sylvie-2024

commit 51fd4d9ce2f44702304116c8a4c07e1dbf7d0c80
parent 1c971cce0dcfebe013456b3f2e28d57d437fd4d0
Author: khanumballz <[email protected]>
Date:   Thu, 16 May 2024 07:33:29 +0100

Piper tts and Ollama Bindings

Diffstat:
A
python/animation/voice-and-motor.py
 | 
72
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

1 file changed, 72 insertions(+), 0 deletions(-)
diff --git a/python/animation/voice-and-motor.py b/python/animation/voice-and-motor.py
@@ -0,0 +1,72 @@
+import serial
+import threading
+import time
+from dimits import Dimits
+import ollama
+
+def send_command(ser, command):
+    ser.write(command.encode())
+    ser.write(b'\n')  # Assuming commands are terminated with a newline character
+
+def play_tts(text, command_type):
+    # Initialize Dimits with the desired voice model
+    dt = Dimits("en_US-amy-medium")
+    # Convert text to audio and play it using the aplay engine
+    dt.text_2_speech(text, engine="aplay", command=command_type)
+
+def main():
+    # Open serial port
+    ser = serial.Serial('/dev/ttyUSB0', 115200, timeout=1)
+
+    if not ser.is_open:
+        print("Failed to open serial port.")
+        return
+
+    print("Serial port opened successfully.")
+
+    prompt_text = input("Enter the prompt for the robot: ")
+
+    stream = ollama.chat(
+        model='dolphin-phi2-usb',
+        messages=[{'role': 'user', 'content': prompt_text}],
+        stream=True,
+    )
+
+    generated_text = ""
+    for chunk in stream:
+      generated_text += chunk['message']['content']
+      #print(generated_text)
+      if len(generated_text.split()) >= 15:
+        print(generated_text)
+        user_text = generated_text
+        break
+
+    # Prompt the user to enter text for TTS
+    #user_text = input("Enter the text for speech: ")
+
+    # Determine the command type based on the length of the input
+    if len(user_text.split()) > 7:  # Check if the input has more than 5 words
+        command_type = "1sm"  # Use long command for speech
+    elif len(user_text.split()) > 4:
+        command_type = "1so"  # Use short command for speech
+    else:
+        command_type = "1sn"
+
+    # Create a thread to play the TTS audio with the user's text
+    tts_thread = threading.Thread(target=play_tts, args=(user_text, command_type))
+    tts_thread.start()
+
+    # Delay to ensure synchronization (adjust as needed)
+    time.sleep(3.5)  # Adjust the delay based on the duration of the TTS audio
+
+    # Initiate animation sequence after the delay
+    send_command(ser, "1s3")
+    send_command(ser, command_type)  # Use variable command_type
+    print("Sent motor command")
+    time.sleep(12)
+
+    # Wait for the TTS thread to finish before exiting
+    tts_thread.join()
+
+if __name__ == "__main__":
+    main()