Improve the lovelace yaml parser for Home Assistant.

See my floorplan card for an example of what's possible.

• Copy the contents of custom_components/lovelace_gen/ to <your config dir>/custom_components/lovelace_gen/.
• Add the following to your configuration.yaml:

lovelace_gen:

lovelace:
  mode: yaml

• Restart Home Assistant

This integration changes the way Home Assistant parses your ui_lovelace.yaml before sending the information off to the lovelace frontend in your browser. It's obviously only useful if you are using YAML mode.

To rerender the frontend, use the Refresh option from the three-dots-menu in Lovelace

Any yaml file that is to be processed with lovelace_gen MUST have the following as its first line:

# lovelace_gen

Important: For some reason, which I can't seem to nail down, things stop working if you add # lovelace_gen to ui-lovelace.yaml. Adding it to any file included from ui-lovelace.yaml works, though.

The changes from the default generation include

You can now use Jinja2 templates in your lovelace configuration.

This can be used e.g. to

• Set and use variables

{% set my_lamp = "light.bed_light" %}

type: entities
entities:
 - {{ my_lamp }}

• Loop over lists

{% set lights = ['light.bed_light', 'light.kitchen_lights', 'light.ceiling_lights'] %}

- type: entities
  entities:
  {% for l in lights %}
    - {{ l }}
  {% endfor %}

- type: horizontal-stack
  cards:
    {% for l in lights %}
    - type: light
      entity: {{ l }}
    {% endfor %}

• Use macros

{% macro button(entity) -%}
  - type: entity-button
    entity: {{ entity }}
    tap_action:
      action: more-info
    hold_action:
      action: toggle
{%- endmacro %}

type: horizontal-stack
cards:
  {{ button("light.bed_light") }}
  {{ button("light.ceiling_lights") }}
  {{ button("light.kitchen_lights") }}

Please note that for this to work, the indentation of the code in the macro block must be equal to what it should be where it's called.

• Add conditional parts

{% if myvariable == true %}
Do something
{% endif %}

This is NOT dynamic. The values of variables are locked down when you rerender the interface.

This might make conditions seem pointless... but they work well with the next feature.

Normally, you can include a file in your lovelace configuration using

view:
  - !include lovelace/my_view.yaml

lovelace_gen lets you add a second argument to that function. This second argument is a dictionary of variables and their values, that will be set for all jinja2 templates in the new file:

type: horizontal-stack
cards:
  - !include
    - button_card.yaml
    - entity: light.bed_light
  - !include
    - button_card.yaml
    - entity: switch.decorative_lights
  - !include
    - button_card.yaml
    - entity: light.ceiling_lights
      name: LIGHT!

button_card.yaml

# lovelace_gen
{% if entity.startswith("light") %}
type: light
{% else %}
type: entity-button
{% endif %}
entity: {{ entity }}
name: {{ name }}

Be careful about the syntax here. Note that the arguments are given as a list and is indented under the !include statement. The second item in the list is a dictionary.

Note: If you want to pass a dictionary of values into a file, you need to convert it to json first:

{% set mydict = {"a": 1, "b": 2} %}
variable: {{ mydict | tojson }}

And then convert it back from json inside the file:

content: The value of a is {{ (variable | fromjson)['a'] }}

The fromjson filter is a feature of lovelace_gen and not normally included in jinja.

If you use lots of custom lovelace cards, chances are that you have run into caching problems at one point or another.

I.e. you refresh your page after updating the custom card, but nothing changes.

The answer is often to add a counter after the URL of your resource, e.g.

resources:
  - url: /local/card-mod.js?v=2
    type: module

lovelace_gen introduces a !file command that handles this for you.

resources:
  - url: !file /local/card-mod.js
    type: module

After this, lovelace_gen will automatically add a random version number to your URL every time you rerender the frontend. You won't have to worry about cache ever again.

This can also be used for pictures.

ui_lovelace.yaml

# lovelace_gen
resources:
  # When you regenerate, the browser cache for this file will be invalidated
  - url: !file /local/card-mod.js
    type: module
...

views:
 - ! include lovelace/my_cool_view.yaml

lovelace/my_cool_view.yaml

# lovelace_gen
{% set my_lights = ["light.bed_light", "light.kitchen_lights", "light.ceiling_lights"] %}
title: My view
cards:
  - type: entities
    entities:
{% for light in my_lights %}
      - {{ light }}
{% endfor %}

  # Include files with arguments
  # NB: JSON format for arguments
  # And NO SPACE after the colons!
  - !include
    -floorplan.yaml
    - lamps: true
      title: With Lamps

# Use this if you want lovelace_gen to ignore the jinja
{% raw %}
  - type: markdown
    content: |
      # Coming soon(?)

      A built-inmarkdown card with jinja templating.
      So I can tell that my light is {{ states('light.bed_light') }}!
{% endraw %}

  - !include
    - floorplan.yaml
    - title: No lights

lovelace/floorplan.yaml

# lovelace_gen
{% macro lamp(entity, x, y) -%}
{% if lamps %}
- type: state-icon
  entity: {{ entity }}
{% else %}
- type: custom:gap-card
{% endif %}
  style:
    left: {{ x }}%
    top: {{ y }}%
{%- endmacro %}

type: picture-elements
title: {{ title }}
image: https://placekitten.com/800/600
elements:
  {{ lamp('light.bed_light', 25, 25) }}
  {{ lamp('light.kitchen_lights', 50, 25) }}
  {{ lamp('light.ceiling_lights', 50, 50) }}

With lovelace_gen installed, you'll be able to redefine node anchors in the same file. A feature in the YAML specification, but an error in the engine Home Assistant normally uses...

You can add variables to the lovelace_gen configuration in configuration.yaml and then refernce them in lovelace using {{ _global }}.

E.g.:

lovelace_gen:
  rooms:
    - living_room
    - kitchen
    - bed_room

type: entities
entities:
  {% for room in _global.rooms %}
  - type: custom:auto-entities
    card:
      type: custom:fold-entity-row
      head:
        type: section
        label: {{ room|capitalize }}
    filter:
      include:
        - area: {{ room }}
  {% endfor %}

It's called lovelace_gen for a reason... That being said - it might work. Or it might not. There's really no way to tell. It depends on what parts of the configuration are loaded before or after lovelace_gen itself.

I'd advice you not to try it.

Use the {% raw %} and {% endraw %} tags. There's an example above.

Not automatically, but you can do something like

{% macro button(entity, ws) %}
{{"  "*ws}}- type: entity-button
{{"  "*ws}}  entity: {{ entity }}
{{"  "*ws}}  tap_action:
{{"  "*ws}}    action: more-info
{{"  "*ws}}  hold_action:
{{"  "*ws}}    action: toggle
{%- endmacro %}

  - type: horizontal-stack
    cards:
      {{ button('light.bed_light', 3) }}

  {{ button('light.bed_light', 1) }}

Note that included files don't have this problem.