Sync from internal repo (2024/06/12)

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changelog
 
+## [4.4.7]
+
+- Rename `TranscriptService.redactions` function to `TranscriptService.redactedAudio`.
+- Add `TranscriptService.redactedAudioFile` function.
+- Add `workerd` export to fix `cache` issue with `fetch` on Cloudflare Workers.
+
 ## [4.4.6]
 
 - Fix Rollup exports so \_\_SDK_VERSION\_\_ is properly replaced with the version of the SDK.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "assemblyai",
-  "version": "4.4.6",
+  "version": "4.4.7",
   "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"
@@ -16,7 +16,7 @@
         "types": "./dist/exports/index.d.ts",
         "default": "./dist/deno.mjs"
       },
-      "workerd": "./dist/index.mjs",
+      "workerd": "./dist/workerd.mjs",
       "browser": "./dist/browser.mjs",
       "react-native": "./dist/browser.mjs",
       "node": {
@@ -38,6 +38,10 @@
     "./package.json": "./package.json"
   },
   "imports": {
+    "#fetch": {
+      "workerd": "./src/polyfills/fetch/workerd.ts",
+      "default": "./src/polyfills/fetch/default.ts"
+    },
     "#fs": {
       "node": "./src/polyfills/fs/node.ts",
       "bun": "./src/polyfills/fs/bun.ts",

rollup.config.js

@@ -117,6 +117,21 @@ module.exports = [
       },
     ],
   },
+  {
+    input: "src/exports/index.ts",
+    plugins: [
+      replacePlugin,
+      ts({ compilerOptions: { customConditions: ["workerd"] } }),
+    ],
+    external: ["ws"],
+    output: [
+      {
+        file: `./dist/workerd.mjs`,
+        format: "es",
+        exports: "named",
+      },
+    ],
+  },
   // Browser ESM build
   {
     input: "src/exports/index.ts",

src/polyfills/fetch/default.ts

@@ -0,0 +1,3 @@
+export const DEFAULT_FETCH_INIT: Record<string, unknown> = {
+  cache: "no-store",
+};

src/polyfills/fetch/workerd.ts

@@ -0,0 +1 @@
+export const DEFAULT_FETCH_INIT: Record<string, unknown> = {};

src/services/base.ts

@@ -1,5 +1,5 @@
-import { BaseServiceParams } from "..";
-import { Error as JsonError } from "..";
+import { DEFAULT_FETCH_INIT } from "#fetch";
+import { BaseServiceParams, Error as JsonError } from "..";
 import { buildUserAgent } from "../utils/userAgent";
 
 /**
@@ -22,13 +22,14 @@ export abstract class BaseService {
     input: string,
     init?: RequestInit | undefined,
   ): Promise<Response> {
-    init = init ?? {};
-    let headers = init.headers ?? {};
-    headers = {
+    init = { ...DEFAULT_FETCH_INIT, ...init };
+    let headers = {
       Authorization: this.params.apiKey,
       "Content-Type": "application/json",
-      ...init.headers,
     };
+    if (DEFAULT_FETCH_INIT?.headers)
+      headers = { ...headers, ...DEFAULT_FETCH_INIT.headers };
+    if (init?.headers) headers = { ...headers, ...init.headers };
 
     if (this.userAgent) {
       (headers as Record<string, string>)["User-Agent"] = this.userAgent;
@@ -40,7 +41,6 @@ export abstract class BaseService {
     }
     init.headers = headers;
 
-    init.cache = "no-store";
     if (!input.startsWith("http")) input = this.params.baseUrl + input;
 
     const response = await fetch(input, init);

src/services/transcripts/index.ts

@@ -16,6 +16,7 @@ import {
   TranscribeOptions,
   SubmitParams,
   SpeechModel,
+  RedactedAudioFile,
 } from "../..";
 import { FileService } from "../files";
 import { getPath } from "../../utils/path";
@@ -244,15 +245,47 @@ export class TranscriptService extends BaseService {
   }
 
   /**
-   * Retrieve redactions of a transcript.
+   * Retrieve the redacted audio URL of a transcript.
    * @param id - The identifier of the transcript.
-   * @returns A promise that resolves to the subtitles text.
+   * @returns A promise that resolves to the details of the redacted audio.
+   * @deprecated Use `redactedAudio` instead.
    */
   redactions(id: string): Promise<RedactedAudioResponse> {
+    return this.redactedAudio(id);
+  }
+
+  /**
+   * Retrieve the redacted audio URL of a transcript.
+   * @param id - The identifier of the transcript.
+   * @returns A promise that resolves to the details of the redacted audio.
+   */
+  redactedAudio(id: string): Promise<RedactedAudioResponse> {
     return this.fetchJson<RedactedAudioResponse>(
       `/v2/transcript/${id}/redacted-audio`,
     );
   }
+
+  /**
+   * Retrieve the redacted audio file of a transcript.
+   * @param id - The identifier of the transcript.
+   * @returns A promise that resolves to the fetch HTTP response of the redacted audio file.
+   */
+  async redactedAudioFile(id: string): Promise<RedactedAudioFile> {
+    const { redacted_audio_url, status } = await this.redactedAudio(id);
+    if (status !== "redacted_audio_ready") {
+      throw new Error(`Redacted audio status is ${status}`);
+    }
+    const response = await fetch(redacted_audio_url);
+    if (!response.ok) {
+      throw new Error(`Failed to fetch redacted audio: ${response.statusText}`);
+    }
+    return {
+      arrayBuffer: response.arrayBuffer.bind(response),
+      blob: response.blob.bind(response),
+      body: response.body,
+      bodyUsed: response.bodyUsed,
+    };
+  }
 }
 
 function deprecateConformer2(params: { speech_model?: SpeechModel | null }) {

src/types/transcripts/index.ts

@@ -54,3 +54,13 @@ export type SubmitParams = TranscribeParams;
  * The options to transcribe an audio file, including polling options.
  */
 export type TranscribeOptions = PollingOptions;
+
+/**
+ * The PII redacted audio file, transmitted over the network.
+ */
+export type RedactedAudioFile = {
+  arrayBuffer: () => Promise<ArrayBuffer>;
+  blob: () => Promise<Blob>;
+  body: ReadableStream<Uint8Array> | null;
+  bodyUsed: boolean;
+};

tests/integration/transcript.test.ts

@@ -123,7 +123,7 @@ describe("transcript", () => {
     expect(subtitle).toBeTruthy();
   });
 
-  it("should get redactions", async () => {
+  it("should get redacted audio", async () => {
     const transcript = await client.transcripts.transcribe({
       audio: remoteAudioUrl,
       redact_pii: true,
@@ -132,19 +132,33 @@ describe("transcript", () => {
       redact_pii_policies: ["medical_condition"],
       redact_pii_sub: "hash",
     });
-    const redactedAudioResponse = await client.transcripts.redactions(
+    const redactedAudioResponse = await client.transcripts.redactedAudio(
       transcript.id,
     );
     expect(redactedAudioResponse.status).toBe("redacted_audio_ready");
     expect(redactedAudioResponse.redacted_audio_url).toBeTruthy();
-  });
+  }, 30_000);
+
+  it("should get redacted audio file", async () => {
+    const transcript = await client.transcripts.transcribe({
+      audio: remoteAudioUrl,
+      redact_pii: true,
+      redact_pii_audio: true,
+      redact_pii_audio_quality: "wav",
+      redact_pii_policies: ["medical_condition"],
+      redact_pii_sub: "hash",
+    });
+    const redactedAudioResponse = await client.transcripts.redactedAudioFile(
+      transcript.id,
+    );
+    expect((await redactedAudioResponse.blob()).size).toBeGreaterThan(0);
+  }, 60_000);
 
   it("should word search", async () => {
     const searchResponse = await client.transcripts.wordSearch(
       knownTranscriptId,
       ["Giants"],
     );
-    console.log(searchResponse);
     expect(searchResponse.id).toBe(knownTranscriptId);
     expect(searchResponse.total_count).toBeGreaterThan(0);
     expect(Array.isArray(searchResponse.matches)).toBeTruthy();

tests/unit/transcript.test.ts

@@ -419,7 +419,7 @@ describe("transcript", () => {
     expect(subtitle).toBeTruthy();
   });
 
-  it("should get redactions", async () => {
+  it("should get redacted audio", async () => {
     fetchMock.doMockOnceIf(
       requestMatches({
         url: `/v2/transcript/${transcriptId}/redacted-audio`,
@@ -431,11 +431,28 @@ describe("transcript", () => {
       }),
     );
     const redactedAudioResponse =
-      await assembly.transcripts.redactions(transcriptId);
+      await assembly.transcripts.redactedAudio(transcriptId);
     expect(redactedAudioResponse.status).toBe("redacted_audio_ready");
     expect(redactedAudioResponse.redacted_audio_url).toBeTruthy();
   });
 
+  it("should get redacted audio file", async () => {
+    fetchMock.doMockOnceIf(
+      requestMatches({
+        url: `/v2/transcript/${transcriptId}/redacted-audio`,
+        method: "GET",
+      }),
+      JSON.stringify({
+        status: "redacted_audio_ready",
+        redacted_audio_url: "https://some-url.com",
+      }),
+    );
+    fetchMock.doMockOnceIf("https://some-url.com/", "audio data");
+    const redactedAudioResponse =
+      await assembly.transcripts.redactedAudioFile(transcriptId);
+    expect(redactedAudioResponse).toBeTruthy();
+  });
+
   it("should word search", async () => {
     fetchMock.mockResponseOnce(
       JSON.stringify({