Assume there is an entity with entity_picture defined as
/api/media_player_proxy/media_player.kodi_rpi_1?token=b684340411dbbb3e884c394b311ce6a57c5d09d430d6bb6335bce3b4e04ab2ad&cache=b6e3f2f3fb9fbc95
or simply as
/local/images/image.jpg
How to save this as an image file in some directory?
For the first one you could create a generic camera and use the snapshot service.
That is already an image file in some directory. if you want to copy or move it use a shell command.
Thanks for a feedback.
Initial goal was to create a Logbook for a template sensor which has a dynamically changing “entity_picture”.
Particularly - I need a list of media content watched in Kodi (movies, music, …).
I can get a current “entity_picture” from some “media_player” entity & then use it for that template sensor (updated on every new played content).
But my attempt to see pictures in Logbook failed.
So I started looking for “ways”.
I speculated that may be storing these images locally in some folder may help me somehow.
Problem is that this address “/api/media_player_proxy/media_player.kodi_rpi_1?token=xxxx” is different for every played media content, i.e. not static. So I have no idea how to save these images in some kind of automation.
templates are allowed in the generic camera, so you should be able to point it to the media player url with the changing token. then do the snapshot when the content of that changes
Thanks for advising a Generic camera, started learning…
Generic camera was moved to UI from yaml, omg.
What a great joy to define templates in UI…
First step: testing a locally stored image.
Assume an image is stored in “/local/images/test/pink.jpg”.
Config for camera - a challenge with such a “convenient” UI:
Address is
http://localhost:8123/local/images/test/pink.jpg
We cannot use a “/local” address here:
Renamed the created camera entity to “camera.test_static_pink”.
Created a directory “snapshots” in “www/images/”.
Added permissions:
sudo chmod -Rc 777 www/images/snapshots
Added to “allowlist_external_dirs”: (update 26.11.24 - not needed since I already had “/config” in “allowlist_external_dirs”)
homeassistant:
...
allowlist_external_dirs:
...
- /config/www/images/snapshots
Failed to define a “/local” path here:
Called a service:
service: camera.snapshot
target:
entity_id: camera.test_static_pink
data:
filename: /config/www/images/snapshots/2.jpg
Failed to use a “/local” path here:
And got my jpg file.
did you get it all to work? wasn’t clear since you posted a lot of errors but then in the end you said you got your jpg.
if it helps, this is what i have in one of my generic cameras:
http://homeassistant.local:8123{{ state_attr('camera.ring', 'entity_picture') }}
and that attribute points to
/api/camera_proxy/camera.ring?token=xxxxxxx
of course i verified in my browser that the above url works fine… presume you did as well for your localhost address.
and url is like
/api/media_player_proxy/media_player.kodi_rpi_1?token=65d56d55d52a2ca689a05abf9016933986e736c555218742bae20c553de02a43&cache=b6e3f2f3fb9fbc95
I am testing now with
http://localhost:8123{{state_attr('media_player.kodi_rpi_1','entity_picture')}}
it creates a camera entity fine.
Now I am making a test automation to save snapshots on every new media played…
Will publish here a ready result.
not sure if this helps you, but the snapshot can be made in local, and on the /media share if you want
- service: camera.snapshot
data:
filename: >
/media/snapshots/oprit_{{now().strftime('%Y_%m_%d_%H_%M')}}.jpg
target:
entity_id: camera.oprit
- service: camera.snapshot
data:
filename: /media/snapshots/camera_oprit.jpg
target:
entity_id: camera.oprit
- service: camera.snapshot
data:
filename: /config/www/images/snapshots/camera_oprit.jpg
target:
entity_id: camera.oprit
this creates a file for each individual motion detection, one fixed named file that is overwritten (so we can use it in a template image, and also a fixed one in the /local folder
camera:
- platform: local_file
name: Snapshot Oprit
file_path: /media/snapshots/camera_oprit.jpg
##########################################################################################
# Templates
##########################################################################################
template:
- trigger:
- platform: state
entity_id:
- binary_sensor.oprit_vehicle_detected
- binary_sensor.oprit_person_detected
to: 'on'
for:
seconds: 4
- platform: event
event_type: delayed_homeassistant_start
- platform: event
event_type: event_template_reloaded
image:
- unique_id: last_snapshot_oprit_camera
url: !secret snapshot_camera_oprit
the secret is an url in the form of:
snapshot_camera_oprit: http://ha.local:<port>/local/images/snapshots/camera_oprit.jpg
Thanks for the example.
As I already explained - I failed to use “local” here:
and then failed to add to “allowlist” here:
That is why I had to use a “/config” method.
could it be the path you set? I have only this:
homeassistant:
packages: !include_dir_named packages
allowlist_external_dirs:
- /config
- /media
No…
Seems that “local” cannot be used here.
if you test with the dev container, the camera will have an entity picture, but the snapshot will not be saved.
It seems to be properly saved.
Yes, I am using Docker.
Ofc I am making 1st steps here - so could easy miss smth.
Well, step two: testing with an “entity_picture” from Kodi
Create camera entity:
http://localhost:8123{{state_attr('media_player.kodi_rpi_1','entity_picture')}}
Rename it to “camera.test_kodi_rpi_1”.
Create a helper:
input_text:
kodi_last_image:
max: 255
which is used in automation to save the latest “entity_picture”.
Create an automation:
alias: "test: Kodi: store images"
id: test_kodi_store_images
mode: single
trigger:
- platform: state
entity_id: media_player.kodi_rpi_1
to:
- playing
- idle
- paused
action:
- choose:
- conditions:
- condition: template
value_template: >-
{{ trigger.to_state.state == 'idle' or
trigger.to_state.state == 'paused' }}
sequence:
- service: input_text.set_value
data:
value: >-
{% if trigger.from_state.attributes.entity_picture is defined -%}
{{ trigger.from_state.attributes.entity_picture }}
{%- else -%}
'empty'
{%- endif %}
target:
entity_id: input_text.kodi_last_image
- conditions:
- condition: template
value_template: >-
{{ trigger.to_state.state == 'playing' }}
sequence:
- if:
- condition: template
value_template: >-
{{ trigger.to_state.attributes.entity_picture != states('input_text.kodi_last_image') }}
then:
- service: camera.snapshot
target:
entity_id: camera.testing_kodi_rpi_1
data:
filename: >-
{% set FILENAME = (now()|string).split('.')[0].replace('-','').replace(' ','').replace(':','') -%}
/config/www/images/snapshots/{{FILENAME}}.jpg
Here:
1.When a player is paused or stopped - the latest image is saved in input_text.
2.When a player is started - compare a current image with that previously saved image: if not same → save as a file.
This is done to prevent consecutive storing same image in cases when:
– media is paused
– a music album is played
Result files have names like “YYYYMMDDHHMMSS”.
Plus - add a periodical purging old files in another automation, smth like
service: delete.files_in_folder
data:
folder: "/config/www/images/snapshots"
time: >-
{{ VALUE_PURGE_KEEP_DAYS | int(default=1) * 86400 }}
no_warn: true
Goal - learning. See not a lot of practical need.
fwiw, Ive also learnt…
cant use shell_commands on any other than /config.
so, I moved my timed snapshots to the /local aka /www folder.
- service: camera.snapshot
data:
filename: >
/config/www/snapshots/timed/oprit_{{now().strftime('%Y_%m_%d_%H_%M')}}.jpg
changed the shell_command to
shell_command:
delete_snapshots: >
find /config/www/snapshots/timed/ -mtime +{{states('input_number.ouder_dan')|int(600)}} -exec rm {} \;
and can add that service: shell_command.delete_snapshots to the automation, or a button in the dashboard:
- type: button
tap_action:
action: call-service
service: shell_command.delete_snapshots
name: Delete snapshots
icon: mdi:delete-forever
execution of the delete command is immediate like this, so, if you need extra options, change the shell_command
And no need for yet another CC that was not maintained for 2 years+
Agree, thinking about using a stock “shell_command”.
Never used it before, so 1st chose that CC way.
only issue with that (the /www option) is that if you do backups and have a lot of camera snapshots, those are all stored in the backup too…
it was the main reason I chose the /media folder before, so I could exclude that from the backups.
Now, I can no longer do that, so I have to make a choice… not sure which is best yet
I am making backups manually ((
Lots to learn & implement in HA, no time for all these toys…
