feat(client-polly): Add new engine - long-form - dedicated for longer…

… content, such as news articles, training materials, or marketing videos.
aws · Nov 16, 2023 · 367e933 · 367e933
1 parent 473c3e9
commit 367e933
Show file tree

Hide file tree

Showing 7 changed files with 54 additions and 35 deletions.
diff --git a/clients/client-polly/src/commands/DescribeVoicesCommand.ts b/clients/client-polly/src/commands/DescribeVoicesCommand.ts
@@ -60,7 +60,7 @@ export interface DescribeVoicesCommandOutput extends DescribeVoicesOutput, __Met
  * // const { PollyClient, DescribeVoicesCommand } = require("@aws-sdk/client-polly"); // CommonJS import
  * const client = new PollyClient(config);
  * const input = { // DescribeVoicesInput
- *   Engine: "standard" || "neural",
+ *   Engine: "standard" || "neural" || "long-form",
  *   LanguageCode: "arb" || "cmn-CN" || "cy-GB" || "da-DK" || "de-DE" || "en-AU" || "en-GB" || "en-GB-WLS" || "en-IN" || "en-US" || "es-ES" || "es-MX" || "es-US" || "fr-CA" || "fr-FR" || "is-IS" || "it-IT" || "ja-JP" || "hi-IN" || "ko-KR" || "nb-NO" || "nl-NL" || "pl-PL" || "pt-BR" || "pt-PT" || "ro-RO" || "ru-RU" || "sv-SE" || "tr-TR" || "en-NZ" || "en-ZA" || "ca-ES" || "de-AT" || "yue-CN" || "ar-AE" || "fi-FI" || "en-IE" || "nl-BE" || "fr-BE",
  *   IncludeAdditionalLanguageCodes: true || false,
  *   NextToken: "STRING_VALUE",
@@ -79,7 +79,7 @@ export interface DescribeVoicesCommandOutput extends DescribeVoicesOutput, __Met
  * //         "arb" || "cmn-CN" || "cy-GB" || "da-DK" || "de-DE" || "en-AU" || "en-GB" || "en-GB-WLS" || "en-IN" || "en-US" || "es-ES" || "es-MX" || "es-US" || "fr-CA" || "fr-FR" || "is-IS" || "it-IT" || "ja-JP" || "hi-IN" || "ko-KR" || "nb-NO" || "nl-NL" || "pl-PL" || "pt-BR" || "pt-PT" || "ro-RO" || "ru-RU" || "sv-SE" || "tr-TR" || "en-NZ" || "en-ZA" || "ca-ES" || "de-AT" || "yue-CN" || "ar-AE" || "fi-FI" || "en-IE" || "nl-BE" || "fr-BE",
  * //       ],
  * //       SupportedEngines: [ // EngineList
- * //         "standard" || "neural",
+ * //         "standard" || "neural" || "long-form",
  * //       ],
  * //     },
  * //   ],

diff --git a/clients/client-polly/src/commands/GetSpeechSynthesisTaskCommand.ts b/clients/client-polly/src/commands/GetSpeechSynthesisTaskCommand.ts
@@ -54,7 +54,7 @@ export interface GetSpeechSynthesisTaskCommandOutput extends GetSpeechSynthesisT
  * const response = await client.send(command);
  * // { // GetSpeechSynthesisTaskOutput
  * //   SynthesisTask: { // SynthesisTask
- * //     Engine: "standard" || "neural",
+ * //     Engine: "standard" || "neural" || "long-form",
  * //     TaskId: "STRING_VALUE",
  * //     TaskStatus: "scheduled" || "inProgress" || "completed" || "failed",
  * //     TaskStatusReason: "STRING_VALUE",

diff --git a/clients/client-polly/src/commands/ListSpeechSynthesisTasksCommand.ts b/clients/client-polly/src/commands/ListSpeechSynthesisTasksCommand.ts
@@ -57,7 +57,7 @@ export interface ListSpeechSynthesisTasksCommandOutput extends ListSpeechSynthes
  * //   NextToken: "STRING_VALUE",
  * //   SynthesisTasks: [ // SynthesisTasks
  * //     { // SynthesisTask
- * //       Engine: "standard" || "neural",
+ * //       Engine: "standard" || "neural" || "long-form",
  * //       TaskId: "STRING_VALUE",
  * //       TaskStatus: "scheduled" || "inProgress" || "completed" || "failed",
  * //       TaskStatusReason: "STRING_VALUE",

diff --git a/clients/client-polly/src/commands/StartSpeechSynthesisTaskCommand.ts b/clients/client-polly/src/commands/StartSpeechSynthesisTaskCommand.ts
@@ -54,7 +54,7 @@ export interface StartSpeechSynthesisTaskCommandOutput extends StartSpeechSynthe
  * // const { PollyClient, StartSpeechSynthesisTaskCommand } = require("@aws-sdk/client-polly"); // CommonJS import
  * const client = new PollyClient(config);
  * const input = { // StartSpeechSynthesisTaskInput
- *   Engine: "standard" || "neural",
+ *   Engine: "standard" || "neural" || "long-form",
  *   LanguageCode: "arb" || "cmn-CN" || "cy-GB" || "da-DK" || "de-DE" || "en-AU" || "en-GB" || "en-GB-WLS" || "en-IN" || "en-US" || "es-ES" || "es-MX" || "es-US" || "fr-CA" || "fr-FR" || "is-IS" || "it-IT" || "ja-JP" || "hi-IN" || "ko-KR" || "nb-NO" || "nl-NL" || "pl-PL" || "pt-BR" || "pt-PT" || "ro-RO" || "ru-RU" || "sv-SE" || "tr-TR" || "en-NZ" || "en-ZA" || "ca-ES" || "de-AT" || "yue-CN" || "ar-AE" || "fi-FI" || "en-IE" || "nl-BE" || "fr-BE",
  *   LexiconNames: [ // LexiconNameList
  *     "STRING_VALUE",
@@ -75,7 +75,7 @@ export interface StartSpeechSynthesisTaskCommandOutput extends StartSpeechSynthe
  * const response = await client.send(command);
  * // { // StartSpeechSynthesisTaskOutput
  * //   SynthesisTask: { // SynthesisTask
- * //     Engine: "standard" || "neural",
+ * //     Engine: "standard" || "neural" || "long-form",
  * //     TaskId: "STRING_VALUE",
  * //     TaskStatus: "scheduled" || "inProgress" || "completed" || "failed",
  * //     TaskStatusReason: "STRING_VALUE",

diff --git a/clients/client-polly/src/commands/SynthesizeSpeechCommand.ts b/clients/client-polly/src/commands/SynthesizeSpeechCommand.ts
@@ -57,7 +57,7 @@ export interface SynthesizeSpeechCommandOutput extends Omit<SynthesizeSpeechOutp
  * // const { PollyClient, SynthesizeSpeechCommand } = require("@aws-sdk/client-polly"); // CommonJS import
  * const client = new PollyClient(config);
  * const input = { // SynthesizeSpeechInput
- *   Engine: "standard" || "neural",
+ *   Engine: "standard" || "neural" || "long-form",
  *   LanguageCode: "arb" || "cmn-CN" || "cy-GB" || "da-DK" || "de-DE" || "en-AU" || "en-GB" || "en-GB-WLS" || "en-IN" || "en-US" || "es-ES" || "es-MX" || "es-US" || "fr-CA" || "fr-FR" || "is-IS" || "it-IT" || "ja-JP" || "hi-IN" || "ko-KR" || "nb-NO" || "nl-NL" || "pl-PL" || "pt-BR" || "pt-PT" || "ro-RO" || "ru-RU" || "sv-SE" || "tr-TR" || "en-NZ" || "en-ZA" || "ca-ES" || "de-AT" || "yue-CN" || "ar-AE" || "fi-FI" || "en-IE" || "nl-BE" || "fr-BE",
  *   LexiconNames: [ // LexiconNameList
  *     "STRING_VALUE",

diff --git a/clients/client-polly/src/models/models_0.ts b/clients/client-polly/src/models/models_0.ts
@@ -70,6 +70,7 @@ export class ServiceFailureException extends __BaseException {
  * @enum
  */
 export const Engine = {
+  LONG_FORM: "long-form",
   NEURAL: "neural",
   STANDARD: "standard",
 } as const;
@@ -136,8 +137,9 @@ export type LanguageCode = (typeof LanguageCode)[keyof typeof LanguageCode];
 export interface DescribeVoicesInput {
   /**
    * @public
-   * <p>Specifies the engine (<code>standard</code> or <code>neural</code>)
-   *       used by Amazon Polly when processing input text for speech synthesis. </p>
+   * <p>Specifies the engine (<code>standard</code>, <code>neural</code> or
+   *       <code>long-form</code>) used by Amazon Polly when processing input text for
+   *       speech synthesis. </p>
    */
   Engine?: Engine;
 
@@ -342,8 +344,8 @@ export interface Voice {
 
   /**
    * @public
-   * <p>Specifies which engines (<code>standard</code> or <code>neural</code>)
-   *       that are supported by a given voice.</p>
+   * <p>Specifies which engines (<code>standard</code>, <code>neural</code> or
+   *       <code>long-form</code>) are supported by a given voice.</p>
    */
   SupportedEngines?: Engine[];
 }
@@ -589,10 +591,10 @@ export type TextType = (typeof TextType)[keyof typeof TextType];
 export interface SynthesisTask {
   /**
    * @public
-   * <p>Specifies the engine (<code>standard</code> or <code>neural</code>)
-   *       for Amazon Polly to use when processing input text for speech synthesis. Using a
-   *       voice that is not supported for the engine selected will result in an
-   *       error.</p>
+   * <p>Specifies the engine (<code>standard</code>, <code>neural</code> or
+   *       <code>long-form</code>) for Amazon Polly to use when processing input text for
+   *       speech synthesis. Using a voice that is not supported for the engine
+   *       selected will result in an error.</p>
    */
   Engine?: Engine;
 
@@ -661,7 +663,8 @@ export interface SynthesisTask {
    * <p>The audio frequency specified in Hz.</p>
    *          <p>The valid values for mp3 and ogg_vorbis are "8000", "16000", "22050",
    *       and "24000". The default value for standard voices is "22050". The default
-   *       value for neural voices is "24000".</p>
+   *       value for neural voices is "24000". The default value for long-form voices
+   *       is "24000".</p>
    *          <p>Valid values for pcm are "8000" and "16000" The default value is
    *       "16000". </p>
    */
@@ -1176,10 +1179,10 @@ export class SsmlMarksNotSupportedForTextTypeException extends __BaseException {
 export interface StartSpeechSynthesisTaskInput {
   /**
    * @public
-   * <p>Specifies the engine (<code>standard</code> or <code>neural</code>)
-   *       for Amazon Polly to use when processing input text for speech synthesis. Using a
-   *       voice that is not supported for the engine selected will result in an
-   *       error.</p>
+   * <p>Specifies the engine (<code>standard</code>, <code>neural</code> or
+   *       <code>long-form</code>) for Amazon Polly to use when processing input text for
+   *       speech synthesis. Using a voice that is not supported for the engine
+   *       selected will result in an error.</p>
    */
   Engine?: Engine;
 
@@ -1229,7 +1232,8 @@ export interface StartSpeechSynthesisTaskInput {
    * <p>The audio frequency specified in Hz.</p>
    *          <p>The valid values for mp3 and ogg_vorbis are "8000", "16000", "22050",
    *       and "24000". The default value for standard voices is "22050". The default
-   *       value for neural voices is "24000".</p>
+   *       value for neural voices is "24000". The default value for long-form voices
+   *       is "24000".</p>
    *          <p>Valid values for pcm are "8000" and "16000" The default value is
    *       "16000". </p>
    */
@@ -1313,19 +1317,27 @@ export class TextLengthExceededException extends __BaseException {
 export interface SynthesizeSpeechInput {
   /**
    * @public
-   * <p>Specifies the engine (<code>standard</code> or <code>neural</code>)
-   *       for Amazon Polly to use when processing input text for speech synthesis. For
-   *       information on Amazon Polly voices and which voices are available in
-   *       standard-only, NTTS-only, and both standard and NTTS formats, see <a href="https://docs.aws.amazon.com/polly/latest/dg/voicelist.html">Available Voices</a>.</p>
+   * <p>Specifies the engine (<code>standard</code>, <code>neural</code> or
+   *       <code>long-form</code>) for Amazon Polly to use when processing input text for
+   *       speech synthesis. For information on Amazon Polly voices and which voices are
+   *       available for each engine, see <a href="https://docs.aws.amazon.com/polly/latest/dg/voicelist.html">Available Voices</a>.</p>
    *          <p>
    *             <b>NTTS-only voices</b>
    *          </p>
    *          <p>When using NTTS-only voices such as Kevin (en-US), this parameter is
    *       required and must be set to <code>neural</code>. If the engine is not
    *       specified, or is set to <code>standard</code>, this will result in an
    *       error. </p>
+   *          <p>
+   *             <b>long-form-only voices</b>
+   *          </p>
+   *          <p>When using long-form-only voices such as Danielle (en-US), this
+   *       parameter is required and must be set to <code>long-form</code>. If the
+   *       engine is not specified, or is set to <code>standard</code> or
+   *       <code>neural</code>, this will result in an error. </p>
    *          <p>Type: String</p>
-   *          <p>Valid Values: <code>standard</code> | <code>neural</code>
+   *          <p>Valid Values: <code>standard</code> | <code>neural</code> |
+   *       <code>long-form</code>
    *          </p>
    *          <p>Required: Yes</p>
    *          <p>
@@ -1375,7 +1387,8 @@ export interface SynthesizeSpeechInput {
    * <p>The audio frequency specified in Hz.</p>
    *          <p>The valid values for mp3 and ogg_vorbis are "8000", "16000", "22050",
    *       and "24000". The default value for standard voices is "22050". The default
-   *       value for neural voices is "24000".</p>
+   *       value for neural voices is "24000". The default value for long-form voices
+   *       is "24000".</p>
    *          <p>Valid values for pcm are "8000" and "16000" The default value is
    *       "16000". </p>
    */