bingo!
https://my_domain.duckdns.org/api/config?api_password=[password_encoded] works! in the browser, as does https://hassio.local/api/config?api_password=[password_encoded]
thanks!
havent found my way in yet trying this as a curl command…
1 Like
Hey @Jer78 this is pretty cool. I know you helped me a couple months back with some sonos groupings before that was really available. I was actually just going back through that thread trying to set it up again, Im curious if your still doing it that way or if there is a cleaner way now?
Id even like to add the ability of grouping rooms using voice. It looks like you need to specify a ‘master’ though when using the sonos_join service. I think that might get in the way, unless I can come up with a sensor and script to do that for me…hmmmm. I will work on this.
Thanks
Hi @Darbos yes I’ll try to update the Sonos grouping project with this sometime this weekend. I’ll post in the other thread and tag you if/when I get it working correctly.
I do have a script that joins and unjoins Sonos speakers by voice. I’ll post that for you as well. But basically once you have the sensor that tells you the master you can use that dynamically in the script.
My man. Thank you!
hello, thanks for sharing this solution. at every restart of Hassio I start an automation that populates my input_select, but unfortunately doing so I lose the current selection, do you know if there is a way to keep the previous setting on reboot?
can you please tell me what this HTTP password? I keep getting the following error.
Login attempt or request with invalid authentication from 127.0.0.1
update: solution is here
1 Like
Sorry - I only just found your post now, haven’t logged in for a while. Yeah since HA has has long lived tokens, I use them religiously now 
that’s the way to go
This string helped me a lot - thanks gents. I still can’t get this to work. Goal is to dynamically update input_select… Gives me a perfectly formatted array in the developer tools template.
rest_command:
populate_input_select_options:
url: "http://192.168.1.100:8123/api/services/input_select/set_options"
method: POST
headers:
x-ha-access: !secret http_password
content-type: application/json
payload: >-
{
"entity_id": "input_select.meat_type"
"options": [
{%- for state in states.variable if state_attr(state.entity_id, 'meat_type')==states['input_select.meat_type'].state -%}
"{{state_attr(state.entity_id, 'meat_cut')}}"
{%- if not loop.last %}, {%-endif-%}
{%- endfor -%}]
}
Help much appreciated. Dave
It’s giving you a string that only looks like an array.
The output of any template is a string.
Hi Dave,
A few suggestions:
1: Paste your rest_command config into the template area with the HASS UI, then paste the resulting payload here or at the jsonlint site to test if it’s actually valid json
2: it looks like you are using an HTTP password, try creating long life tokens instead. (The ‘bearer: ####’ examples from above)
3: call the rest command a few times then check the main log. It should tell you what is going wrong.
Hi everybody,
Im newbie with HA but i like it 
I would like too populate an input_select depending of the value of a first input_select.
input_select:
who_cooks:
name: Who cooks today
options:
- Select
- Paulus
- Anne Therese
initial: Select
icon: mdi:panda
living_room_preset:
options:
- Visitors
- Visitors with kids
- Home Alone
automation:
- alias: change select
trigger:
platform: state
entity_id: input_select.who_cooks
action:
service: input_select.set_options
data:
entity_id: input_select.living_room_preset
options: >
{% if trigger.to_state.state == 'Paulus' %}
["Item A", "Item B", "Item C"]
{% elif trigger.to_state.state == 'radio' %}
["19", "20", "21"]
{% endif %}
I understand that template returns string, also i would like use the workaroud.set_options to populate the second input_select.
BUT i need the values are in a list written by me for some temperature value.
If 1st input == ‘Comfort’ , the second input populates with temperature number 19, 20, 21…
Like
automation:
- alias: change input_select_Second
trigger:
platform: state
entity_id: input_select.first
action:
service: input_select.set_options
data_template:
entity_id: input_select.second
options: >
{% if trigger.to_state.state == 'Comfort' %}
["20", "20.5", "21", "21.5", "22"]
{% elif trigger.to_state.state == 'Eco' %}
["14", "14.5", "15", "15.5"]
{% endif %}
Can you help me please ?
Thanks
With 0.117 and the new native types for templates, you should be able to use the automation as you have written.
Just make sure you have
homeassistant:
legacy_templates: false
in your configuration.yaml
Thanks @Olen
I know but it is in beta and i dont want to risk modify the other template that i use.
It’s why i would like to do this with the rest_command
Generally beta doesnt like me
The chances of anything breaking are pretty slim, especially if you have just started using ha, and your templates are reasonably standard. And if they do, it is pretty easy to fix.
I have hundreds of automations, scripts and sensors using templates built over several years, and nothing serious broke here.
Not using something that works well and will be default in the next release or so is imho just causing yourself more trouble at a later date.
Hi @Olen
He i cant wait and i do the modification.
It works like a charm yeah !!
But i want testing the rest solution for populating.
What add in configuration for testing the api? Please
Thanks for thé job for all admins of ha
I know this thread is old, and my reply doesn’t exactly address the original question, but I have solved a similar problem that someone might find useful.
The use case is: maintain a dropdown in the front end containing a (relatively short) list of movies available for my Kodi media player. The list must be able to be updated dynamically. My solution uses pyscript modules to glue everything together (https://github.com/custom-components/pyscript).
The dropdown is defined as an input_select in yaml, with a static dummy option:
kodi_movies:
name: Kodi Movies
options:
- "Select"
Next is a script to set or update the actual list - the key service is kodi.call_method:
kodi_get_all_movies:
alias: Kodi Get All Movies
sequence:
- condition: state
entity_id: media_player.shield_tv
state: 'off'
- service: input_boolean.turn_off
data:
entity_id: input_boolean.kodi_specific_movie
- service: media_player.turn_on
data:
entity_id: media_player.shield_tv
- delay: 00:00:03
- service: androidtv.adb_command
data:
entity_id: media_player.shield_tv
command: "am start -a android.intent.action.MAIN -c android.intent.category.LEANBACK_LAUNCHER -n 'org.xbmc.kodi/.Splash'"
- delay: 00:00:02
- service: kodi.call_method
data:
entity_id: media_player.kodi
method: VideoLibrary.GetMovies
- delay: 00:00:02
- service: remote.send_command
entity_id: remote.harmony_hub
data:
command: 'Home'
device: NVIDIA Shield TV
- delay: 00:00:02
- service: media_player.turn_off
data:
entity_id: media_player.shield_tv
This drives an automation to process the response, which is a list of 2 dictionaries (“limit” describes the number of entries, “movies” contains all the movies and movie IDs). I am just using the movie title at the moment, then calling Kodi again to get the ID when I play it. I might try to save the IDs at this point, but I think that would get a bit messy:
- alias: Kodi All Movies Result
id: kodi_all_movies_result
trigger:
- platform: event
event_type: kodi_call_method_result
event_data:
result_ok: true
input:
method: VideoLibrary.GetMovies
condition:
- condition: state
entity_id: input_boolean.kodi_specific_movie
state: 'off'
action:
- service: pyscript.populate_all_movies
data:
entity: input_select.kodi_movies
result: "{{ trigger.event.data.result }}"
Finally, here is the pyscript module called by the above automation:
#!/usr/bin/python3
# Populate the given input select from the result of a Kodi all movies query.
import sys
# File I/O is done in a called module - ensure the path includes it. Note that this only needs to be added
# to one file ... pyscript picks it up when loading the files.
if "/config/pyscript_modules" not in sys.path:
sys.path.append("/config/pyscript_modules")
import file_ops
# This runs as a service, and is normally run from an automation.
@service
def populate_all_movies(entity=None, result=None):
log.debug(f"got entity {entity}, result {result}")
# Inputs are required. They are defined as optional so we can just log an error instead of encountering
# an exception.
if entity is None or result is None:
log.error("entity and result are both required, exiting")
# Domain must be input_select.
elif not entity.startswith("input_select."):
log.error("input entity must be domain input_select, exiting")
# All seems well.
else:
# List of all movies
movies = result['movies']
all_movies = []
for movie in movies:
all_movies.append(movie['label'])
all_movies.sort()
log.debug(f"movie list {all_movies}")
# Populate the input select.
input_select.set_options(entity_id=entity, options=all_movies)
# Write the list to a file so we can restore the input select after a restart.
task.executor(file_ops.write_movie_file, all_movies)
Well, I said “finally” but not really. Note the last instruction above. When HA restarts, the input_select reverts back to its initial, static, value. So the write_movie_file() above persists the movie list using a native Python module. Note that when pyscript calls native Python, using task.executor(), the native Python module must exist in a different directory, python_modules in my case. At the top of the above file is a bit of code to modify the Python search path to include this directory. This is covered in the pyscript doc, but I struggled with it a bit. Here is the file containing write_movie_file(), as well as read_movie_file(), to be described in a bit:
#!/usr/bin/python3
# This native Python module is called by pyscript modules to perform file operations.
import os
import pickle
# Constants
MOVIE_BACKUP_FILE = "/config/all_movies_backup"
# Write the movie list to a file.
def write_movie_file(all_movies):
with open(MOVIE_BACKUP_FILE, "wb") as outfile:
pickle.dump(all_movies, outfile)
# Read and return the movie list.
def read_movie_file():
with open(MOVIE_BACKUP_FILE, "rb") as infile:
return pickle.load(infile)
To fully restore the input_select we need not only the movie list, but the selected option as well. For my use this is the movie to be shown for my “feature presentation” mode (work in progress), so I don’t want to lose it over a restart. This requires an input_text (not shown since it just has a name), and an automation to set the text when a movie is selected from the dropdown in the front end:
- alias: Save Feature Movie Name
id: save_feature_movie_name
trigger:
- platform: state
entity_id: input_select.kodi_movies
condition:
- condition: template
value_template: "{{ ( trigger.to_state.state ) != ( 'Select' ) }}"
action:
- service: input_text.set_value
data_template:
entity_id: input_text.current_feature
value: "{{ ( trigger.to_state.state ) }}"
Here is the restart automation to restore the input_select:
- alias: Restore Movie List On Start
id: restore_movie_list_on_start
trigger:
- platform: homeassistant
event: start
action:
- service: pyscript.restore_all_movies
data:
entity: input_select.kodi_movies
- service: input_select.select_option
data:
entity_id: input_select.kodi_movies
option: "{{ states('input_text.current_feature') }}"
And the pyscript module:
#!/usr/bin/python3
# Restore the given input select from the previously saved list of all movies.
import file_ops
# This runs as a service, and is normally run from an automation.
@service
def restore_all_movies(entity=None):
log.debug(f"got entity {entity}")
# Input entity is required. It is defined as optional so we can just log an error instead of encountering
# an exception.
if entity is None:
log.error("entity is required, exiting")
# Domain must be input_select.
elif not entity.startswith("input_select."):
log.error("input entity must be domain input_select, exiting")
# All seems well.
else:
# Read the backup file.
try:
all_movies = task.executor(file_ops.read_movie_file)
except FileNotFoundError:
log.error(f"saved movie file not found, exiting")
sys.exit(f"file not found")
except pickle.UnpicklingError:
log.error(f"saved movie file not usable, exiting")
sys.exit(f"file not usable")
log.debug(f"movie list {all_movies}")
# Populate the input select.
input_select.set_options(entity_id=entity, options=all_movies)
Apologies for the massively long reply
Hey it’s always great when someone takes the time to post their solution. There is nearly always something to be learned from the ingenuity of others. Thanks!
I thought I might share the direction I have gone since this thread… so when I need something reflected into HASS based on a scripted payload, I push it straight to HASS from the code.
Here is some old BASH code I really need to convert to Python. The services function allows me to update most things editable. The sensors thing has been really valuable too.
hassApi='https://mypersonal.dns.com/api/'
# where I keep my long lived access tokens
# the directory business is so I can use these functions from within the HASS docker,
# or from my server
if [ -f /config/llat.txt ]; then
source /config/llat.txt
elif [ -f /volume1/docker/hass/llat.txt ]; then
source /volume1/docker/hass/llat.txt
fi
function callHASSservice {
domain=$1
service=$2
json=$3
_service='services/'
serviceurl=$hassApi$_service
url="$serviceurl$domain/$service"
post=$(curl -s -X POST -H "Authorization: Bearer $llat" -H "Content-Type: application/json" -d "$json" $url)
echo $post
}
function callHASSsensor {
# I use a state-on, pause then state-off so that I can more easily control when automations trigger.. probably uneccessary
stateon=$1
json=$2
sensor=$3
stateoff=$4
_sensor='states/sensor.'
status='{"state":"'$stateon'","attributes":'$json'}'
sensorurl=$hassApi$_sensor$sensor
send=$(curl -X POST -s -H "Authorization: Bearer $llat" -H "Content-Type: application/json" -d "$status" $sensorurl)
if [ ! -z $4 ]; then
sleep 3
status='{"state":"'$stateoff'","attributes":'$json'}'
send=$(curl -s -X POST -H "Authorization: Bearer $llat" -H "Content-Type: application/json" -d "$status" $sensorurl)
echo $send
fi
}
This works !!!
action:
- service: input_select.set_options
target:
entity_id: input_select.di_zhi
data:
options: >
{% set zonelist = namespace(numbers=[]) %} {%- for zone in states.zone -%} {%- set zonelist.numbers = zonelist.numbers + [zone.entity_id.replace('zone.','') ] -%} {%- endfor -%} {{zonelist.numbers}}
If you simply want to create a list of each zone’s object_id (the object_id in zone.office is office) then you can do that, without using a for-loop, like this:
action:
- service: input_select.set_options
target:
entity_id: input_select.di_zhi
data:
options: "{{ states.zone | map(attribute='object_id') | list }}"
3 Likes
I decided to try out the new native types support for templates (beta) feature .
I was able to populate the list of my favorite Sonos stations in an input_select helper with this code:
service: input_select.set_options
target:
entity_id: input_select.sonos_favorites
data_template:
options: >
[{%- for key, value in state_attr('sensor.sonos_favorites', 'items').items() -%}
"{{ value }}",
{%- endfor %}]
1 Like