From 0ec5a9ee303f91f6d66da9e91ab726ca27f28534 Mon Sep 17 00:00:00 2001
From: Daniel Faria <danielfariati@gmail.com>
Date: Sun, 12 Nov 2023 18:22:08 -0300
Subject: [PATCH] add support to audio/createSpeech API (#392)

---
 .../openai/audio/CreateSpeechRequest.java     | 45 +++++++++++++++++++
 .../theokanning/openai/client/OpenAiApi.java  |  4 ++
 .../openai/service/OpenAiService.java         |  5 +++
 .../theokanning/openai/service/AudioTest.java | 19 ++++++++
 4 files changed, 73 insertions(+)
 create mode 100644 api/src/main/java/com/theokanning/openai/audio/CreateSpeechRequest.java

diff --git a/api/src/main/java/com/theokanning/openai/audio/CreateSpeechRequest.java b/api/src/main/java/com/theokanning/openai/audio/CreateSpeechRequest.java
new file mode 100644
index 00000000..6d2e69ac
--- /dev/null
+++ b/api/src/main/java/com/theokanning/openai/audio/CreateSpeechRequest.java
@@ -0,0 +1,45 @@
+package com.theokanning.openai.audio;
+
+import com.fasterxml.jackson.annotation.JsonProperty;
+
+import lombok.AllArgsConstructor;
+import lombok.Builder;
+import lombok.Data;
+import lombok.NoArgsConstructor;
+import lombok.NonNull;
+
+@Builder
+@NoArgsConstructor
+@AllArgsConstructor
+@Data
+public class CreateSpeechRequest {
+
+    /**
+     * The name of the model to use.
+     */
+    @NonNull
+    String model;
+
+    /**
+     * The text to generate audio for. The maximum length is 4096 characters.
+     */
+    @NonNull
+    String input;
+
+    /**
+     * The voice to use when generating the audio.
+     */
+    @NonNull
+    String voice;
+
+    /**
+     * The format to audio in. Supported formats are mp3, opus, aac, and flac. Defaults to mp3.
+     */
+    @JsonProperty("response_format")
+    String responseFormat;
+
+    /**
+     * The speed of the generated audio. Select a value from 0.25 to 4.0. Defaults to 1.0.
+     */
+    Double speed;
+}
diff --git a/client/src/main/java/com/theokanning/openai/client/OpenAiApi.java b/client/src/main/java/com/theokanning/openai/client/OpenAiApi.java
index 497dacd5..f2665ee2 100644
--- a/client/src/main/java/com/theokanning/openai/client/OpenAiApi.java
+++ b/client/src/main/java/com/theokanning/openai/client/OpenAiApi.java
@@ -2,6 +2,7 @@
 
 import com.theokanning.openai.DeleteResult;
 import com.theokanning.openai.OpenAiResponse;
+import com.theokanning.openai.audio.CreateSpeechRequest;
 import com.theokanning.openai.audio.TranscriptionResult;
 import com.theokanning.openai.audio.TranslationResult;
 import com.theokanning.openai.billing.BillingUsage;
@@ -149,6 +150,9 @@ public interface OpenAiApi {
     @POST("/v1/audio/translations")
     Single<TranslationResult> createTranslation(@Body RequestBody requestBody);
 
+    @POST("/v1/audio/speech")
+    Single<ResponseBody> createSpeech(@Body CreateSpeechRequest requestBody);
+
     @POST("/v1/moderations")
     Single<ModerationResult> createModeration(@Body ModerationRequest request);
 
diff --git a/service/src/main/java/com/theokanning/openai/service/OpenAiService.java b/service/src/main/java/com/theokanning/openai/service/OpenAiService.java
index cedbd805..ea59417e 100644
--- a/service/src/main/java/com/theokanning/openai/service/OpenAiService.java
+++ b/service/src/main/java/com/theokanning/openai/service/OpenAiService.java
@@ -8,6 +8,7 @@
 import com.theokanning.openai.DeleteResult;
 import com.theokanning.openai.OpenAiError;
 import com.theokanning.openai.OpenAiHttpException;
+import com.theokanning.openai.audio.CreateSpeechRequest;
 import com.theokanning.openai.audio.CreateTranscriptionRequest;
 import com.theokanning.openai.audio.CreateTranslationRequest;
 import com.theokanning.openai.audio.TranscriptionResult;
@@ -347,6 +348,10 @@ public ModerationResult createModeration(ModerationRequest request) {
         return execute(api.createModeration(request));
     }
 
+    public ResponseBody createSpeech(CreateSpeechRequest request) {
+        return execute(api.createSpeech(request));
+    }
+
     /**
      * Calls the Open AI api, returns the response, and parses error messages if the request fails
      */
diff --git a/service/src/test/java/com/theokanning/openai/service/AudioTest.java b/service/src/test/java/com/theokanning/openai/service/AudioTest.java
index d5a54a23..9cb083de 100644
--- a/service/src/test/java/com/theokanning/openai/service/AudioTest.java
+++ b/service/src/test/java/com/theokanning/openai/service/AudioTest.java
@@ -1,13 +1,18 @@
 package com.theokanning.openai.service;
 
+import com.theokanning.openai.audio.CreateSpeechRequest;
 import com.theokanning.openai.audio.CreateTranscriptionRequest;
 import com.theokanning.openai.audio.CreateTranslationRequest;
 import com.theokanning.openai.audio.TranscriptionResult;
 import com.theokanning.openai.audio.TranslationResult;
 import org.junit.jupiter.api.Test;
 
+import java.io.IOException;
 import java.time.Duration;
 
+import okhttp3.MediaType;
+import okhttp3.ResponseBody;
+
 import static org.junit.jupiter.api.Assertions.*;
 
 
@@ -69,4 +74,18 @@ void createTranslationVerbose() {
         assertTrue(result.getDuration() > 0);
         assertEquals(1, result.getSegments().size());
     }
+
+    @Test
+    void createSpeech() throws IOException {
+        CreateSpeechRequest createSpeechRequest = CreateSpeechRequest.builder()
+                .model("tts-1")
+                .input("Hello World.")
+                .voice("alloy")
+                .build();
+
+        final ResponseBody speech = service.createSpeech(createSpeechRequest);
+        assertNotNull(speech);
+        assertEquals(MediaType.get("audio/mpeg"), speech.contentType());
+        assertTrue(speech.bytes().length > 0);
+    }
 }