7.4 KiB
layout, title, description, date, sidebar, comments, sharing, footer
layout | title | description | date | sidebar | comments | sharing | footer |
---|---|---|---|---|---|---|---|
page | Splitting up the configuration | Splitting the configuration.yaml into several files. | 2016-03-25 23:30 | false | false | true | true |
So you've been using Home Assistant for a while now and your configuration.yaml file brings people to tears or you simply want to start off with the distributed approach, here's how to "split the configuration.yaml" into more manageable (read: humanly readable) pieces.
First off, several community members have sanitized (read: without api keys/passwords etc) versions of their configurations available for viewing, you can see a list of them here.
As commenting code doesn't always happen, please read on for the details.
Now despite the logical assumption that the configuration.yaml
will be replaced by this process it will in fact remain all be it in a much less cluttered form.
In this lighter version we will still need what could be called the core snippet:
homeassistant:
# Name of the location where Home Assistant is running
name: My Hass Instance
# Location required to calculate the time the sun rises and sets
latitude: 37
longitude: -121
# C for Celsius, F for Fahrenheit
temperature_unit: F
# Pick yours from here: http://en.wikipedia.org/wiki/List_of_tz_database_time_zones
time_zone: America/Los_Angeles
customize: !include customize.yaml
Note that each line after homeassistant:
is indented two (2) spaces. Since the configuration files in Home Assistant are based on the YAML language, indentation and spacing are important. Also note that seemingly strange entry under customize:
.
!include filename.yaml
is the statement that tells Home Assistant to insert the contents of filename.yaml
at that point. This is how we are going to break a monolithic and hard to read file (when it gets big) into more manageable chunks.
Now before we start splitting out the different components, let's look at the other components (in our example) that will stay in the base file:
history:
frontend:
logbook:
http:
api_password: ImNotTelling!
ifttt:
key: [nope]
wink:
access_token: [wouldn't you]
refresh_token: [like to know]
zwave:
usb_path: /dev/ttyUSB0
config_path: /usr/local/share/python-openzwave/config
polling_interval: 10000
#zigbee:
# device: /dev/ttyUSB1
# baud: 115200
mqtt:
broker: 127.0.0.1
As with the core snippet, indentation makes a difference. The component headers (mqtt:
) should be fully left aligned (aka no indent), and the parameters (port:
) should be indented two (2) spaces.
While some of these components can technically be moved to a separate file they are so small or "one off's" where splitting them off is superfluous. Also, you'll notice the # symbol (hash/pound). This represents a "comment" as far as the commands are interpreted. Put another way, any line prefixed with a #
will be ignored. This makes breaking up files for human readability really convenient, not to mention turning off features while leaving the entry intact. (Look at the zigbee:
entry above and the b entry further down)
Now, lets assume that a blank file has been created in the hass configuration directory for each of the following:
automation.yaml
zones.yaml
sensors.yaml
switches.yaml
device_tracker.yaml
customize.yaml
automation.yaml
will hold all the automation component details. zones.yaml
will hold the zone component details and so forth. These files can be called anything but giving them names that match their function will make things easier to keep track of.
Inside the base configuration file add the following entries:
automation: !include automation.yaml
zone: !include zones.yaml
sensor: !include sensors.yaml
switch: !include switches.yaml
device_tracker: !include device_tracker.yaml
Note that there can only be one !include:
for each component so chaining them isn't going to work. If that sounds like greek, don't worry about it.
Alright, so we've got the single components and the include statements in the base file, what goes in those extra files?
Let's look at the device_tracker.yaml
file from our example:
- platform: owntracks
- platform: nmap_tracker
hosts: 192.168.2.0/24
home_interval: 3
track_new_devices: yes
interval_seconds: 40
consider_home: 120
This small example illustrates how the "split" files work. In this case, we start with a "comment block" identifying the file followed by two (2) device tracker entries (owntracks
and nmap
). These files follow "style 1" that is to say a fully left aligned leading entry (- platform: owntracks
) followed by the parameter entries indented two (2) spaces.
This (large) sensor configuration gives us another example:
### sensors.yaml
##############################################################
### METEOBRIDGE ####
##############################################################
- platform: tcp
name: 'Outdoor Temp (Meteobridge)'
host: 192.168.2.82
timeout: 6
payload: "Content-type: text/xml; charset=UTF-8\n\n"
value_template: "{{value.split (' ')[2]}}"
unit: C
- platform: tcp
name: 'Outdoor Humidity (Meteobridge)'
host: 192.168.2.82
port: 5556
timeout: 6
payload: "Content-type: text/xml; charset=UTF-8\n\n"
value_template: "{{value.split (' ')[3]}}"
unit: Percent
- platform: tcp
name: 'Outdoor Dewpoint (Meteobridge)'
host: 192.168.2.82
port: 5556
timeout: 6
payload: "Content-type: text/xml; charset=UTF-8\n\n"
value_template: "{{value.split (' ')[4] }}"
unit: C
###################################
#### STEAM FRIENDS ####
##################################
- platform: steam_online
api_key: [not telling]
accounts:
- 76561198012067051
##################################
#### TIME/DATE ####
##################################
- platform: time_date
display_options:
- 'time'
- 'date'
- platform: worldclock
time_zone: Etc/UTC
name: 'UTC'
- platform: worldclock
time_zone: America/New_York
name: 'Ann Arbor'
- platform: worldclock
time_zone: Europe/Vienna
name: 'Innsbruck'
- platform: worldclock
time_zone: America/New_York
name: 'Ann Arbor'
You'll notice that this example includes a secondary parameter section (under the steam section) as well as a better example of the way comments can be used to break down files into sections.
That about wraps it up.
If you have issues checkout home-assistant.log
in the configuration directory as well as your indentations. If all else fails, head over to the Gitter Chatroom and ask away.
{% linkable_title Advanced usage %}
We offer four advanced options to include whole directories at once.
!include_dir_list
will return content of a directory as a list with each file content being an entry in the list.
!include_dir_named
will return content of a directory as a dictionary which maps filename => content of file.
!include_dir_merge_list
will return content of a directory as a list by merging all files (which should contain a list) into 1 big list.
!include_dir_merge_named
will return content of a directory as a dictionary by loading each file and merging it into 1 big dictionary.