"""Support for displaying the minimal and the maximal value."""
import logging

import voluptuous as vol

import homeassistant.helpers.config_validation as cv
from homeassistant.components.sensor import PLATFORM_SCHEMA
from homeassistant.const import (
    CONF_NAME,
    STATE_UNKNOWN,
    STATE_UNAVAILABLE,
    CONF_TYPE,
    ATTR_UNIT_OF_MEASUREMENT,
)
from homeassistant.core import callback
from homeassistant.helpers.entity import Entity
from homeassistant.helpers.event import async_track_state_change

_LOGGER = logging.getLogger(__name__)

ATTR_MIN_VALUE = "min_value"
ATTR_MAX_VALUE = "max_value"
ATTR_COUNT_SENSORS = "count_sensors"
ATTR_MEAN = "mean"
ATTR_LAST = "last"

ATTR_TO_PROPERTY = [
    ATTR_COUNT_SENSORS,
    ATTR_MAX_VALUE,
    ATTR_MEAN,
    ATTR_MIN_VALUE,
    ATTR_LAST,
]

CONF_ENTITY_IDS = "entity_ids"
CONF_ROUND_DIGITS = "round_digits"

ICON = "mdi:calculator"

SENSOR_TYPES = {
    ATTR_MIN_VALUE: "min",
    ATTR_MAX_VALUE: "max",
    ATTR_MEAN: "mean",
    ATTR_LAST: "last",
}

PLATFORM_SCHEMA = PLATFORM_SCHEMA.extend(
    {
        vol.Optional(CONF_TYPE, default=SENSOR_TYPES[ATTR_MAX_VALUE]): vol.All(
            cv.string, vol.In(SENSOR_TYPES.values())
        ),
        vol.Optional(CONF_NAME): cv.string,
        vol.Required(CONF_ENTITY_IDS): cv.entity_ids,
        vol.Optional(CONF_ROUND_DIGITS, default=2): vol.Coerce(int),
    }
)


async def async_setup_platform(hass, config, async_add_entities, discovery_info=None):
    """Set up the min/max/mean sensor."""
    entity_ids = config.get(CONF_ENTITY_IDS)
    name = config.get(CONF_NAME)
    sensor_type = config.get(CONF_TYPE)
    round_digits = config.get(CONF_ROUND_DIGITS)

    async_add_entities(
        [MinMaxSensor(hass, entity_ids, name, sensor_type, round_digits)], True
    )
    return True


def calc_min(sensor_values):
    """Calculate min value, honoring unknown states."""
    val = None
    for sval in sensor_values:
        if sval != STATE_UNKNOWN:
            if val is None or val > sval:
                val = sval
    return val


def calc_max(sensor_values):
    """Calculate max value, honoring unknown states."""
    val = None
    for sval in sensor_values:
        if sval != STATE_UNKNOWN:
            if val is None or val < sval:
                val = sval
    return val


def calc_mean(sensor_values, round_digits):
    """Calculate mean value, honoring unknown states."""
    val = 0
    count = 0
    for sval in sensor_values:
        if sval != STATE_UNKNOWN:
            val += sval
            count += 1
    if count == 0:
        return None
    return round(val / count, round_digits)


class MinMaxSensor(Entity):
    """Representation of a min/max sensor."""

    def __init__(self, hass, entity_ids, name, sensor_type, round_digits):
        """Initialize the min/max sensor."""
        self._hass = hass
        self._entity_ids = entity_ids
        self._sensor_type = sensor_type
        self._round_digits = round_digits

        if name:
            self._name = name
        else:
            self._name = "{} sensor".format(
                next(v for k, v in SENSOR_TYPES.items() if self._sensor_type == v)
            ).capitalize()
        self._unit_of_measurement = None
        self._unit_of_measurement_mismatch = False
        self.min_value = self.max_value = self.mean = self.last = None
        self.count_sensors = len(self._entity_ids)
        self.states = {}

        @callback
        def async_min_max_sensor_state_listener(entity, old_state, new_state):
            """Handle the sensor state changes."""
            if new_state.state is None or new_state.state in [
                STATE_UNKNOWN,
                STATE_UNAVAILABLE,
            ]:
                self.states[entity] = STATE_UNKNOWN
                hass.async_add_job(self.async_update_ha_state, True)
                return

            if self._unit_of_measurement is None:
                self._unit_of_measurement = new_state.attributes.get(
                    ATTR_UNIT_OF_MEASUREMENT
                )

            if self._unit_of_measurement != new_state.attributes.get(
                ATTR_UNIT_OF_MEASUREMENT
            ):
                _LOGGER.warning(
                    "Units of measurement do not match for entity %s", self.entity_id
                )
                self._unit_of_measurement_mismatch = True

            try:
                self.states[entity] = float(new_state.state)
                self.last = float(new_state.state)
            except ValueError:
                _LOGGER.warning(
                    "Unable to store state. " "Only numerical states are supported"
                )

            hass.async_add_job(self.async_update_ha_state, True)

        async_track_state_change(hass, entity_ids, async_min_max_sensor_state_listener)

    @property
    def name(self):
        """Return the name of the sensor."""
        return self._name

    @property
    def state(self):
        """Return the state of the sensor."""
        if self._unit_of_measurement_mismatch:
            return None
        return getattr(
            self, next(k for k, v in SENSOR_TYPES.items() if self._sensor_type == v)
        )

    @property
    def unit_of_measurement(self):
        """Return the unit the value is expressed in."""
        if self._unit_of_measurement_mismatch:
            return "ERR"
        return self._unit_of_measurement

    @property
    def should_poll(self):
        """No polling needed."""
        return False

    @property
    def device_state_attributes(self):
        """Return the state attributes of the sensor."""
        state_attr = {
            attr: getattr(self, attr)
            for attr in ATTR_TO_PROPERTY
            if getattr(self, attr) is not None
        }
        return state_attr

    @property
    def icon(self):
        """Return the icon to use in the frontend, if any."""
        return ICON

    async def async_update(self):
        """Get the latest data and updates the states."""
        sensor_values = [self.states[k] for k in self._entity_ids if k in self.states]
        self.min_value = calc_min(sensor_values)
        self.max_value = calc_max(sensor_values)
        self.mean = calc_mean(sensor_values, self._round_digits)