Files
sideline/engine/sensors/__init__.py
David Gwilliam b27ddbccb8 fix(sensors): add inlet/outlet types to SensorStage
- Add DataType properties to SensorStage
- Fix MicSensor import issues (remove conflicting TYPE_CHECKING)
- Add numpy to main dependencies for type hints
2026-03-16 15:40:09 -07:00

204 lines
5.0 KiB
Python

"""
Sensor framework - PureData-style real-time input system.
Sensors are data sources that emit values over time, similar to how
PureData objects emit signals. Effects can bind to sensors to modulate
their parameters dynamically.
Architecture:
- Sensor: Base class for all sensors (mic, camera, ntfy, OSC, etc.)
- SensorRegistry: Global registry for sensor discovery
- SensorStage: Pipeline stage wrapper for sensors
- Effect param_bindings: Declarative sensor-to-param routing
Example:
class GlitchEffect(EffectPlugin):
param_bindings = {
"intensity": {"sensor": "mic", "transform": "linear"},
}
This binds the mic sensor to the glitch intensity parameter.
"""
from abc import ABC, abstractmethod
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any
if TYPE_CHECKING:
from engine.pipeline.core import PipelineContext
@dataclass
class SensorValue:
"""A sensor reading with metadata."""
sensor_name: str
value: float
timestamp: float
unit: str = ""
class Sensor(ABC):
"""Abstract base class for sensors.
Sensors are real-time data sources that emit values. They can be:
- Physical: mic, camera, joystick, MIDI, OSC
- Virtual: ntfy, timer, random, noise
Each sensor has a name and emits SensorValue objects.
"""
name: str
unit: str = ""
@property
def available(self) -> bool:
"""Whether the sensor is currently available."""
return True
@abstractmethod
def read(self) -> SensorValue | None:
"""Read current sensor value.
Returns:
SensorValue if available, None if sensor is not ready.
"""
...
@abstractmethod
def start(self) -> bool:
"""Start the sensor.
Returns:
True if started successfully.
"""
...
@abstractmethod
def stop(self) -> None:
"""Stop the sensor and release resources."""
...
class SensorRegistry:
"""Global registry for sensors.
Provides:
- Registration of sensor instances
- Lookup by name
- Global start/stop
"""
_sensors: dict[str, Sensor] = {}
_started: bool = False
@classmethod
def register(cls, sensor: Sensor) -> None:
"""Register a sensor instance."""
cls._sensors[sensor.name] = sensor
@classmethod
def get(cls, name: str) -> Sensor | None:
"""Get a sensor by name."""
return cls._sensors.get(name)
@classmethod
def list_sensors(cls) -> list[str]:
"""List all registered sensor names."""
return list(cls._sensors.keys())
@classmethod
def start_all(cls) -> bool:
"""Start all sensors.
Returns:
True if all sensors started successfully.
"""
if cls._started:
return True
all_started = True
for sensor in cls._sensors.values():
if sensor.available and not sensor.start():
all_started = False
cls._started = all_started
return all_started
@classmethod
def stop_all(cls) -> None:
"""Stop all sensors."""
for sensor in cls._sensors.values():
sensor.stop()
cls._started = False
@classmethod
def read_all(cls) -> dict[str, float]:
"""Read all sensor values.
Returns:
Dict mapping sensor name to current value.
"""
result = {}
for name, sensor in cls._sensors.items():
value = sensor.read()
if value:
result[name] = value.value
return result
class SensorStage:
"""Pipeline stage wrapper for sensors.
Provides sensor data to the pipeline context.
Sensors don't transform data - they inject sensor values into context.
"""
def __init__(self, sensor: Sensor, name: str | None = None):
self._sensor = sensor
self.name = name or sensor.name
self.category = "sensor"
self.optional = True
@property
def stage_type(self) -> str:
return "sensor"
@property
def inlet_types(self) -> set:
from engine.pipeline.core import DataType
return {DataType.ANY}
@property
def outlet_types(self) -> set:
from engine.pipeline.core import DataType
return {DataType.ANY}
@property
def capabilities(self) -> set[str]:
return {f"sensor.{self.name}"}
@property
def dependencies(self) -> set[str]:
return set()
def init(self, ctx: "PipelineContext") -> bool:
return self._sensor.start()
def process(self, data: Any, ctx: "PipelineContext") -> Any:
value = self._sensor.read()
if value:
ctx.set_state(f"sensor.{self.name}", value.value)
ctx.set_state(f"sensor.{self.name}.full", value)
return data
def cleanup(self) -> None:
self._sensor.stop()
def create_sensor_stage(sensor: Sensor, name: str | None = None) -> SensorStage:
"""Create a pipeline stage from a sensor."""
return SensorStage(sensor, name)