dune-weaver/modules/connection/connection_manager.py

import threading
import time
import logging
import serial
import serial.tools.list_ports
import websocket
import asyncio
import os

from modules.core import pattern_manager
from modules.core.state import state
from modules.led.led_interface import LEDInterface
from modules.led.idle_timeout_manager import idle_timeout_manager

logger = logging.getLogger(__name__)

IGNORE_PORTS = ['/dev/cu.debug-console', '/dev/cu.Bluetooth-Incoming-Port']


async def _check_table_is_idle() -> bool:
    """Helper function to check if table is idle."""
    return not state.current_playing_file or state.pause_requested


def _start_idle_led_timeout():
    """Start idle LED timeout if enabled."""
    if not state.dw_led_idle_timeout_enabled or state.dw_led_idle_timeout_minutes <= 0:
        return

    logger.debug(f"Starting idle LED timeout: {state.dw_led_idle_timeout_minutes} minutes")
    idle_timeout_manager.start_idle_timeout(
        timeout_minutes=state.dw_led_idle_timeout_minutes,
        state=state,
        check_idle_callback=_check_table_is_idle
    )


###############################################################################
# Connection Abstraction
###############################################################################

class BaseConnection:
    """Abstract base class for a connection."""
    def send(self, data: str) -> None:
        raise NotImplementedError

    def flush(self) -> None:
        raise NotImplementedError

    def readline(self) -> str:
        raise NotImplementedError

    def in_waiting(self) -> int:
        raise NotImplementedError

    def is_connected(self) -> bool:
        raise NotImplementedError

    def close(self) -> None:
        raise NotImplementedError

###############################################################################
# Serial Connection Implementation
###############################################################################

class SerialConnection(BaseConnection):
    def __init__(self, port: str, baudrate: int = 115200, timeout: int = 2):
        self.port = port
        self.baudrate = baudrate
        self.timeout = timeout
        self.lock = threading.RLock()
        logger.info(f'Connecting to Serial port {port}')
        self.ser = serial.Serial(port, baudrate, timeout=timeout)
        state.port = port
        logger.info(f'Connected to Serial port {port}')

    def send(self, data: str) -> None:
        with self.lock:
            self.ser.write(data.encode())
            self.ser.flush()

    def flush(self) -> None:
        with self.lock:
            self.ser.flush()

    def readline(self) -> str:
        with self.lock:
            return self.ser.readline().decode().strip()

    def in_waiting(self) -> int:
        with self.lock:
            return self.ser.in_waiting

    def is_connected(self) -> bool:
        return self.ser is not None and self.ser.is_open

    def close(self) -> None:
        # Save current state synchronously first (critical for position persistence)
        try:
            state.save()
        except Exception as e:
            logger.error(f"Error saving state on close: {e}")

        # Schedule async position update if event loop exists, otherwise skip
        # This avoids creating nested event loops which causes RuntimeError
        try:
            loop = asyncio.get_running_loop()
            # We're in async context - schedule as task (fire-and-forget)
            asyncio.create_task(update_machine_position())
            logger.debug("Scheduled async machine position update")
        except RuntimeError:
            # No running event loop - we're in sync context
            # Position was already saved above, skip async update to avoid nested loop
            logger.debug("No event loop running, skipping async position update")

        with self.lock:
            if self.ser.is_open:
                self.ser.close()
        # Release the lock resources
        self.lock = None

###############################################################################
# WebSocket Connection Implementation
###############################################################################

class WebSocketConnection(BaseConnection):
    def __init__(self, url: str, timeout: int = 5):
        self.url = url
        self.timeout = timeout
        self.lock = threading.RLock()
        self.ws = None
        self.connect()

    def connect(self):
        logger.info(f'Connecting to Websocket {self.url}')
        self.ws = websocket.create_connection(self.url, timeout=self.timeout)
        state.port = self.url
        logger.info(f'Connected to Websocket {self.url}')
        
    def send(self, data: str) -> None:
        with self.lock:
            self.ws.send(data)

    def flush(self) -> None:
        # WebSocket sends immediately; nothing to flush.
        pass

    def readline(self) -> str:
        with self.lock:
            data = self.ws.recv()
            # Decode bytes to string if necessary
            if isinstance(data, bytes):
                data = data.decode('utf-8')
            return data.strip()

    def in_waiting(self) -> int:
        return 0  # Not applicable for WebSocket

    def is_connected(self) -> bool:
        return self.ws is not None

    def close(self) -> None:
        # Save current state synchronously first (critical for position persistence)
        try:
            state.save()
        except Exception as e:
            logger.error(f"Error saving state on close: {e}")

        # Schedule async position update if event loop exists, otherwise skip
        # This avoids creating nested event loops which causes RuntimeError
        try:
            loop = asyncio.get_running_loop()
            # We're in async context - schedule as task (fire-and-forget)
            asyncio.create_task(update_machine_position())
            logger.debug("Scheduled async machine position update")
        except RuntimeError:
            # No running event loop - we're in sync context
            # Position was already saved above, skip async update to avoid nested loop
            logger.debug("No event loop running, skipping async position update")

        with self.lock:
            if self.ws:
                self.ws.close()
        # Release the lock resources
        self.lock = None
                
def list_serial_ports():
    """Return a list of available serial ports."""
    ports = serial.tools.list_ports.comports()
    available_ports = [port.device for port in ports if port.device not in IGNORE_PORTS]
    logger.debug(f"Available serial ports: {available_ports}")
    return available_ports

def device_init(homing=True):
    try:
        if get_machine_steps():
            logger.info(f"x_steps_per_mm: {state.x_steps_per_mm}, y_steps_per_mm: {state.y_steps_per_mm}, gear_ratio: {state.gear_ratio}")
        else: 
            logger.fatal("Failed to get machine steps")
            state.conn.close()
            return False
    except:
        logger.fatal("Not GRBL firmware")
        state.conn.close()
        return False

    machine_x, machine_y = get_machine_position()
    if machine_x != state.machine_x or machine_y != state.machine_y:
        logger.info(f'x, y; {machine_x}, {machine_y}')
        logger.info(f'State x, y; {state.machine_x}, {state.machine_y}')
        if homing:
            success = home()
            if not success:
                logger.error("Homing failed during device initialization")
    else:
        logger.info('Machine position known, skipping home')
        logger.info(f'Theta: {state.current_theta}, rho: {state.current_rho}')
        logger.info(f'x, y; {machine_x}, {machine_y}')
        logger.info(f'State x, y; {state.machine_x}, {state.machine_y}')

    time.sleep(2)  # Allow time for the connection to establish


def connect_device(homing=True):
    # Initialize LED interface based on configured provider
    if state.led_provider == "wled" and state.wled_ip:
        state.led_controller = LEDInterface(provider="wled", ip_address=state.wled_ip)
    elif state.led_provider == "hyperion" and state.hyperion_ip:
        state.led_controller = LEDInterface(
            provider="hyperion",
            ip_address=state.hyperion_ip,
            port=state.hyperion_port
        )
    else:
        state.led_controller = None

    # Show loading effect
    if state.led_controller:
        state.led_controller.effect_loading()

    ports = list_serial_ports()

    # Priority for auto-connect:
    # 1. Preferred port (user's explicit choice) if available
    # 2. Last used port if available
    # 3. First available port as fallback
    if state.preferred_port and state.preferred_port in ports:
        logger.info(f"Connecting to preferred port: {state.preferred_port}")
        state.conn = SerialConnection(state.preferred_port)
    elif state.port and state.port in ports:
        logger.info(f"Connecting to last used port: {state.port}")
        state.conn = SerialConnection(state.port)
    elif ports:
        logger.info(f"Connecting to first available port: {ports[0]}")
        state.conn = SerialConnection(ports[0])
    else:
        logger.error("Auto connect failed: No serial ports available")
        # state.conn = WebSocketConnection('ws://fluidnc.local:81')

    if (state.conn.is_connected() if state.conn else False):
        # Check for alarm state and unlock if needed before initializing
        if not check_and_unlock_alarm():
            logger.error("Failed to unlock device from alarm state")
            # Still proceed with device_init but log the issue

        device_init(homing)

    # Show connected effect, then transition to configured idle effect
    if state.led_controller:
        logger.info("Showing LED connected effect (green flash)")
        state.led_controller.effect_connected()
        # Set the configured idle effect after connection
        logger.info(f"Setting LED to idle effect: {state.dw_led_idle_effect}")
        state.led_controller.effect_idle(state.dw_led_idle_effect)
        _start_idle_led_timeout()

def check_and_unlock_alarm():
    """
    Check if GRBL is in alarm state and unlock it with $X if needed.
    Returns True if device is ready (unlocked or no alarm), False on error.

    Note: If sensors are physically triggered (Pn:XY), the alarm may persist
    but we still return True to allow homing to proceed.
    """
    try:
        logger.info("Checking device status for alarm state...")

        # Clear any pending data in buffer first
        while state.conn.in_waiting() > 0:
            state.conn.readline()

        # Send status query
        state.conn.send('?\n')
        time.sleep(0.2)

        # Read response with timeout
        max_attempts = 10
        response = None

        for attempt in range(max_attempts):
            if state.conn.in_waiting() > 0:
                response = state.conn.readline()
                logger.debug(f"Status response: {response}")
                if response and ('<' in response or 'Alarm' in response or 'Idle' in response):
                    break  # Got a valid status response
            time.sleep(0.1)

        if not response:
            logger.warning("No status response received, proceeding anyway")
            return True

        # Check for alarm state
        if "Alarm" in response:
            logger.warning(f"Device in ALARM state: {response}")

            # Send unlock command
            logger.info("Sending $X to unlock...")
            state.conn.send('$X\n')
            time.sleep(1.0)  # Give more time for unlock to process

            # Clear buffer before verification
            while state.conn.in_waiting() > 0:
                discarded = state.conn.readline()
                logger.debug(f"Discarded response: {discarded}")

            # Verify unlock succeeded
            state.conn.send('?\n')
            time.sleep(0.3)

            verify_response = None
            for attempt in range(max_attempts):
                if state.conn.in_waiting() > 0:
                    verify_response = state.conn.readline()
                    logger.debug(f"Verification response: {verify_response}")
                    if verify_response and '<' in verify_response:
                        break
                time.sleep(0.1)

            if verify_response and "Alarm" in verify_response:
                # Check if pins are physically triggered (Pn: in response)
                if "Pn:" in verify_response:
                    logger.warning(f"Alarm persists due to triggered sensors: {verify_response}")
                    logger.warning("Proceeding anyway - homing may clear the sensor state")
                    return True  # Let homing attempt to proceed
                else:
                    logger.error("Failed to unlock device from alarm state")
                    return False
            else:
                logger.info("Device successfully unlocked")
                return True
        else:
            logger.info("Device not in alarm state, proceeding normally")
            return True

    except Exception as e:
        logger.error(f"Error checking/unlocking alarm: {e}")
        return False

def get_status_response() -> str:
    """
    Send a status query ('?') and return the response if available.
    """
    while True:
        try:
            state.conn.send('?')
            response = state.conn.readline()
            if "MPos" in response:
                logger.debug(f"Status response: {response}")
                return response
        except Exception as e:
            logger.error(f"Error getting status response: {e}")
            return False
        time.sleep(1)
        
def parse_machine_position(response: str):
    """
    Parse the work position (MPos) from a status response.
    Expected format: "<...|MPos:-994.869,-321.861,0.000|...>"
    Returns a tuple (work_x, work_y) if found, else None.
    """
    if "MPos:" not in response:
        return None
    try:
        wpos_section = next((part for part in response.split("|") if part.startswith("MPos:")), None)
        if wpos_section:
            wpos_str = wpos_section.split(":", 1)[1]
            wpos_values = wpos_str.split(",")
            work_x = float(wpos_values[0])
            work_y = float(wpos_values[1])
            return work_x, work_y
    except Exception as e:
        logger.error(f"Error parsing work position: {e}")
    return None


async def send_grbl_coordinates(x, y, speed=600, timeout=2, home=False):
    """
    Send a G-code command to FluidNC and wait for an 'ok' response.
    If no response after set timeout, sets state to stop and disconnects.
    """
    logger.debug(f"Sending G-code: X{x} Y{y} at F{speed}")

    # Track overall attempt time
    overall_start_time = time.time()

    while True:
        try:
            gcode = f"$J=G91 G21 Y{y} F{speed}" if home else f"G1 X{x} Y{y} F{speed}"
            # Use asyncio.to_thread for both send and receive operations to avoid blocking
            await asyncio.to_thread(state.conn.send, gcode + "\n")
            logger.debug(f"Sent command: {gcode}")
            start_time = time.time()
            while True:
                # Use asyncio.to_thread for blocking I/O operations
                response = await asyncio.to_thread(state.conn.readline)
                logger.debug(f"Response: {response}")
                if response.lower() == "ok":
                    logger.debug("Command execution confirmed.")
                    return
        except Exception as e:
            # Store the error string inside the exception block
            error_str = str(e)
            logger.warning(f"Error sending command: {error_str}")

            # Immediately return for device not configured errors
            if "Device not configured" in error_str or "Errno 6" in error_str:
                logger.error(f"Device configuration error detected: {error_str}")
                state.stop_requested = True
                state.conn = None
                state.is_connected = False
                logger.info("Connection marked as disconnected due to device error")
                return False


        logger.warning(f"No 'ok' received for X{x} Y{y}, speed {speed}. Retrying...")
        await asyncio.sleep(0.1)
    
    # If we reach here, the timeout has occurred
    logger.error(f"Failed to receive 'ok' response after {max_total_attempt_time} seconds. Stopping and disconnecting.")
    
    # Set state to stop
    state.stop_requested = True
    
    # Set connection status to disconnected
    if state.conn:
        try:
            state.conn.disconnect()
        except:
            pass
        state.conn = None
        
    # Update the state connection status
    state.is_connected = False
    logger.info("Connection marked as disconnected due to timeout")
    return False

def get_machine_steps(timeout=10):
    """
    Get machine steps/mm from the GRBL controller.
    Returns True if successful, False otherwise.
    """
    if not state.conn or not state.conn.is_connected():
        logger.error("Cannot get machine steps: No connection available")
        return False

    x_steps_per_mm = None
    y_steps_per_mm = None
    start_time = time.time()

    # Clear any pending data in the buffer
    try:
        while state.conn.in_waiting() > 0:
            state.conn.readline()
    except Exception as e:
        logger.warning(f"Error clearing buffer: {e}")

    # Send the command to request all settings
    try:
        logger.info("Requesting GRBL settings with $$ command")
        state.conn.send("$$\n")
        time.sleep(1.0)  # Give GRBL time to process and respond (ESP32/FluidNC may need longer)
    except Exception as e:
        logger.error(f"Error sending $$ command: {e}")
        return False

    # Wait for and process responses
    settings_complete = False
    last_retry_time = start_time  # Track when we last sent $$ for retry logic
    while time.time() - start_time < timeout and not settings_complete:
        try:
            # Attempt to read a line from the connection
            if state.conn.in_waiting() > 0:
                response = state.conn.readline()
                logger.debug(f"Raw response: {response}")

                # Process the line
                if response.strip():  # Only process non-empty lines
                    for line in response.splitlines():
                        line = line.strip()
                        logger.debug(f"Config response: {line}")
                        if line.startswith("$100="):
                            x_steps_per_mm = float(line.split("=")[1])
                            state.x_steps_per_mm = x_steps_per_mm
                            logger.info(f"X steps per mm: {x_steps_per_mm}")
                        elif line.startswith("$101="):
                            y_steps_per_mm = float(line.split("=")[1])
                            state.y_steps_per_mm = y_steps_per_mm
                            logger.info(f"Y steps per mm: {y_steps_per_mm}")
                        elif line.startswith("$22="):
                            # $22 reports if the homing cycle is enabled
                            # returns 0 if disabled, 1 if enabled
                            # Note: We only log this, we don't overwrite state.homing
                            # because user preference (saved in state.json) should take precedence
                            firmware_homing = int(line.split('=')[1])
                            logger.info(f"Firmware homing setting ($22): {firmware_homing}, using user preference: {state.homing}")

                # Check if we've received all the settings we need
                if x_steps_per_mm is not None and y_steps_per_mm is not None:
                    settings_complete = True
            else:
                # No data waiting, small sleep to prevent CPU thrashing
                time.sleep(0.1)

                # Retry every 3 seconds if no response received
                if time.time() - last_retry_time > 3:
                    logger.warning("No response yet, sending $$ command again")
                    state.conn.send("$$\n")
                    last_retry_time = time.time()

        except Exception as e:
            logger.error(f"Error getting machine steps: {e}")
            time.sleep(0.5)
    
    # Process results and determine table type
    if settings_complete:
        if y_steps_per_mm == 180 and x_steps_per_mm == 256:
            state.table_type = 'dune_weaver_mini'
        elif y_steps_per_mm == 210 and x_steps_per_mm == 256:
            state.table_type = 'dune_weaver_mini_pro_byj'
        elif y_steps_per_mm == 270 and x_steps_per_mm == 200:
            state.table_type = 'dune_weaver_gold'
        elif y_steps_per_mm == 287:
            state.table_type = 'dune_weaver'
        elif y_steps_per_mm == 164:
            state.table_type = 'dune_weaver_mini_pro'
        elif y_steps_per_mm >= 320:
            state.table_type = 'dune_weaver_pro'
        else:
            state.table_type = None
            logger.warning(f"Unknown table type with Y steps/mm: {y_steps_per_mm}")

        # Use override if set, otherwise use detected table type
        effective_table_type = state.table_type_override or state.table_type

        # Set gear ratio based on effective table type (hardcoded)
        if effective_table_type in ['dune_weaver_mini', 'dune_weaver_mini_pro', 'dune_weaver_mini_pro_byj', 'dune_weaver_gold']:
            state.gear_ratio = 6.25
        else:
            state.gear_ratio = 10

        # Check for environment variable override
        gear_ratio_override = os.getenv('GEAR_RATIO')
        if gear_ratio_override is not None:
            try:
                state.gear_ratio = float(gear_ratio_override)
                logger.info(f"Machine type detected: {state.table_type}, effective: {effective_table_type}, gear ratio: {state.gear_ratio} (from GEAR_RATIO env var)")
            except ValueError:
                logger.error(f"Invalid GEAR_RATIO env var value: {gear_ratio_override}, using default: {state.gear_ratio}")
                logger.info(f"Machine type detected: {state.table_type}, effective: {effective_table_type}, gear ratio: {state.gear_ratio} (hardcoded)")
        elif state.table_type_override:
            logger.info(f"Machine type detected: {state.table_type}, overridden to: {effective_table_type}, gear ratio: {state.gear_ratio}")
        else:
            logger.info(f"Machine type detected: {state.table_type}, gear ratio: {state.gear_ratio} (hardcoded)")

        return True
    else:
        missing = []
        if x_steps_per_mm is None: missing.append("X steps/mm")
        if y_steps_per_mm is None: missing.append("Y steps/mm")
        logger.error(f"Failed to get all machine parameters after {timeout}s. Missing: {', '.join(missing)}")
        return False

def home(timeout=90):
    """
    Perform homing sequence based on configured mode:

    Mode 0 (Crash):
        - Y axis moves -22mm (or -30mm for mini) until physical stop
        - Set theta=0, rho=0 (no x0 y0 command)

    Mode 1 (Sensor):
        - Send $H command to home both X and Y axes
        - Wait for [MSG:Homed:X] and [MSG:Homed:Y] messages
        - Send x0 y0 to zero positions
        - Set theta to compass offset, rho=0

    Args:
        timeout: Maximum time in seconds to wait for homing to complete (default: 90)
    """
    import threading
    import math

    # Check for alarm state before homing and unlock if needed
    if not check_and_unlock_alarm():
        logger.error("Failed to unlock device from alarm state, cannot proceed with homing")
        return False

    # Flag to track if homing completed
    homing_complete = threading.Event()
    homing_success = False

    def home_internal():
        nonlocal homing_success
        effective_table_type = state.table_type_override or state.table_type
        homing_speed = 400
        if effective_table_type == 'dune_weaver_mini':
            homing_speed = 100
        try:
            if state.homing == 1:
                # Mode 1: Sensor-based homing using $H
                logger.info("Using sensor-based homing mode ($H)")

                # Clear any pending responses
                state.homed_x = False
                state.homed_y = False

                # Send $H command
                state.conn.send("$H\n")
                logger.info("Sent $H command, waiting for homing messages...")

                # Wait for [MSG:Homed:X] and [MSG:Homed:Y] messages
                max_wait_time = 30  # 30 seconds timeout for homing messages
                start_time = time.time()

                while (time.time() - start_time) < max_wait_time:
                    try:
                        response = state.conn.readline()
                        if response:
                            logger.debug(f"Homing response: {response}")

                            # Check for homing messages
                            if "[MSG:Homed:X]" in response:
                                state.homed_x = True
                                logger.info("Received [MSG:Homed:X]")
                            if "[MSG:Homed:Y]" in response:
                                state.homed_y = True
                                logger.info("Received [MSG:Homed:Y]")

                            # Break if we've received both messages
                            if state.homed_x and state.homed_y:
                                logger.info("Received both homing confirmation messages")
                                break
                    except Exception as e:
                        logger.error(f"Error reading homing response: {e}")

                    time.sleep(0.1)

                if not (state.homed_x and state.homed_y):
                    logger.warning(f"Did not receive all homing messages (X:{state.homed_x}, Y:{state.homed_y}), unlocking and continuing...")
                    # Unlock machine to clear any alarm state
                    state.conn.send("$X\n")
                    time.sleep(0.5)

                # Wait for idle state after $H
                logger.info("Waiting for device to reach idle state after $H...")
                idle_reached = check_idle()

                if not idle_reached:
                    logger.error("Device did not reach idle state after $H command")
                    homing_complete.set()
                    return

                # Send x0 y0 to zero both positions using send_grbl_coordinates
                logger.info(f"Zeroing positions with x0 y0 f{homing_speed}")

                # Run async function in new event loop
                loop = asyncio.new_event_loop()
                asyncio.set_event_loop(loop)
                try:
                    # Send G1 X0 Y0 F{homing_speed}
                    result = loop.run_until_complete(send_grbl_coordinates(0, 0, homing_speed))
                    if result == False:
                        logger.error("Position zeroing failed - send_grbl_coordinates returned False")
                        homing_complete.set()
                        return
                    logger.info("Position zeroing completed successfully")
                finally:
                    loop.close()

                # Wait for device to reach idle state after zeroing movement
                logger.info("Waiting for device to reach idle state after zeroing...")
                idle_reached = check_idle()

                if not idle_reached:
                    logger.error("Device did not reach idle state after zeroing")
                    homing_complete.set()
                    return

                # Set current position based on compass reference point (sensor mode only)
                # Only set AFTER x0 y0 is confirmed and device is idle
                offset_radians = math.radians(state.angular_homing_offset_degrees)
                state.current_theta = offset_radians
                state.current_rho = 0

                logger.info(f"Sensor homing completed - theta set to {state.angular_homing_offset_degrees}° ({offset_radians:.3f} rad), rho=0")

            else:
                logger.info(f"Using crash homing mode at {homing_speed} mm/min")

                # Run async function in new event loop
                loop = asyncio.new_event_loop()
                asyncio.set_event_loop(loop)
                try:
                    if effective_table_type == 'dune_weaver_mini':
                        result = loop.run_until_complete(send_grbl_coordinates(0, -30, homing_speed, home=True))
                        if result == False:
                            logger.error("Crash homing failed - send_grbl_coordinates returned False")
                            homing_complete.set()
                            return
                        state.machine_y -= 30
                    else:
                        result = loop.run_until_complete(send_grbl_coordinates(0, -22, homing_speed, home=True))
                        if result == False:
                            logger.error("Crash homing failed - send_grbl_coordinates returned False")
                            homing_complete.set()
                            return
                        state.machine_y -= 22
                finally:
                    loop.close()

                # Wait for device to reach idle state after crash homing
                logger.info("Waiting for device to reach idle state after crash homing...")
                idle_reached = check_idle()

                if not idle_reached:
                    logger.error("Device did not reach idle state after crash homing")
                    homing_complete.set()
                    return

                # Crash homing just sets theta and rho to 0 (no x0 y0 command)
                state.current_theta = 0
                state.current_rho = 0

                logger.info("Crash homing completed - theta=0, rho=0")

            # Update machine position from hardware after homing
            logger.info("Updating machine position after homing...")
            try:
                pos = get_machine_position()
                if pos and pos[0] is not None and pos[1] is not None:
                    state.machine_x, state.machine_y = pos
                    state.save()
                    logger.info(f"Machine position updated after homing: X={state.machine_x}, Y={state.machine_y}")
                else:
                    logger.warning("Could not get machine position after homing")
            except Exception as e:
                logger.error(f"Error updating machine position after homing: {e}")

            homing_success = True
            homing_complete.set()

        except Exception as e:
            logger.error(f"Error during homing: {e}")
            homing_complete.set()

    # Start homing in a separate thread
    homing_thread = threading.Thread(target=home_internal)
    homing_thread.daemon = True
    homing_thread.start()

    # Wait for homing to complete or timeout
    if not homing_complete.wait(timeout):
        logger.error(f"Homing timeout after {timeout} seconds")
        # Try to stop any ongoing movement
        try:
            if state.conn and state.conn.is_connected():
                state.conn.send("!\n")  # Send feed hold
                time.sleep(0.1)
                state.conn.send("\x18\n")  # Send reset
        except Exception as e:
            logger.error(f"Error stopping movement after timeout: {e}")
        return False

    if not homing_success:
        logger.error("Homing failed")
        return False

    logger.info("Homing completed successfully")
    return True

def check_idle():
    """
    Continuously check if the device is idle (synchronous version).
    """
    logger.info("Checking idle")
    while True:
        response = get_status_response()
        if response and "Idle" in response:
            logger.info("Device is idle")
            # Schedule async update_machine_position in the existing event loop
            try:
                # Try to schedule in existing event loop if available
                try:
                    loop = asyncio.get_running_loop()
                    # Create a task but don't await it (fire and forget)
                    asyncio.create_task(update_machine_position())
                    logger.debug("Scheduled machine position update task")
                except RuntimeError:
                    # No event loop running, skip machine position update
                    logger.debug("No event loop running, skipping machine position update")
            except Exception as e:
                logger.error(f"Error scheduling machine position update: {e}")
            return True
        time.sleep(1)

async def check_idle_async():
    """
    Continuously check if the device is idle (async version).
    """
    logger.info("Checking idle (async)")
    while True:
        response = await asyncio.to_thread(get_status_response)
        if response and "Idle" in response:
            logger.info("Device is idle")
            try:
                await update_machine_position()
            except Exception as e:
                logger.error(f"Error updating machine position: {e}")
            return True
        await asyncio.sleep(1)

def is_machine_idle() -> bool:
    """
    Single check to see if the machine is currently idle.
    Does not loop - returns immediately with current status.

    Returns:
        True if machine is idle, False otherwise
    """
    if not state.conn or not state.conn.is_connected():
        logger.debug("No connection - machine not idle")
        return False

    try:
        state.conn.send('?')
        response = state.conn.readline()

        if response and "Idle" in response:
            logger.debug("Machine status: Idle")
            return True
        else:
            logger.debug(f"Machine status: {response}")
            return False
    except Exception as e:
        logger.error(f"Error checking machine idle status: {e}")
        return False


def get_machine_position(timeout=5):
    """
    Query the device for its position.
    """
    start_time = time.time()
    while time.time() - start_time < timeout:
        try:
            state.conn.send('?')
            response = state.conn.readline()
            logger.debug(f"Raw status response: {response}")
            if "MPos" in response:
                pos = parse_machine_position(response)
                if pos:
                    machine_x, machine_y = pos
                    logger.debug(f"Machine position: X={machine_x}, Y={machine_y}")
                    return machine_x, machine_y
        except Exception as e:
            logger.error(f"Error getting machine position: {e}")
            return
        time.sleep(0.1)
    logger.warning("Timeout reached waiting for machine position")
    return None, None

async def update_machine_position():
    if (state.conn.is_connected() if state.conn else False):
        try:
            logger.info('Saving machine position')
            state.machine_x, state.machine_y = await asyncio.to_thread(get_machine_position)
            await asyncio.to_thread(state.save)
            logger.info(f'Machine position saved: {state.machine_x}, {state.machine_y}')
        except Exception as e:
            logger.error(f"Error updating machine position: {e}")

def restart_connection(homing=False):
    """
    Restart the connection. If a connection exists, close it and attempt to establish a new one.
    It will try to connect via serial first (if available), otherwise it will fall back to websocket.
    The new connection is saved to state.conn.
    
    Returns:
        True if the connection was restarted successfully, False otherwise.
    """
    try:
        if (state.conn.is_connected() if state.conn else False):
            logger.info("Closing current connection...")
            state.conn.close()
    except Exception as e:
        logger.error(f"Error while closing connection: {e}")

    # Clear the connection reference.
    state.conn = None

    logger.info("Attempting to restart connection...")
    try:
        connect_device(homing)  # This will set state.conn appropriately.
        if (state.conn.is_connected() if state.conn else False):
            logger.info("Connection restarted successfully.")
            return True
        else:
            logger.error("Failed to restart connection.")
            return False
    except Exception as e:
        logger.error(f"Error restarting connection: {e}")
        return False