Add ability to play audio on server

mimic3_http/app.py

@@ -19,6 +19,8 @@ import dataclasses
 import json
 import logging
 import re
+import shlex
+import subprocess
 import typing
 from pathlib import Path
 from queue import Queue
@@ -163,7 +165,7 @@ def get_app(args: argparse.Namespace, request_queue: Queue, temp_dir: str):
         )
 
     @app.route("/api/tts", methods=["GET", "POST"])
-    async def app_tts() -> Response:
+    async def app_tts() -> typing.Union[Response, str]:
         """Speak text to WAV."""
         tts_args: typing.Dict[str, typing.Any] = {
             "length_scale": args.length_scale,
@@ -224,7 +226,15 @@ def get_app(args: argparse.Namespace, request_queue: Queue, temp_dir: str):
             TextToWavParams(text=text, **tts_args), no_cache=no_cache
         )
 
-        return Response(wav_bytes, mimetype="audio/wav")
+        audio_target = request.args.get("audioTarget", "client").strip().lower()
+        if audio_target == "client":
+            return Response(wav_bytes, mimetype="audio/wav")
+
+        # Play audio on server
+        play_cmd = shlex.split(args.play_program)
+        subprocess.run(play_cmd, input=wav_bytes, check=True)
+
+        return "OK"
 
     @app.route("/api/voices", methods=["GET"])
     async def api_voices():

mimic3_http/args.py

@@ -90,6 +90,9 @@ def get_args(argv=None) -> argparse.Namespace:
         "--default-voice",
         help="Default voice key to select in web interface",
     )
+    parser.add_argument(
+        "--play-program", default="aplay -q", help="Program to play WAV audio on server"
+    )
     parser.add_argument(
         "--no-show-openapi", action="store_true", help="Don't show OpenAPI link"
     )

mimic3_http/templates/index.html

@@ -115,6 +115,15 @@
                     <button id="copy-voice" class="btn btn-info btn-sm" title="Copy voice key to clipboard">Copy</button>
                 </div>
             </div>
+            <div class="row mt-3">
+                <div class="col-auto">
+                    <label for="audio-target" title="Audio target">Play audio on:</label>
+                    <select id="audio-target" name="audio-target">
+                      <option value="client" selected>Client</value>
+                      <option value="server">Server</value>
+                    </select>
+                </div>
+            </div>
             <div id="audio-message" class="row mt-3" hidden>
                 <div class="col">
                     <audio id="audio" preload="none" controls autoplay hidden></audio>
@@ -302,6 +311,8 @@
 
              var ssml = q('#ssml').checked || 'false'
 
+             var audioTarget = q('#audio-target').value || 'client'
+
              q('#audio-message').hidden = false
              q('#audio').pause()
 
@@ -313,7 +324,8 @@
                  '&noiseScale=' + encodeURIComponent(noiseScale) +
                  '&noiseW=' + encodeURIComponent(noiseW) +
                  '&lengthScale=' + encodeURIComponent(lengthScale) +
-                 '&ssml=' + encodeURIComponent(ssml),
+                 '&ssml=' + encodeURIComponent(ssml) +
+                 '&audioTarget=' + encodeURIComponent(audioTarget),
                  {cache: 'no-cache'})
 
              if (res.ok) {
@@ -322,8 +334,11 @@
 
                  q('#message').innerHTML = (elapsedTime / 1000).toFixed(3) + ' second(s)'
                  q('#speak-button').disabled = false
-                 q('#audio').src = URL.createObjectURL(blob)
-                 q('#audio').hidden = false
+
+                 if (audioTarget != 'server') {
+                     q('#audio').src = URL.createObjectURL(blob)
+                     q('#audio').hidden = false
+                 }
              } else {
                  message = await res.text()
                  q('#message').textContent = message

mimic3_tts/VERSION

@@ -1 +1 @@
-0.2.0
+0.2.1