refactor(tts): use bitbots_tts consistently (#529)

Author: jaagut

bitbots_misc/bitbots_tts/bitbots_tts/tts.py

@@ -8,6 +8,7 @@
 
 import rclpy
 import requests
+from ament_index_python import get_package_prefix
 from rcl_interfaces.msg import Parameter, SetParametersResult
 from rclpy.node import Node
 from rclpy.publisher import Publisher
@@ -24,6 +25,13 @@ def speak(text: str, publisher: Publisher, priority: int = 20, speaking_active:
         publisher.publish(msg)
 
 
+def say(text: str) -> None:
+    """Start the shell `say.sh` script to output given text with mimic3. Beware: this is blocking."""
+    script_path = os.path.join(get_package_prefix("bitbots_tts"), "lib/bitbots_tts/say.sh")
+    process = subprocess.Popen((script_path, text))
+    process.wait()
+
+
 class Speaker(Node):
     """
     Uses tts to say messages from the speak topic.
@@ -50,17 +58,6 @@ def __init__(self) -> None:
         # Callback for parameter changes
         self.add_on_set_parameters_callback(self.on_set_parameters)
 
-        # Mapping from robot name to voice name
-        self.robot_voice_mapping = {
-            "amy": "en_US/vctk_low",
-            "donna": "en_US/vctk_low",
-            "jack": "en_UK/apope_low",
-            "melody": "en_US/vctk_low",
-            "rory": "en_UK/apope_low",
-        }
-
-        self.robot_speed_mapping = {"amy": 2.2, "donna": 2.2, "jack": 1.0, "melody": 2.2, "rory": 1.0}
-
         # Subscribe to the speak topic
         self.create_subscription(Audio, "speak", self.speak_cb, 10)
 
@@ -92,30 +89,15 @@ def on_set_parameters(self, parameters: List[Parameter]) -> SetParametersResult:
         return SetParametersResult(successful=True)
 
     def run_speaker(self) -> None:
-        """Continously checks the queue and speaks the next message."""
+        """Continuously checks the queue and speaks the next message."""
         # Check if there is a message in the queue
         if len(self.prio_queue) > 0:
             # Get the next message and speak it
             text, _ = self.prio_queue.pop(0)
-            self.say(text)
-
-    def say(self, text: str) -> None:
-        """Speak this specific text."""
-        # Get the voice name from the environment variable ROBOT_NAME or use the default voice if it's not set
-        voice = self.robot_voice_mapping.get(os.getenv("ROBOT_NAME"), "en_US/vctk_low")
-        # Get the speed for the given robot or use the default speed if no robot name is set
-        speed = self.robot_speed_mapping.get(os.getenv("ROBOT_NAME"), 2.2)
-        try:
-            # Generate the speech with mimic
-            mimic_subprocess = subprocess.Popen(
-                ("mimic3", "--remote", "--voice", voice, "--length-scale", str(speed), text), stdout=subprocess.PIPE
-            )
-            # Play the audio from the previous process with aplay
-            aplay_subprocess = subprocess.Popen(("aplay", "-"), stdin=mimic_subprocess.stdout, stdout=subprocess.PIPE)
-            # Wait for the process to finish
-            aplay_subprocess.wait()
-        except OSError:
-            self.get_logger().error(str(traceback.format_exc()))
+            try:
+                say(text)
+            except OSError:
+                self.get_logger().error(str(traceback.format_exc()))
 
     def speak_cb(self, msg: Audio) -> None:
         """Handles incoming msg on speak topic."""

bitbots_misc/bitbots_tts/launch/tts.launch

@@ -1,5 +1,7 @@
 <launch>
-    <executable cmd="mimic3-server --preload-voice en_US/vctk_low --cache-dir $(env HOME)/.cache/mimic3/" name="mimic3-server" output="screen"/>
+    <!-- Do not launch mimic3-server on robot, as it is running as a systemd service -->
+    <executable unless="$(env IS_ROBOT)" cmd="mimic3-server --preload-voice en_US/vctk_low --preload-voice en_UK/apope_low --cache-dir $(env HOME)/.cache/mimic3/" name="mimic3-server" output="screen"/>
+
     <node name="bitbots_tts" pkg="bitbots_tts" exec="tts">
         <param from="$(find-pkg-share bitbots_tts)/config/tts_config.yaml"/>
     </node>

bitbots_misc/bitbots_tts/scripts/say.sh

@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+set -eEuo pipefail
+
+ROBOT_NAME="${ROBOT_NAME:-}"
+
+# Mapping robot name to voice and speed
+case "$ROBOT_NAME" in
+  "jack"|"rory")
+    voice="en_UK/apope_low"
+    speed=1.0
+    ;;
+  "amy"|"donna"|"melody"|"rose")
+    voice="en_US/vctk_low"
+    speed=1.7
+    ;;
+  *)
+    echo "Unknown robot: '$ROBOT_NAME', using default female voice"
+    voice="en_US/vctk_low"
+    speed=1.7
+    ;;
+esac
+
+text="$1"
+if [ -z "$text" ]; then
+  echo "No text provided!"
+  exit 1
+fi
+
+# Generate the speech with mimic and play it with alsa
+mimic3 --remote --voice "$voice" --length-scale "$speed" "$text" | aplay -q -

bitbots_misc/bitbots_tts/scripts/speak_ip.py

@@ -0,0 +1,19 @@
+#!/usr/bin/env python3
+
+import os
+from socket import gethostname
+
+from bitbots_tts.tts import say
+from netifaces import AF_INET, ifaddresses, interfaces
+
+ip_address = "not set"
+
+wifi_interface = next(filter(lambda i: i.startswith("wlp"), interfaces()), None)
+if wifi_interface and AF_INET in ifaddresses(wifi_interface):
+    ip_address = ifaddresses(wifi_interface)[AF_INET][0]["addr"]
+
+ip = " dot ".join(ip_address.split("."))
+
+robot_name = os.getenv("ROBOT_NAME") or gethostname()
+msg = f"Startup complete: {robot_name}. My wifi IP is {ip}."
+say(msg)

bitbots_misc/bitbots_tts/setup.py

@@ -16,7 +16,7 @@
     install_requires=[
         "setuptools",
     ],
-    scripts=["scripts/send_text.py"],
+    scripts=glob.glob("scripts/*"),
     entry_points={
         "console_scripts": [
             "tts = bitbots_tts.tts:main",