assemblyai 4.4.1 → 4.4.2-beta.0

package/dist/node.mjs

@@ -1,5 +1,5 @@
 import { WritableStream } from 'stream/web';
-import WebSocket from '#ws';
+import ws from 'ws';
 import { createReadStream } from 'fs';
 import { Readable } from 'stream';
 
@@ -22,6 +22,7 @@ class BaseService {
             "Content-Type": "application/json",
             ...init.headers,
         };
+        init.cache = "no-store";
         if (!input.startsWith("http"))
             input = this.params.baseUrl + input;
         const response = await fetch(input, init);
@@ -85,6 +86,8 @@ class LemurService extends BaseService {
     }
 }
 
+const factory = (url, params) => new ws(url, params);
+
 var RealtimeErrorType;
 (function (RealtimeErrorType) {
     RealtimeErrorType[RealtimeErrorType["BadSampleRate"] = 4000] = "BadSampleRate";
@@ -180,10 +183,10 @@ class RealtimeTranscriber {
             }
             const url = this.connectionUrl();
             if (this.token) {
-                this.socket = new WebSocket(url.toString());
+                this.socket = factory(url.toString());
             }
             else {
-                this.socket = new WebSocket(url.toString(), {
+                this.socket = factory(url.toString(), {
                     headers: { Authorization: this.apiKey },
                 });
             }
@@ -276,14 +279,14 @@ class RealtimeTranscriber {
         this.send(`{"end_utterance_silence_threshold":${threshold}}`);
     }
     send(data) {
-        if (!this.socket || this.socket.readyState !== WebSocket.OPEN) {
+        if (!this.socket || this.socket.readyState !== this.socket.OPEN) {
             throw new Error("Socket is not open for communication");
         }
         this.socket.send(data);
     }
     async close(waitForSessionTermination = true) {
         if (this.socket) {
-            if (this.socket.readyState === WebSocket.OPEN) {
+            if (this.socket.readyState === this.socket.OPEN) {
                 if (waitForSessionTermination) {
                     const sessionTerminatedPromise = new Promise((resolve) => {
                         this.sessionTerminatedResolve = resolve;
@@ -295,7 +298,7 @@ class RealtimeTranscriber {
                     this.socket.send(terminateSessionMessage);
                 }
             }
-            if ("removeAllListeners" in this.socket)
+            if (this.socket?.removeAllListeners)
                 this.socket.removeAllListeners();
             this.socket.close();
         }
@@ -346,6 +349,8 @@ function getPath(path) {
         return null;
     if (path.startsWith("https"))
         return null;
+    if (path.startsWith("data:"))
+        return null;
     if (path.startsWith("file://"))
         return path.substring(7);
     if (path.startsWith("file:"))
@@ -387,8 +392,13 @@ class TranscriptService extends BaseService {
                     audioUrl = await this.files.upload(path);
                 }
                 else {
-                    // audio is not a local path, assume it's a URL
-                    audioUrl = audio;
+                    if (audio.startsWith("data:")) {
+                        audioUrl = await this.files.upload(audio);
+                    }
+                    else {
+                        // audio is not a local path, and not a data-URI, assume it's a normal URL
+                        audioUrl = audio;
+                    }
                 }
             }
             else {
@@ -565,8 +575,14 @@ class FileService extends BaseService {
      */
     async upload(input) {
         let fileData;
-        if (typeof input === "string")
-            fileData = await readFile(input);
+        if (typeof input === "string") {
+            if (input.startsWith("data:")) {
+                fileData = dataUrlToBlob(input);
+            }
+            else {
+                fileData = await readFile(input);
+            }
+        }
         else
             fileData = input;
         const data = await this.fetchJson("/v2/upload", {
@@ -580,6 +596,17 @@ class FileService extends BaseService {
         return data.upload_url;
     }
 }
+function dataUrlToBlob(dataUrl) {
+    const arr = dataUrl.split(",");
+    const mime = arr[0].match(/:(.*?);/)[1];
+    const bstr = atob(arr[1]);
+    let n = bstr.length;
+    const u8arr = new Uint8Array(n);
+    while (n--) {
+        u8arr[n] = bstr.charCodeAt(n);
+    }
+    return new Blob([u8arr], { type: mime });
+}
 
 const defaultBaseUrl = "https://api.assemblyai.com";
 class AssemblyAI {

package/dist/polyfills/websocket/browser.d.ts

@@ -0,0 +1,3 @@
+import { PolyfillWebSocketFactory } from ".";
+export { PolyfillWebSocket } from ".";
+export declare const factory: PolyfillWebSocketFactory;

package/dist/polyfills/websocket/default.d.ts

@@ -0,0 +1,3 @@
+import { PolyfillWebSocketFactory } from ".";
+export { PolyfillWebSocket } from ".";
+export declare const factory: PolyfillWebSocketFactory;

package/dist/polyfills/websocket/index.d.ts

@@ -0,0 +1,27 @@
+/// <reference types="node" />
+import ws, { Event, ErrorEvent, CloseEvent, MessageEvent } from "ws";
+export type PolyfillWebSocket = {
+    OPEN: typeof ws.OPEN;
+    binaryType: string;
+    onopen: ((event: Event) => void) | null;
+    onerror: ((event: ErrorEvent) => void) | null;
+    onclose: ((event: CloseEvent) => void) | null;
+    onmessage: ((event: MessageEvent) => void) | null;
+    readonly readyState: typeof ws.CONNECTING | typeof ws.OPEN | typeof ws.CLOSING | typeof ws.CLOSED;
+    removeAllListeners?: () => void;
+    send(data: string | number | Buffer | DataView | ArrayBufferView | Uint8Array | ArrayBuffer | SharedArrayBuffer | readonly unknown[] | readonly number[] | {
+        valueOf(): ArrayBuffer;
+    } | {
+        valueOf(): SharedArrayBuffer;
+    } | {
+        valueOf(): Uint8Array;
+    } | {
+        valueOf(): readonly number[];
+    } | {
+        valueOf(): string;
+    } | {
+        [Symbol.toPrimitive](hint: string): string;
+    }): unknown;
+    close(): unknown;
+};
+export type PolyfillWebSocketFactory = (url: string, params?: unknown) => PolyfillWebSocket;

package/dist/types/openapi.generated.d.ts

@@ -721,7 +721,7 @@ export type LemurQuestionAnswerResponse = LemurBaseResponse & {
  * ```js
  * {
  *   "transcript_ids": [
- *     "64nygnr62k-405c-4ae8-8a6b-d90b40ff3cce"
+ *     "47b95ba5-8889-44d8-bc80-5de38306e582"
  *   ],
  *   "context": "This is an interview about wildfires.",
  *   "final_model": "default",
@@ -828,7 +828,6 @@ export type ListTranscriptParams = {
     throttled_only?: boolean;
 };
 /**
- * Details of the transcript page.
  * Details of the transcript page. Transcripts are sorted from newest to oldest. The previous URL always points to a page with older transcripts.
  * @example
  * ```js
@@ -851,12 +850,10 @@ export type PageDetails = {
      */
     limit: number;
     /**
-     * The URL to the next page of transcripts
      * The URL to the next page of transcripts. The next URL always points to a page with newer transcripts.
      */
     next_url: string | null;
     /**
-     * The URL to the previous page of transcripts
      * The URL to the next page of transcripts. The previous URL always points to a page with older transcripts.
      */
     prev_url: string | null;
@@ -1124,7 +1121,7 @@ export type SentencesResponse = {
 };
 export type Sentiment = "POSITIVE" | "NEUTRAL" | "NEGATIVE";
 /**
- * The result of the sentiment analysis model
+ * The result of the Sentiment Analysis model
  * @example
  * ```js
  * {
@@ -2114,7 +2111,7 @@ export type Transcript = {
     auto_highlights: boolean;
     /**
      * An array of results for the Key Phrases model, if it is enabled.
-     * See {@link https://www.assemblyai.com/docs/models/key-phrases | Key phrases } for more information.
+     * See {@link https://www.assemblyai.com/docs/models/key-phrases | Key Phrases } for more information.
      */
     auto_highlights_result?: AutoHighlightsResult | null;
     /**
@@ -2236,7 +2233,7 @@ export type Transcript = {
     sentiment_analysis?: boolean | null;
     /**
      * An array of results for the Sentiment Analysis model, if it is enabled.
-     * See {@link https://www.assemblyai.com/docs/models/sentiment-analysis | Sentiment analysis } for more information.
+     * See {@link https://www.assemblyai.com/docs/models/sentiment-analysis | Sentiment Analysis } for more information.
      */
     sentiment_analysis_results?: SentimentAnalysisResult[] | null;
     /**
@@ -2313,7 +2310,7 @@ export type Transcript = {
      */
     webhook_status_code?: number | null;
     /**
-     * The URL to which we send webhooks upon trancription completion
+     * The URL to which we send webhooks upon transcription completion
      */
     webhook_url?: string | null;
     /**
@@ -2358,9 +2355,8 @@ export type TranscriptCustomSpelling = {
  *
  * @defaultValue "en_us
  */
-export type TranscriptLanguageCode = "en" | "en_au" | "en_uk" | "en_us" | "es" | "fr" | "de" | "it" | "pt" | "nl" | "hi" | "ja" | "zh" | "fi" | "ko" | "pl" | "ru" | "tr" | "uk" | "vi";
+export type TranscriptLanguageCode = "en" | "en_au" | "en_uk" | "en_us" | "es" | "fr" | "de" | "it" | "pt" | "nl" | "af" | "sq" | "am" | "ar" | "hy" | "as" | "az" | "ba" | "eu" | "be" | "bn" | "bs" | "br" | "bg" | "my" | "ca" | "zh" | "hr" | "cs" | "da" | "et" | "fo" | "fi" | "gl" | "ka" | "el" | "gu" | "ht" | "ha" | "haw" | "he" | "hi" | "hu" | "is" | "id" | "ja" | "jw" | "kn" | "kk" | "km" | "ko" | "lo" | "la" | "lv" | "ln" | "lt" | "lb" | "mk" | "mg" | "ms" | "ml" | "mt" | "mi" | "mr" | "mn" | "ne" | "no" | "nn" | "oc" | "pa" | "ps" | "fa" | "pl" | "ro" | "ru" | "sa" | "sr" | "sn" | "sd" | "si" | "sk" | "sl" | "so" | "su" | "sw" | "sv" | "tl" | "tg" | "ta" | "tt" | "te" | "th" | "bo" | "tr" | "tk" | "uk" | "ur" | "uz" | "vi" | "cy" | "yi" | "yo";
 /**
- * A list of transcripts
  * A list of transcripts. Transcripts are sorted from newest to oldest. The previous URL always points to a page with older transcripts.
  * @example
  * ```js
@@ -2498,7 +2494,7 @@ export type TranscriptOptionalParams = {
      */
     auto_chapters?: boolean;
     /**
-     * Whether Key Phrases is enabled, either true or false
+     * Enable Key Phrases, either true or false
      */
     auto_highlights?: boolean;
     /**
@@ -2510,7 +2506,7 @@ export type TranscriptOptionalParams = {
      */
     content_safety?: boolean;
     /**
-     * The confidence threshold for content moderation. Values must be between 25 and 100.
+     * The confidence threshold for the Content Moderation model. Values must be between 25 and 100.
      */
     content_safety_confidence?: number;
     /**
@@ -2518,7 +2514,7 @@ export type TranscriptOptionalParams = {
      */
     custom_spelling?: TranscriptCustomSpelling[];
     /**
-     * Whether custom topics is enabled, either true or false
+     * Enable custom topics, either true or false
      */
     custom_topics?: boolean;
     /**
@@ -2551,7 +2547,7 @@ export type TranscriptOptionalParams = {
      */
     language_code?: LiteralUnion<TranscriptLanguageCode, string> | null;
     /**
-     * Whether {@link https://www.assemblyai.com/docs/models/speech-recognition#automatic-language-detection | Automatic language detection } was enabled in the transcription request, either true or false.
+     * Enable {@link https://www.assemblyai.com/docs/models/speech-recognition#automatic-language-detection | Automatic language detection }, either true or false.
      */
     language_detection?: boolean;
     /**
@@ -2619,7 +2615,7 @@ export type TranscriptOptionalParams = {
      */
     summary_type?: SummaryType;
     /**
-     * The list of custom topics provided, if custom topics is enabled
+     * The list of custom topics
      */
     topics?: string[];
     /**
@@ -2633,7 +2629,7 @@ export type TranscriptOptionalParams = {
      */
     webhook_auth_header_value?: string | null;
     /**
-     * The URL to which AssemblyAI send webhooks upon trancription completion
+     * The URL to which AssemblyAI send webhooks upon transcription completion
      */
     webhook_url?: string;
     /**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.4.1",
+  "version": "4.4.2-beta.0",
   "description": "The AssemblyAI JavaScript SDK provides an easy-to-use interface for interacting with the AssemblyAI API, which supports async and real-time transcription, as well as the latest LeMUR models.",
   "engines": {
     "node": ">=18"
@@ -17,7 +17,8 @@
         "default": "./dist/deno.mjs"
       },
       "workerd": "./dist/index.mjs",
-      "browser": "./dist/index.mjs",
+      "browser": "./dist/browser.mjs",
+      "react-native": "./dist/browser.mjs",
       "node": {
         "types": "./dist/index.d.ts",
         "import": "./dist/node.mjs",
@@ -40,14 +41,10 @@
       "node": "./src/polyfills/streams/node.ts",
       "default": "./src/polyfills/streams/index.ts"
     },
-    "#ws": {
-      "types": "./src/polyfills/ws/index.d.ts",
-      "browser": "./src/polyfills/ws/browser.mjs",
-      "default": {
-        "types": "./src/polyfills/ws/index.d.ts",
-        "import": "./src/polyfills/ws/index.mjs",
-        "require": "./src/polyfills/ws/index.cjs"
-      }
+    "#websocket": {
+      "browser": "./src/polyfills/websocket/browser.ts",
+      "node": "./src/polyfills/websocket/default.ts",
+      "default": "./src/polyfills/websocket/default.ts"
     }
   },
   "type": "commonjs",
@@ -61,7 +58,7 @@
     "url": "git+https://github.com/AssemblyAI/assemblyai-node-sdk.git"
   },
   "publishConfig": {
-    "tag": "latest",
+    "tag": "beta",
     "access": "public",
     "registry": "https://registry.npmjs.org/"
   },
@@ -70,7 +67,7 @@
     "clean": "rimraf dist/* && rimraf temp/* && rimraf temp-docs/*",
     "lint": "eslint -c .eslintrc.json '{src,tests}/**/*.{js,ts}' && publint && tsc --noEmit -p tsconfig.json",
     "test": "pnpm run test:unit && pnpm run test:integration",
-    "test:unit": "jest --config jest.unit.config.js",
+    "test:unit": "jest --config jest.unit.config.js --testTimeout 1000",
     "test:integration": "jest --config jest.integration.config.js --testTimeout 360000",
     "format": "prettier '**/*' --write",
     "generate:types": "tsx ./scripts/generate-types.ts && prettier 'src/types/*.generated.ts' --write",

package/src/polyfills/websocket/browser.ts

@@ -0,0 +1,18 @@
+import { PolyfillWebSocketFactory, PolyfillWebSocket } from ".";
+export { PolyfillWebSocket } from ".";
+
+const PolyfillWebSocket =
+  WebSocket ?? global?.WebSocket ?? window?.WebSocket ?? self?.WebSocket;
+
+export const factory: PolyfillWebSocketFactory = (
+  url: string,
+  params?: unknown,
+) => {
+  if (params) {
+    return new PolyfillWebSocket(
+      url,
+      params as string | string[],
+    ) as unknown as PolyfillWebSocket;
+  }
+  return new PolyfillWebSocket(url) as unknown as PolyfillWebSocket;
+};

package/src/polyfills/websocket/default.ts

@@ -0,0 +1,8 @@
+import ws from "ws";
+import { PolyfillWebSocket, PolyfillWebSocketFactory } from ".";
+export { PolyfillWebSocket } from ".";
+
+export const factory: PolyfillWebSocketFactory = (
+  url: string,
+  params?: unknown,
+) => new ws(url, params as ws.ClientOptions) as unknown as PolyfillWebSocket;

package/src/polyfills/websocket/index.ts

@@ -0,0 +1,41 @@
+import ws, { Event, ErrorEvent, CloseEvent, MessageEvent } from "ws";
+
+export type PolyfillWebSocket = {
+  OPEN: typeof ws.OPEN;
+  binaryType: string;
+  onopen: ((event: Event) => void) | null;
+  onerror: ((event: ErrorEvent) => void) | null;
+  onclose: ((event: CloseEvent) => void) | null;
+  onmessage: ((event: MessageEvent) => void) | null;
+  readonly readyState:
+    | typeof ws.CONNECTING
+    | typeof ws.OPEN
+    | typeof ws.CLOSING
+    | typeof ws.CLOSED;
+  removeAllListeners?: () => void;
+  send(
+    data:
+      | string
+      | number
+      | Buffer
+      | DataView
+      | ArrayBufferView
+      | Uint8Array
+      | ArrayBuffer
+      | SharedArrayBuffer
+      | readonly unknown[]
+      | readonly number[]
+      | { valueOf(): ArrayBuffer }
+      | { valueOf(): SharedArrayBuffer }
+      | { valueOf(): Uint8Array }
+      | { valueOf(): readonly number[] }
+      | { valueOf(): string }
+      | { [Symbol.toPrimitive](hint: string): string },
+  ): unknown;
+  close(): unknown;
+};
+
+export type PolyfillWebSocketFactory = (
+  url: string,
+  params?: unknown,
+) => PolyfillWebSocket;

package/src/services/base.ts

@@ -21,6 +21,7 @@ export abstract class BaseService {
       "Content-Type": "application/json",
       ...init.headers,
     };
+    init.cache = "no-store";
     if (!input.startsWith("http")) input = this.params.baseUrl + input;
 
     const response = await fetch(input, init);

package/src/services/files/index.ts

@@ -10,8 +10,13 @@ export class FileService extends BaseService {
    */
   async upload(input: FileUploadParams): Promise<string> {
     let fileData: FileUploadData;
-    if (typeof input === "string") fileData = await readFile(input);
-    else fileData = input;
+    if (typeof input === "string") {
+      if (input.startsWith("data:")) {
+        fileData = dataUrlToBlob(input);
+      } else {
+        fileData = await readFile(input);
+      }
+    } else fileData = input;
 
     const data = await this.fetchJson<UploadedFile>("/v2/upload", {
       method: "POST",
@@ -24,3 +29,15 @@ export class FileService extends BaseService {
     return data.upload_url;
   }
 }
+
+function dataUrlToBlob(dataUrl: string) {
+  const arr = dataUrl.split(",");
+  const mime = arr[0].match(/:(.*?);/)![1];
+  const bstr = atob(arr[1]);
+  let n = bstr.length;
+  const u8arr = new Uint8Array(n);
+  while (n--) {
+    u8arr[n] = bstr.charCodeAt(n);
+  }
+  return new Blob([u8arr], { type: mime });
+}

package/src/services/realtime/service.ts

@@ -1,5 +1,8 @@
 import { WritableStream } from "#streams";
-import WebSocket from "#ws";
+import {
+  PolyfillWebSocket,
+  factory as polyfillWebSocketFactory,
+} from "#websocket";
 import { ErrorEvent, MessageEvent, CloseEvent } from "ws";
 import {
   RealtimeEvents,
@@ -52,7 +55,7 @@ export class RealtimeTranscriber {
   private endUtteranceSilenceThreshold?: number;
   private disablePartialTranscripts?: boolean;
 
-  private socket?: WebSocket;
+  private socket?: PolyfillWebSocket;
   private listeners: RealtimeListeners = {};
   private sessionTerminatedResolve?: () => void;
 
@@ -136,15 +139,15 @@ export class RealtimeTranscriber {
       const url = this.connectionUrl();
 
       if (this.token) {
-        this.socket = new WebSocket(url.toString());
+        this.socket = polyfillWebSocketFactory(url.toString());
       } else {
-        this.socket = new WebSocket(url.toString(), {
+        this.socket = polyfillWebSocketFactory(url.toString(), {
           headers: { Authorization: this.apiKey },
         });
       }
-      this.socket.binaryType = "arraybuffer";
+      this.socket!.binaryType = "arraybuffer";
 
-      this.socket.onopen = () => {
+      this.socket!.onopen = () => {
         if (
           this.endUtteranceSilenceThreshold === undefined ||
           this.endUtteranceSilenceThreshold === null
@@ -156,7 +159,7 @@ export class RealtimeTranscriber {
         );
       };
 
-      this.socket.onclose = ({ code, reason }: CloseEvent) => {
+      this.socket!.onclose = ({ code, reason }: CloseEvent) => {
         if (!reason) {
           if (code in RealtimeErrorType) {
             reason = RealtimeErrorMessages[code as RealtimeErrorType];
@@ -165,12 +168,12 @@ export class RealtimeTranscriber {
         this.listeners.close?.(code, reason);
       };
 
-      this.socket.onerror = (event: ErrorEvent) => {
+      this.socket!.onerror = (event: ErrorEvent) => {
         if (event.error) this.listeners.error?.(event.error as Error);
         else this.listeners.error?.(new Error(event.message));
       };
 
-      this.socket.onmessage = ({ data }: MessageEvent) => {
+      this.socket!.onmessage = ({ data }: MessageEvent) => {
         const message = JSON.parse(data.toString()) as RealtimeMessage;
         if ("error" in message) {
           this.listeners.error?.(new RealtimeError(message.error));
@@ -242,7 +245,7 @@ export class RealtimeTranscriber {
   }
 
   private send(data: BufferLike) {
-    if (!this.socket || this.socket.readyState !== WebSocket.OPEN) {
+    if (!this.socket || this.socket.readyState !== this.socket.OPEN) {
       throw new Error("Socket is not open for communication");
     }
     this.socket.send(data);
@@ -250,7 +253,7 @@ export class RealtimeTranscriber {
 
   async close(waitForSessionTermination = true) {
     if (this.socket) {
-      if (this.socket.readyState === WebSocket.OPEN) {
+      if (this.socket.readyState === this.socket.OPEN) {
         if (waitForSessionTermination) {
           const sessionTerminatedPromise = new Promise<void>((resolve) => {
             this.sessionTerminatedResolve = resolve;
@@ -261,7 +264,7 @@ export class RealtimeTranscriber {
           this.socket.send(terminateSessionMessage);
         }
       }
-      if ("removeAllListeners" in this.socket) this.socket.removeAllListeners();
+      if (this.socket?.removeAllListeners) this.socket.removeAllListeners();
       this.socket.close();
     }
 

package/src/services/transcripts/index.ts

@@ -60,8 +60,12 @@ export class TranscriptService extends BaseService {
           // audio is local path, upload local file
           audioUrl = await this.files.upload(path);
         } else {
-          // audio is not a local path, assume it's a URL
-          audioUrl = audio;
+          if (audio.startsWith("data:")) {
+            audioUrl = await this.files.upload(audio);
+          } else {
+            // audio is not a local path, and not a data-URI, assume it's a normal URL
+            audioUrl = audio;
+          }
         }
       } else {
         // audio is of uploadable type