assemblyai 4.4.0 → 4.4.2-beta.0

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,
@@ -50,10 +53,9 @@ export class RealtimeTranscriber {
   private apiKey?: string;
   private token?: string;
   private endUtteranceSilenceThreshold?: number;
-  private enableExtraSessionInformation?: boolean;
   private disablePartialTranscripts?: boolean;
 
-  private socket?: WebSocket;
+  private socket?: PolyfillWebSocket;
   private listeners: RealtimeListeners = {};
   private sessionTerminatedResolve?: () => void;
 
@@ -63,7 +65,6 @@ export class RealtimeTranscriber {
     this.wordBoost = params.wordBoost;
     this.encoding = params.encoding;
     this.endUtteranceSilenceThreshold = params.endUtteranceSilenceThreshold;
-    this.enableExtraSessionInformation = params.enableExtraSessionInformation;
     this.disablePartialTranscripts = params.disablePartialTranscripts;
     if ("token" in params && params.token) this.token = params.token;
     if ("apiKey" in params && params.apiKey) this.apiKey = params.apiKey;
@@ -91,12 +92,9 @@ export class RealtimeTranscriber {
     if (this.encoding) {
       searchParams.set("encoding", this.encoding);
     }
-    if (this.enableExtraSessionInformation) {
-      searchParams.set(
-        "enable_extra_session_information",
-        this.enableExtraSessionInformation.toString(),
-      );
-    }
+
+    searchParams.set("enable_extra_session_information", "true");
+
     if (this.disablePartialTranscripts) {
       searchParams.set(
         "disable_partial_transcripts",
@@ -141,15 +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
@@ -161,7 +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];
@@ -170,12 +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));
@@ -247,7 +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);
@@ -255,7 +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;
@@ -266,7 +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

package/src/types/openapi.generated.ts

@@ -789,7 +789,7 @@ 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",
@@ -901,7 +901,6 @@ export type ListTranscriptParams = {
 };
 
 /**
- * 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
@@ -924,12 +923,10 @@ 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;
@@ -1234,7 +1231,7 @@ 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
  * {
@@ -2239,7 +2236,7 @@ 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;
   /**
@@ -2361,7 +2358,7 @@ 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;
   /**
@@ -2438,7 +2435,7 @@ 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;
   /**
@@ -2497,19 +2494,100 @@ export type TranscriptLanguageCode =
   | "it"
   | "pt"
   | "nl"
-  | "hi"
-  | "ja"
+  | "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"
-  | "vi";
+  | "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
@@ -2649,7 +2727,7 @@ 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;
   /**
@@ -2661,7 +2739,7 @@ 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;
   /**
@@ -2669,7 +2747,7 @@ 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;
   /**
@@ -2702,7 +2780,7 @@ 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;
   /**
@@ -2770,7 +2848,7 @@ export type TranscriptOptionalParams = {
    */
   summary_type?: SummaryType;
   /**
-   * The list of custom topics provided, if custom topics is enabled
+   * The list of custom topics
    */
   topics?: string[];
   /**
@@ -2784,7 +2862,7 @@ 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/src/types/realtime/index.ts

@@ -37,7 +37,8 @@ type CreateRealtimeTranscriberParams = {
   /**
    * Enable extra session information.
    * Set to `true` to receive the `session_information` message before the session ends. Defaults to `false`.
-   * @defaultValue false
+   * @defaultValue true
+   * @deprecated This parameter is now ignored and will be removed. Session information will always be sent.
    */
   enableExtraSessionInformation?: boolean;
 } & (
@@ -91,7 +92,8 @@ type RealtimeTranscriberParams = {
   /**
    * Enable extra session information.
    * Set to `true` to receive the `session_information` message before the session ends. Defaults to `false`.
-   * @defaultValue false
+   * @defaultValue true
+   * @deprecated This parameter is now ignored and will be removed. Session information will always be sent.
    */
   enableExtraSessionInformation?: boolean;
 } & (

package/src/utils/path.ts

@@ -1,6 +1,7 @@
 export function getPath(path: string) {
   if (path.startsWith("http")) 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:")) return path.substring(5);
   return path;

package/src/polyfills/ws/browser.mjs

@@ -1,15 +0,0 @@
-var ws = null;
-
-if (typeof WebSocket !== "undefined") {
-  ws = WebSocket;
-} else if (typeof MozWebSocket !== "undefined") {
-  ws = MozWebSocket;
-} else if (typeof global !== "undefined") {
-  ws = global.WebSocket || global.MozWebSocket;
-} else if (typeof window !== "undefined") {
-  ws = window.WebSocket || window.MozWebSocket;
-} else if (typeof self !== "undefined") {
-  ws = self.WebSocket || self.MozWebSocket;
-}
-
-export default ws;

package/src/polyfills/ws/index.cjs

@@ -1 +0,0 @@
-module.exports = require("ws");

package/src/polyfills/ws/index.d.ts

@@ -1,2 +0,0 @@
-import ws from "ws";
-export default ws;

package/src/polyfills/ws/index.mjs

@@ -1,2 +0,0 @@
-import ws from "ws";
-export default ws;