This platform creates a composite device tracker from one or more other device trackers and/or binary sensors. It will update whenever one of the watched entities updates, taking the last_seen/last_updated (and possibly GPS and battery) data from the changing entity. The result can be a more accurate and up-to-date device tracker if the "input" device tracker's update irregularly.
Currently device_tracker's with a source_type of bluetooth, bluetooth_le, gps or router are supported, as well as binary_sensor's.
Follow the installation instructions below. Then add the desired configuration. Here is an example of a typical configuration:
device_tracker:
- platform: composite
name: me
time_as: device_or_local
entity_id:
- device_tracker.platform1_me
- device_tracker.platform2_me
- binary_sensor.i_am_home
Place a copy of:
__init__.py
at <config>/custom_components/composite/__init__.py
const.py
at <config>/custom_components/composite/const.py
device_tracker.py
at <config>/custom_components/composite/device_tracker.py
manifest.json
at <config>/custom_components/composite/manifest.json
where <config>
is your Home Assistant configuration directory.
NOTE: Do not download the file by using the link above directly. Rather, click on it, then on the page that comes up use the
Raw
button.
You can use HACS to manage installation and updates by adding this repo as a custom repository and then searching for and installing the "Composite" integration.
To determine time zone from GPS coordinates (see time_as
configuration variable below) the package timezonefinderL (by default) is used. That package requires the package numpy. These will both be installed automatically by HA. Note, however, that numpy on Pi usually requires libatlas to be installed. (See this web page for more details.) It can be installed using this command:
sudo apt install libatlas3-base
Note: This is the same step that would be required if using a standard HA component that uses numpy (such as the Trend Binary Sensor), and is only required if you use
device_or_utc
ordevice_or_local
fortime_as
.
- tz_finder (Optional): Specifies which
timezonefinder
package, and possibly version, to install. Must be formatted as required bypip
. Default istimezonefinderL==4.0.2
. Other common values:
timezonefinderL==2.0.1
timezonefinder
timezonefinder<6
timezonefinder==4.2.0
- tz_finder_class (Optional): Specifies which class to use. Only applies when using
timezonefinder
package. Valid options areTimezoneFinder
andTimezoneFinderL
. The default isTimezoneFinder
.
Note: Starting with release 4.4.0 the
timezonefinder
package provides two classes to choose from: the originalTimezoneFinder
class, and a new class namedTimezoneFinderL
, which effectively replaces the functionality of thetimezonefinderL
package.
- entity_id: Entity IDs of watched device tracker devices. Can be a single entity ID, a list of entity IDs, or a string containing multiple entity IDs separated by commas. Another option is to specify a dictionary with
entity
specifying the entity ID, andall_states
specifying a boolean value that controls whether or not to use all states of the entity (rather than just the "Home" state, which is the default.) - name: Object ID (i.e., part of entity ID after the dot) of composite device. For example,
NAME
would result in an entity ID ofdevice_tracker.NAME
. - require_movement (Optional):
true
orfalse
. Default isfalse
. Iftrue
, will skip update from a GPS-based tracker if it has not moved. Specifically, if circle defined by new GPS coordinates and accuracy overlaps circle defined by previous GPS coordinates and accuracy then update will be ignored. - time_as (Optional): One of
utc
,local
,device_or_utc
ordevice_or_local
. Default isutc
which shows time attributes in UTC.local
shows time attributes per HA'stime_zone
configuration.device_or_utc
anddevice_or_local
attempt to determine the time zone in which the device is located based on its GPS coordinates. The name of the time zone (orunknown
) will be shown in a new attribute namedtime_zone
. If the time zone can be determined, then time attributes will be shown in that time zone. If the time zone cannot be determined, then time attributes will be shown in UTC ifdevice_or_utc
is selected, or in HA's local time zone ifdevice_or_local
is selected.
Watched GPS-based devices must have, at a minimum, the following attributes: latitude
, longitude
and gps_accuracy
. If they don't they will not be used.
For watched non-GPS-based devices, which states are used and whether any GPS data (if present) is used depends on several factors. E.g., if GPS-based devices are in use then the 'not_home'/'off' state of non-GPS-based devices will be ignored (unless all_states
was specified as true
for that entity.) If only non-GPS-based devices are in use, then the composite device will be 'home' if any of the watched devices are 'home'/'on', and will be 'not_home' only when all the watched devices are 'not_home'/'off'.
If a watched device has a last_seen
attribute, that will be used in the composite device. If not, then last_updated
from the entity's state will be used instead.
If a watched device has a battery
or battery_level
attribute, that will be used to update the composite device's battery
attribute. If it has a battery_charging
or charging
attribute, that will be used to udpate the composite device's battery_charging
attribute.
The watched devices, and the composite device, should all have track
set to true
.
It is recommended to not use the native merge feature of the device tracker component (i.e., do not add the MAC address from network-based trackers to a GPS-based tracker. See more details in the Device Tracker doc page.)
Attribute | Description |
---|---|
battery | Battery level (in percent, if available.) |
battery_charging | Battery charging status (True/False, if available.) |
entity_id | IDs of entities that have contributed to the state of the composite device. |
gps_accuracy | GPS accuracy radius (in meters, if available.) |
last_entity_id | ID of the last entity to update the composite device. |
last_seen | Date and time when current location information was last updated. |
latitude | Latitude of current location (if available.) |
longitude | Longitude of current location (if available.) |
source_type | Source of current location information: binary_sensor , bluetooth , bluetooth_le , gps or router . |
time_zone | The name of the time zone in which the device is located, or unknown if it cannot be determined. Only exists if device_or_utc or device_or_local is chosen for time_as . |
composite:
tz_finder: timezonefinder<6
tz_finder_class: TimezoneFinderL
device_tracker:
- platform: composite
name: me
time_as: device_or_local
require_movement: true
entity_id:
- device_tracker.platform1_me
- device_tracker.platform2_me
- device_tracker.router_my_device
- entity: binary_sensor.i_am_home
all_states: true
This example assumes time_as
is set to device_or_utc
or device_or_local
. It determines the difference between the time zone in which the device is located and the time_zone
in HA's configuration. A positive value means the device's time zone is ahead of (or later than, or east of) the local time zone.
sensor:
- platform: template
sensors:
my_tz_offset:
friendly_name: My time zone offset
unit_of_measurement: hr
value_template: >
{% set state = states.device_tracker.me %}
{% if state.attributes is defined and
state.attributes.time_zone is defined and
state.attributes.time_zone != 'unknown' %}
{% set n = now() %}
{{ (n.astimezone(state.attributes.last_seen.tzinfo).utcoffset() -
n.utcoffset()).total_seconds()/3600 }}
{% else %}
unknown
{% endif %}
This example converts a time attribute to the local time zone. It works no matter which time zone the attribute is in.
sensor:
- platform: template
sensors:
my_last_seen_local:
friendly_name: My last_seen time in local time zone
value_template: >
{{ state_attr('device_tracker.me', last_seen').astimezone(now().tzinfo) }}