Merge pull request #67 from AssemblyAI/E07417BDFEA3614F5967B1520F8B2F61

Sync from internal repo (2024/08/26)

CHANGELOG.md

@@ -1,5 +1,22 @@
 # Changelog
 
+## [4.7.0]
+
+- Add `language_confidence_threshold` to `Transcript`, `TranscriptParams`, and `TranscriptOptionalParams`.
+  > The confidence threshold for the automatically detected language.
+  > An error will be returned if the langauge confidence is below this threshold.
+- Add `language_confidence` to `Transcript`
+  > The confidence score for the detected language, between 0.0 (low confidence) and 1.0 (high confidence)
+
+Using these new fields you can determine the confidence of the language detection model (enable by setting `language_detection` to `true`), and fail the transcript if it doesn't meet your desired threshold.
+
+[Learn more about the new automatic language detection model and feature improvements on our blog.](https://www.assemblyai.com/blog/ald-improvements)
+
+## [4.6.2]
+
+- Change `RealtimeErrorType` from enum to const object.
+- Add `RealtimeErrorTypeCodes` which is a union of `RealtimeErrorType` values
+
 ## [4.6.1]
 
 - Remove `conformer-2` from `SpeechModel` union type.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.6.1",
+  "version": "4.7.0",
   "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"

src/services/realtime/service.ts

@@ -17,11 +17,8 @@ import {
   AudioData,
   SessionInformation,
 } from "../..";
-import {
-  RealtimeError,
-  RealtimeErrorMessages,
-  RealtimeErrorType,
-} from "../../utils/errors";
+import { RealtimeError, RealtimeErrorMessages } from "../../utils/errors";
+import { RealtimeErrorTypeCodes } from "../../utils/errors/realtime";
 
 const defaultRealtimeUrl = "wss://api.assemblyai.com/v2/realtime/ws";
 const forceEndOfUtteranceMessage = `{"force_end_utterance":true}`;
@@ -213,8 +210,8 @@ export class RealtimeTranscriber {
 
       this.socket!.onclose = ({ code, reason }: CloseEvent) => {
         if (!reason) {
-          if (code in RealtimeErrorType) {
-            reason = RealtimeErrorMessages[code as RealtimeErrorType];
+          if (code in RealtimeErrorMessages) {
+            reason = RealtimeErrorMessages[code as RealtimeErrorTypeCodes];
           }
         }
         this.listeners.close?.(code, reason);

src/types/openapi.generated.ts

@@ -2406,6 +2406,16 @@ export type Transcript = {
    * The default value is 'en_us'.
    */
   language_code?: LiteralUnion<TranscriptLanguageCode, string>;
+  /**
+   * The confidence score for the detected language, between 0.0 (low confidence) and 1.0 (high confidence)
+   */
+  language_confidence: number | null;
+  /**
+   * The confidence threshold for the automatically detected language.
+   * An error will be returned if the langauge confidence is below this threshold.
+   * Defaults to 0.
+   */
+  language_confidence_threshold: number | null;
   /**
    * Whether {@link https://www.assemblyai.com/docs/models/speech-recognition#automatic-language-detection | Automatic language detection } is enabled, either true or false
    */
@@ -2542,7 +2552,7 @@ export type Transcript = {
 };
 
 /**
- * The word boost parameter value
+ * How much to boost specified words
  */
 export type TranscriptBoostParam = "low" | "default" | "high";
 
@@ -2823,7 +2833,7 @@ export type TranscriptOptionalParams = {
    */
   auto_highlights?: boolean;
   /**
-   * The word boost parameter value
+   * How much to boost specified words
    */
   boost_param?: TranscriptBoostParam;
   /**
@@ -2871,6 +2881,12 @@ export type TranscriptOptionalParams = {
    * The default value is 'en_us'.
    */
   language_code?: LiteralUnion<TranscriptLanguageCode, string> | null;
+  /**
+   * The confidence threshold for the automatically detected language.
+   * An error will be returned if the langauge confidence is below this threshold.
+   * Defaults to 0.
+   */
+  language_confidence_threshold?: number;
   /**
    * Enable {@link https://www.assemblyai.com/docs/models/speech-recognition#automatic-language-detection | Automatic language detection }, either true or false.
    */
@@ -2914,7 +2930,7 @@ export type TranscriptOptionalParams = {
    */
   speakers_expected?: number | null;
   /**
-   * The speech model to use for the transcription. When `null`, the default model is used.
+   * The speech model to use for the transcription. When `null`, the "best" model is used.
    * @defaultValue null
    */
   speech_model?: SpeechModel | null;

src/utils/errors/realtime.ts

@@ -1,28 +1,44 @@
-enum RealtimeErrorType {
-  BadSampleRate = 4000,
-  AuthFailed = 4001,
-  // Both InsufficientFunds and FreeAccount error use 4002
-  InsufficientFundsOrFreeAccount = 4002,
-  NonexistentSessionId = 4004,
-  SessionExpired = 4008,
-  ClosedSession = 4010,
-  RateLimited = 4029,
-  UniqueSessionViolation = 4030,
-  SessionTimeout = 4031,
-  AudioTooShort = 4032,
-  AudioTooLong = 4033,
-  BadJson = 4100,
-  BadSchema = 4101,
-  TooManyStreams = 4102,
-  Reconnected = 4103,
-  ReconnectAttemptsExhausted = 1013,
-}
+const RealtimeErrorType = {
+  BadSampleRate: 4000,
+  AuthFailed: 4001,
+  /**
+   * @deprecated Use InsufficientFunds or FreeTierUser instead
+   */
+  InsufficientFundsOrFreeAccount: 4002,
+  InsufficientFunds: 4002,
+  FreeTierUser: 4003,
+  NonexistentSessionId: 4004,
+  SessionExpired: 4008,
+  ClosedSession: 4010,
+  RateLimited: 4029,
+  UniqueSessionViolation: 4030,
+  SessionTimeout: 4031,
+  AudioTooShort: 4032,
+  AudioTooLong: 4033,
+  AudioTooSmallToTranscode: 4034,
+  /**
+   * @deprecated Don't use
+   */
+  BadJson: 4100,
+  BadSchema: 4101,
+  TooManyStreams: 4102,
+  Reconnected: 4103,
+  /**
+   * @deprecated Don't use
+   */
+  ReconnectAttemptsExhausted: 1013,
+  WordBoostParameterParsingFailed: 4104,
+} as const;
 
-const RealtimeErrorMessages: Record<RealtimeErrorType, string> = {
+type RealtimeErrorTypeCodes =
+  (typeof RealtimeErrorType)[keyof typeof RealtimeErrorType];
+
+const RealtimeErrorMessages: Record<RealtimeErrorTypeCodes, string> = {
   [RealtimeErrorType.BadSampleRate]: "Sample rate must be a positive integer",
   [RealtimeErrorType.AuthFailed]: "Not Authorized",
-  [RealtimeErrorType.InsufficientFundsOrFreeAccount]:
-    "Insufficient funds or you are using a free account. This feature is paid-only and requires you to add a credit card. Please visit https://assemblyai.com/dashboard/ to add a credit card to your account.",
+  [RealtimeErrorType.InsufficientFunds]: "Insufficient funds",
+  [RealtimeErrorType.FreeTierUser]:
+    "This feature is paid-only and requires you to add a credit card. Please visit https://app.assemblyai.com/ to add a credit card to your account.",
   [RealtimeErrorType.NonexistentSessionId]: "Session ID does not exist",
   [RealtimeErrorType.SessionExpired]: "Session has expired",
   [RealtimeErrorType.ClosedSession]: "Session is closed",
@@ -31,14 +47,23 @@ const RealtimeErrorMessages: Record<RealtimeErrorType, string> = {
   [RealtimeErrorType.SessionTimeout]: "Session Timeout",
   [RealtimeErrorType.AudioTooShort]: "Audio too short",
   [RealtimeErrorType.AudioTooLong]: "Audio too long",
+  [RealtimeErrorType.AudioTooSmallToTranscode]: "Audio too small to transcode",
   [RealtimeErrorType.BadJson]: "Bad JSON",
   [RealtimeErrorType.BadSchema]: "Bad schema",
   [RealtimeErrorType.TooManyStreams]: "Too many streams",
-  [RealtimeErrorType.Reconnected]: "Reconnected",
+  [RealtimeErrorType.Reconnected]:
+    "This session has been reconnected. This WebSocket is no longer valid.",
   [RealtimeErrorType.ReconnectAttemptsExhausted]:
     "Reconnect attempts exhausted",
+  [RealtimeErrorType.WordBoostParameterParsingFailed]:
+    "Could not parse word boost parameter",
 };
 
 class RealtimeError extends Error {}
 
-export { RealtimeError, RealtimeErrorType, RealtimeErrorMessages };
+export {
+  RealtimeError,
+  RealtimeErrorType,
+  RealtimeErrorTypeCodes,
+  RealtimeErrorMessages,
+};