[lvgl] Build container_schema and widget update schemas lazily

Voluptuous schema construction inside container_schema() and the
register_action() chain in WidgetType.__init__ dominates cold-validate
time for any YAML that imports the lvgl component. cProfile of a cold
`import esphome.components.lvgl` shows ~225ms cumulative inside
container_schema (9 callers) and ~196ms cumulative inside
base_update_schema (30 callers, one per widget type), out of ~400ms
total.

Defer both: container_schema now returns a validator that builds its
underlying schema on first call, and WidgetType registers a lazy
validator with register_action so base_update_schema is only evaluated
when an `lvgl.<widget>.update` action is actually validated.

Measured on M-series Mac, fresh subprocess, 5 cold imports:

  Before: 389, 385, 389, 384, 380 ms (avg 385ms)
  After:  229, 217, 218, 215, 215 ms (avg 219ms)

A 43% reduction in `import esphome.components.lvgl`. Real
`esphome -q config` on tests/components/lvgl/test.host.yaml drops from
~770ms to ~550ms (warm-cache subprocess). Extrapolated to slower
hardware (Celeron N5105 dashboard hosts) this is ~1.5s saved per cold
LVGL save.

Schema output is unchanged: lvgl test configs validate identically and
`script/build_language_schema.py` produces a byte-identical lvgl
sub-tree before vs after.
This commit is contained in:
J. Nick Koston
2026-05-16 15:03:33 -07:00
parent 7c5d5f75dc
commit 3cee06e930
2 changed files with 49 additions and 11 deletions

View File

@@ -542,18 +542,27 @@ def container_schema(widget_type: WidgetType, extras=None):
:param extras: Additional options to be made available, e.g. layout properties for children
:return: The schema for this type of widget.
"""
schema = obj_schema(widget_type).extend(
{cv.GenerateID(): cv.declare_id(widget_type.w_type)}
)
if extras:
schema = schema.extend(extras)
# Delayed evaluation for recursion
# Schema construction is deferred until first validation. obj_schema and the
# part_schema -> base_update_schema chain account for ~225ms of cumulative
# work at lvgl module import time across ~9 callers; building lazily keeps
# `esphome config` fast for YAMLs that never reach this widget type.
schema_holder: list = []
schema = schema.extend(widget_type.schema)
def build() -> cv.Schema:
if not schema_holder:
built = obj_schema(widget_type).extend(
{cv.GenerateID(): cv.declare_id(widget_type.w_type)}
)
if extras:
built = built.extend(extras)
# Delayed evaluation for recursion
built = built.extend(widget_type.schema)
schema_holder.append(built)
return schema_holder[0]
def validator(value):
value = value or {}
return append_layout_schema(schema, value)(value)
return append_layout_schema(build(), value)(value)
return validator

View File

@@ -73,6 +73,30 @@ from ..types import (
EVENT_LAMB = "event_lamb__"
def _lazy_update_schema(widget_type: "WidgetType"):
"""Defer construction of a widget's update-action schema until first use.
base_update_schema(...).extend(widget_type.modify_schema) compiles several
voluptuous schemas and is invoked once per WidgetType at import time. The
caller (register_action) only needs a validator, so wrap the build in a
closure that materialises on the first validation and caches the result.
"""
compiled: list = []
def validator(value):
if not compiled:
from ..schemas import base_update_schema
compiled.append(
base_update_schema(widget_type, widget_type.parts).extend(
widget_type.modify_schema
)
)
return compiled[0](value)
return validator
class WidgetType:
"""
Describes a type of Widget, e.g. "bar" or "line"
@@ -113,18 +137,23 @@ class WidgetType:
# Local import to avoid circular import
from ..automation import update_to_code
from ..schemas import WIDGET_TYPES, base_update_schema
from ..schemas import WIDGET_TYPES
if not is_mock:
if self.name in WIDGET_TYPES:
raise EsphomeError(f"Duplicate definition of widget type '{self.name}'")
WIDGET_TYPES[self.name] = self
# Register the update action automatically, adding widget-specific properties
# Register the update action automatically, adding widget-specific
# properties. The update schema is built lazily on first validation:
# base_update_schema involves part_schema / Schema.extend chains that
# cost ~7ms per widget at import time (~200ms total across ~25
# widgets) and is unused for any YAML that never triggers a
# `lvgl.<widget>.update` action.
register_action(
f"lvgl.{self.name}.update",
ObjUpdateAction,
base_update_schema(self, self.parts).extend(self.modify_schema),
_lazy_update_schema(self),
synchronous=True,
)(update_to_code)