The carrier for the signal whose feature bag you want to read — wire in the signal context this stage is evaluating.
Outputs
Buffer (time series)
A number line carrying the named feature’s value on each bar, or the default where the feature is missing.
Read Signal Feature (Scalar) pulls a named numeric value out of a signal’s feature bag and exposes it as a number line you can use downstream. It’s how a later stage reaches a measurement an earlier stage recorded — such as a confidence score or a measured leg size.
How it works
You give the node a `feature_name` and it reads that feature from the signal flowing through the pipeline, wired in via `value`, emitting it as a number. When the named feature isn’t present, the node returns the `default` you set instead of failing.
Configuration
feature_namescalar
The name of the numeric feature to read from the signal’s feature bag. It must match the name the upstream stage recorded; an unknown name returns the default.
defaultscalar
The number returned when the named feature is absent, so downstream math always receives a defined value rather than a gap.
Connects to
Reads from
Feeds into
Examples
Reuse a score: set feature_name to “confidence” so a position stage can size or threshold off the score the signal stage recorded.
Carry a measurement forward: read “atr_at_entry” with default 0 to feed a stored volatility reading into later distance math.
Common mistakes
Misspelling feature_name so the node silently emits the default for every bar, hiding the wiring error.
Leaving default at 0 when the feature is a divisor downstream, which can turn a missing feature into a divide-by-zero.