A main.py

@@ -0,0 +1,83 @@
+import io
+import json
+import subprocess
+import wave
+from datetime import date
+from time import sleep
+
+import pyaudio
+import requests
+from playsound3 import playsound
+from pysystemd import ServiceManager, ServiceStatus
+
+SIAM_DIR = "/home/sika/siam"
+VOICE_PATH = SIAM_DIR + "/voices/ella.safetensors"
+SESSION_ID = "ses_siam_" + date.today().isoformat()
+AGENT = "siam-tts"
+
+services = ["ollama", "pocket-tts"]
+
+
+def ensure_running_services():
+    for service in services:
+        status = ServiceStatus(service)
+        if not status.is_running():
+            print(service + " is not running")
+            manager = ServiceManager(service)
+            manager.start()
+        else:
+            print(service + " already running")
+
+
+def sync_waybar():
+    sleep(0.1)
+    subprocess.run(["pkill", "-RTMIN+8", "waybar"], check=True)
+
+
+def get_hex_status():
+    return json.loads(
+        subprocess.run(
+            ["hex", "status", "--json"], check=True, capture_output=True
+        ).stdout
+    )["capture"]["state"]
+
+
+def tts(text):
+    response = requests.post("http://localhost:8000/tts", files={"text": (None, text)})
+    response.raise_for_status()
+
+    with wave.open(io.BytesIO(response.content), "rb") as wf:
+        p = pyaudio.PyAudio()
+        stream = p.open(
+            format=p.get_format_from_width(wf.getsampwidth()),
+            channels=wf.getnchannels(),
+            rate=wf.getframerate(),
+            output=True,
+        )
+        stream.write(wf.readframes(wf.getnframes()))
+        stream.stop_stream()
+        stream.close()
+        p.terminate()
+
+
+if get_hex_status() == "idle":
+    ensure_running_services()
+    playsound(SIAM_DIR + "/sounds/trigger/oui.wav")
+    subprocess.run(["hex", "dictation", "start"], check=True)
+    sync_waybar()
+else:
+    stt_output = subprocess.run(
+        ["hex", "dictation", "stop", "--print"], check=True, capture_output=True
+    ).stdout.decode()
+    sync_waybar()
+    print("> " + stt_output)
+
+    oc_output = subprocess.run(
+        "opencode run -s " + SESSION_ID + " --agent " + AGENT + ' "' + stt_output + '"',
+        shell=True,
+        check=True,
+        capture_output=True,
+    ).stdout.decode()
+
+    print("< " + oc_output)
+    tts(oc_output)

A requirements.txt

@@ -0,0 +1,4 @@
+playsound3==3.3.2
+pysystemd==1.1
+requests==2.34.2
+pyaudio==0.2.14