added openai

app.py

@@ -15,11 +15,10 @@ import tempfile
 from pydub import AudioSegment
 import requests
 import json
-from google.cloud import texttospeech
 
-from utils import get_credentials, get_google_tts
+from utils import get_google_credentials, get_google_tts, get_openai_tts
 
-os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = get_credentials()
+os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = get_google_credentials()
 
 def match_target_amplitude(sound, target_dBFS):
     change_in_dBFS = target_dBFS - sound.dBFS
@@ -58,7 +57,8 @@ AVAILABLE_MODELS = {
     'KOTOBA-SPEECH-JACOB': 'kotoba-speech-jacob',
     'BLANE-TTS': 'blane-tts',
     'AMITARO-VITS': 'amitaro-vits',
-    'GOOGLE-API': 'google-api'
+    'GOOGLE-TTS': 'google-tts',
+    'OPENAI-TTS': 'openai-tts'
 }
 
 SPACE_ID = os.getenv('SPACE_ID')
@@ -396,7 +396,8 @@ model_names = {
     'kotoba-speech-jacob': 'KOTOBA-SPEECH-v0.1-JACOB',
     'blane-tts': 'BLANE-TTS',
     'amitaro-vits': 'AMITARO-VITS',
-    'google-api': 'google-api'
+    'google-tts': 'google-tts',
+    'openai-tts': 'openai-tts'
     # 'styletts2': 'StyleTTS 2',
 }
 model_licenses = {
@@ -744,9 +745,12 @@ def synthandreturn(text):
                     print(model_args[model])
                     print(model_kwargs[model])
                     result = router.predict(*model_args[model], **model_kwargs[model])
-                elif model == "google-api":
+                elif model == "google-tts":
                     local_filename = '/tmp/' + str(mkuuid(None)) + '.wav'
                     result = get_google_tts(text, local_filename=local_filename)
+                elif model == "openai-tts":
+                    local_filename = '/tmp/' + str(mkuuid(None)) + '.wav'
+                    result = get_openai_tts(text, local_filename=local_filename)
                 else:
                     result = get_tts_file(text, model)
                     # URL to download the file from

requirements.txt

@@ -9,3 +9,4 @@ langdetect
 pydub
 gradio-space-ci @ git+https://huggingface.co/spaces/Wauplin/gradio-space-ci@0.2.1
 google-cloud-texttospeech
+openai

utils.py

@@ -3,8 +3,23 @@ import json
 import tempfile
 from google.cloud import texttospeech
 
+from pathlib import Path
+from openai import OpenAI
+
+def get_openai_tts(text, local_filename):
+    api_key = os.getenv("OPENAI_KEY")
+    client = OpenAI(api_key=api_key)
+
+    # speech_file_path = Path(__file__).parent / "speech.mp3"
+    response = client.audio.speech.create(
+        model="tts-1",
+        voice="alloy",
+        input=text
+    )
+    response.stream_to_file(local_filename)
+    return local_filename
 
-def get_credentials():
+def get_google_credentials():
     creds_json_str = os.getenv("GCP_CREDENTIAL_JSON")  # get json credentials stored as a string
 
     # create a temporary file