outputFormatSampleRate: 44100,
-    outputFormatBitRate: 128000,
+    "voice[id]": "694f9389-aac1-45b6-b726-9d9369183238",
+    "output_format[container]": "mp3",
+    "output_format[sample_rate]": 44100,
+    "output_format[bit_rate]": 128000,
 });
 for await (const item of response) {
     console.log(item);
@@ -383,7 +378,7 @@ await client.voices.create({
         1, 1, 1, 1, 1, 1, 1,
     ],
     language: "en",
-    baseVoiceId: "string",
+    base_voice_id: "string",
 });
 ```
 
@@ -597,7 +592,7 @@ await client.voices.localize({
         1, 1, 1, 1, 1, 1, 1,
     ],
     language: "en",
-    originalSpeakerGender: "male",
+    original_speaker_gender: "male",
     dialect: "au",
 });
 ```
@@ -689,7 +684,7 @@ await client.voices.mix({
 </dl>
 </details>
 
-<details><summary><code>client.voices.<a href="/src/api/resources/voices/client/Client.ts">cloneFromClip</a>(clip, { ...params }) -> Cartesia.EmbeddingResponse</code></summary>
+<details><summary><code>client.voices.<a href="/src/api/resources/voices/client/Client.ts">clone</a>(clip, { ...params }) -> Cartesia.VoiceMetadata</code></summary>
 <dl>
 <dd>
 
@@ -701,9 +696,11 @@ await client.voices.mix({
 <dl>
 <dd>
 
-Clone a voice from a clip. The clip should be a 15-20 second recording of a person speaking with little to no background noise.
+Clone a voice from an audio clip. This endpoint has two modes, stability and similarity.
+
+Similarity mode clones are more similar to the source clip, but may reproduce background noise. For these, use an audio clip about 5 seconds long.
 
-The endpoint will return an embedding that can either be used directly with text-to-speech endpoints or used to create a new voice.
+Stability mode clones are more stable, but may not sound as similar to the source clip. For these, use an audio clip 10-20 seconds long.
 
 </dd>
 </dl>
@@ -719,7 +716,13 @@ The endpoint will return an embedding that can either be used directly with text
 <dd>
 
 ```typescript
-await client.voices.cloneFromClip(fs.createReadStream("/path/to/your/file"), {});
+await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
+    name: "A high-stability cloned voice",
+    description: "Copied from Cartesia docs",
+    mode: "stability",
+    language: "en",
+    enhance: true,
+});
 ```
 
 </dd>
@@ -743,7 +746,7 @@ await client.voices.cloneFromClip(fs.createReadStream("/path/to/your/file"), {})
 <dl>
 <dd>
 
-**request:** `Cartesia.CloneFromClipRequest`
+**request:** `Cartesia.CloneVoiceRequest`
 
 </dd>
 </dl>

diff --git a/src/api/resources/apiStatus/client/Client.ts b/src/api/resources/apiStatus/client/Client.ts
@@ -5,7 +5,6 @@
 import * as environments from "../../../../environments";
 import * as core from "../../../../core";
 import * as Cartesia from "../../../index";
-import * as serializers from "../../../../serialization/index";
 import * as errors from "../../../../errors/index";
 
 export declare namespace ApiStatus {
@@ -46,8 +45,8 @@ export class ApiStatus {
                 "Cartesia-Version": requestOptions?.cartesiaVersion ?? this._options?.cartesiaVersion ?? "2024-06-10",
                 "X-Fern-Language": "JavaScript",
                 "X-Fern-SDK-Name": "@cartesia/cartesia-js",
-                "X-Fern-SDK-Version": "1.3.1",
-                "User-Agent": "@cartesia/cartesia-js/1.3.1",
+                "X-Fern-SDK-Version": "2.0.0-alpha0",
+                "User-Agent": "@cartesia/cartesia-js/2.0.0-alpha0",
                 "X-Fern-Runtime": core.RUNTIME.type,
                 "X-Fern-Runtime-Version": core.RUNTIME.version,
                 ...(await this._getCustomAuthorizationHeaders()),
@@ -59,13 +58,7 @@ export class ApiStatus {
             abortSignal: requestOptions?.abortSignal,
         });
         if (_response.ok) {
-            return serializers.ApiInfo.parseOrThrow(_response.body, {
-                unrecognizedObjectKeys: "passthrough",
-                allowUnrecognizedUnionMembers: true,
-                allowUnrecognizedEnumValues: true,
-                skipValidation: true,
-                breadcrumbsPrefix: ["response"],
-            });
+            return _response.body as Cartesia.ApiInfo;
         }
 
         if (_response.error.reason === "status-code") {

diff --git a/src/api/resources/tts/client/Client.ts b/src/api/resources/tts/client/Client.ts
@@ -5,10 +5,10 @@
 import * as environments from "../../../../environments";
 import * as core from "../../../../core";
 import * as Cartesia from "../../../index";
-import * as stream from "stream";
-import * as serializers from "../../../../serialization/index";
 import urlJoin from "url-join";
 import * as errors from "../../../../errors/index";
+import * as stream from "stream";
+import * as serializers from "../../../../serialization/index";
 
 export declare namespace Tts {
     interface Options {
@@ -34,8 +34,60 @@ export declare namespace Tts {
 export class Tts {
     constructor(protected readonly _options: Tts.Options = {}) {}
 
-    public async bytes(request: Cartesia.TtsRequest, requestOptions?: Tts.RequestOptions): Promise<stream.Readable> {
-        const _response = await (this._options.fetcher ?? core.fetcher)<stream.Readable>({
+    /**
+     * @param {Cartesia.TtsRequest} request
+     * @param {Tts.RequestOptions} requestOptions - Request-specific configuration.
+     *
+     * @example
+     *     await client.tts.bytes({
+     *         model_id: "sonic-english",
+     *         transcript: "Hello, world!",
+     *         voice: {
+     *             mode: "id",
+     *             id: "694f9389-aac1-45b6-b726-9d9369183238"
+     *         },
+     *         language: "en",
+     *         output_format: {
+     *             container: "mp3",
+     *             sample_rate: 44100,
+     *             bit_rate: 128000
+     *         }
+     *     })
+     *
+     * @example
+     *     await client.tts.bytes({
+     *         model_id: "sonic-english",
+     *         transcript: "Hello, world!",
+     *         voice: {
+     *             mode: "id",
+     *             id: "694f9389-aac1-45b6-b726-9d9369183238"
+     *         },
+     *         language: "en",
+     *         output_format: {
+     *             container: "wav",
+     *             sample_rate: 44100,
+     *             encoding: "pcm_f32le"
+     *         }
+     *     })
+     *
+     * @example
+     *     await client.tts.bytes({
+     *         model_id: "sonic-english",
+     *         transcript: "Hello, world!",
+     *         voice: {
+     *             mode: "id",
+     *             id: "694f9389-aac1-45b6-b726-9d9369183238"
+     *         },
+     *         language: "en",
+     *         output_format: {
+     *             container: "raw",
+     *             sample_rate: 44100,
+     *             encoding: "pcm_f32le"
+     *         }
+     *     })
+     */
+    public async bytes(request: Cartesia.TtsRequest, requestOptions?: Tts.RequestOptions): Promise<void> {
+        const _response = await (this._options.fetcher ?? core.fetcher)({
             url: urlJoin(
                 (await core.Supplier.get(this._options.environment)) ?? environments.CartesiaEnvironment.Production,
                 "/tts/bytes"
@@ -45,22 +97,21 @@ export class Tts {
                 "Cartesia-Version": requestOptions?.cartesiaVersion ?? this._options?.cartesiaVersion ?? "2024-06-10",
                 "X-Fern-Language": "JavaScript",
                 "X-Fern-SDK-Name": "@cartesia/cartesia-js",
-                "X-Fern-SDK-Version": "1.3.1",
-                "User-Agent": "@cartesia/cartesia-js/1.3.1",
+                "X-Fern-SDK-Version": "2.0.0-alpha0",
+                "User-Agent": "@cartesia/cartesia-js/2.0.0-alpha0",
                 "X-Fern-Runtime": core.RUNTIME.type,
                 "X-Fern-Runtime-Version": core.RUNTIME.version,
                 ...(await this._getCustomAuthorizationHeaders()),
             },
             contentType: "application/json",
             requestType: "json",
-            body: serializers.TtsRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
-            responseType: "streaming",
+            body: request,
             timeoutMs: requestOptions?.timeoutInSeconds != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
             maxRetries: requestOptions?.maxRetries,
             abortSignal: requestOptions?.abortSignal,
         });
         if (_response.ok) {
-            return _response.body;
+            return;
         }
 
         if (_response.error.reason === "status-code") {
@@ -99,15 +150,15 @@ export class Tts {
                 "Cartesia-Version": requestOptions?.cartesiaVersion ?? this._options?.cartesiaVersion ?? "2024-06-10",
                 "X-Fern-Language": "JavaScript",
                 "X-Fern-SDK-Name": "@cartesia/cartesia-js",
-                "X-Fern-SDK-Version": "1.3.1",
-                "User-Agent": "@cartesia/cartesia-js/1.3.1",
+                "X-Fern-SDK-Version": "2.0.0-alpha0",
+                "User-Agent": "@cartesia/cartesia-js/2.0.0-alpha0",
                 "X-Fern-Runtime": core.RUNTIME.type,
                 "X-Fern-Runtime-Version": core.RUNTIME.version,
                 ...(await this._getCustomAuthorizationHeaders()),
             },
             contentType: "application/json",
             requestType: "json",
-            body: serializers.TtsRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
+            body: request,
             responseType: "sse",
             timeoutMs: requestOptions?.timeoutInSeconds != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
             maxRetries: requestOptions?.maxRetries,
@@ -127,8 +178,8 @@ export class Tts {
                 },
                 signal: requestOptions?.abortSignal,
                 eventShape: {
-                    type: "sse",
-                    streamTerminator: "[DONE]",
+                    type: "json",
+                    messageTerminator: "\n",
                 },
             });
         }

diff --git a/src/api/resources/tts/types/CancelContextRequest.ts b/src/api/resources/tts/types/CancelContextRequest.ts
@@ -6,7 +6,7 @@ import * as Cartesia from "../../../index";
 
 export interface CancelContextRequest {
     /** The ID of the context to cancel. */
-    contextId: Cartesia.ContextId;
+    context_id: Cartesia.ContextId;
     /** Whether to cancel the context, so that no more messages are generated for that context. */
     cancel: true;
 }