Add sample text for voices

mimic3_http/app.py

@@ -18,6 +18,7 @@ import asyncio
 import dataclasses
 import json
 import logging
+import re
 import typing
 from pathlib import Path
 from queue import Queue
@@ -36,7 +37,7 @@ from quart import (
 from swagger_ui import api_doc
 
 from mimic3_tts import DEFAULT_VOICE, Mimic3Settings, Mimic3TextToSpeechSystem
-from mimic3_tts.utils import LANG_NAMES
+from mimic3_tts.utils import LANG_NAMES, SAMPLE_SENTENCES
 
 from ._resources import _DIR, _PACKAGE
 from .args import _MISSING
@@ -234,6 +235,11 @@ def get_app(args: argparse.Namespace, request_queue: Queue, temp_dir: str):
         # Add more fields to voices
         for voice_dict in voice_dicts:
             voice_lang = voice_dict["language"]
+
+            # en_US => en
+            short_lang = voice_lang.split("_", maxsplit=1)[0]
+
+            # en_US => English (US)
             lang_name = LANG_NAMES.get(voice_lang, voice_lang)
 
             if isinstance(lang_name, str):
@@ -246,6 +252,10 @@ def get_app(args: argparse.Namespace, request_queue: Queue, temp_dir: str):
             voice_dict["language_native"] = native_lang
             voice_dict["language_english"] = english_lang
 
+            sample_text = SAMPLE_SENTENCES.get(short_lang, "")
+            sample_text = re.sub(r"\s+", " ", sample_text)
+            voice_dict["sample_text"] = sample_text
+
         return jsonify(voice_dicts)
 
     @app.route("/process", methods=["GET", "POST"])

mimic3_http/templates/index.html

@@ -235,6 +235,7 @@
         <!-- Bootstrap core JavaScript -->
         <script>
          var voicesInfo = {}
+         var isUserText = false
 
          function q(selector) {return document.querySelector(selector)}
          q('#text').focus()
@@ -378,6 +379,12 @@
                 )
              }
 
+             var textArea = q('#text')
+             if ((textArea.value.length == 0) || !isUserText) {
+                 textArea.value = voice.sample_text
+                 isUserText = false
+             }
+
              // Update inference settings
              if (voice.properties) {
                  q('#length-scale').value = voice.properties.length_scale || 1.0
@@ -433,18 +440,22 @@
                  })
          }
 
-        function copyVoiceKey() {
-            var voiceName = q('#voice-name')
-            var speakerList = q('#speaker-list')
+         function copyVoiceKey() {
+             var voiceName = q('#voice-name')
+             var speakerList = q('#speaker-list')
 
-            var voiceKey = voiceName.options[voiceName.selectedIndex].value
-            if (speakerList.options.length > 1) {
-                voiceKey += '#' + speakerList.options[speakerList.selectedIndex].value
-            }
-             
+             var voiceKey = voiceName.options[voiceName.selectedIndex].value
+             if (speakerList.options.length > 1) {
+                 voiceKey += '#' + speakerList.options[speakerList.selectedIndex].value
+             }
+              
 
-            navigator.clipboard.writeText(voiceKey)
-        }
+             navigator.clipboard.writeText(voiceKey)
+         }
+
+         function textChanged() {
+             isUserText = true
+         }
 
          window.addEventListener('load', function() {
              loadVoices()
@@ -453,6 +464,7 @@
              q('#voice-name').addEventListener('change', nameChanged)
 
              q('#copy-voice').addEventListener('click', copyVoiceKey)
+             q('#text').addEventListener('change', textChanged)
          })
         </script>