Data produced by the Mentalista module
One entity carries a WearableBioSensorComponent built from
the connected client’s prediction events. Raw sensor data - anything other than a prediction,
ready, commandSuccess or commandError event - is received but never processed.
Readings
JSON path in the prediction payload | BiometricReading name |
|---|---|
predictions.emotion.valence | emotion.valence |
predictions.emotion.arousal | emotion.arousal |
predictions.attention.attention_index | attention.attention_index |
predictions.cognitiveLoad.cognitive_index | cognitiveLoad.cognitive_index |
Each is optional independently - a prediction event missing one of these paths omits
that reading, it does not fail the whole event. If none of the four paths are present, no frame
is published for that event at all. Neither the client nor this module states a unit for any of
these four values; treat them as the vendor’s own model output, not a physical measurement with
an agreed unit.
Timestamp
The published capture time comes from the timestamp field in the prediction event’s own JSON
envelope, taken as milliseconds and used directly - this module does not stamp its own wall-clock
time the way most other capture modules do. Its epoch and base are defined by the client, not by
this module or by HEAT Capture.
Entity identity
The entity id is the fixed EntityId setting (default 1) for every published frame - this
module does not distinguish between multiple connected clients. The entity name is EntityName
if set, otherwise the device name from the connected client’s ready event, otherwise the literal string
mentalista.