Create speech.py

speech.py

@@ -0,0 +1,34 @@
+import asyncio, threading
+import speech_recognition as sr
+from TTS.api import TTS
+from logger import logger
+
+# Load TTS
+from config import cfg
+tts = TTS(cfg["tts_model"])
+
+# Async wrapper for TTS
+def tts_speak(text: str, filename: str):
+    try:
+        tts.tts_to_file(text=text, file_path=filename)
+        logger.info(f"TTS generated at {filename}")
+    except Exception as e:
+        logger.error("TTS error", exc_info=e)
+
+def speak_async(text: str, filename="response.wav"):
+    thread = threading.Thread(target=tts_speak, args=(text, filename), daemon=True)
+    thread.start()
+
+# Speech-to-text
+def listen() -> str:
+    recognizer = sr.Recognizer()
+    with sr.Microphone() as source:
+        logger.info("Listening for scenario...")
+        audio = recognizer.listen(source)
+    try:
+        text = recognizer.recognize_google(audio)
+        logger.info(f"Transcribed audio: {text}")
+        return text
+    except Exception as e:
+        logger.error("STT error", exc_info=e)
+        return ""