@pond-ts/fit API Reference
    Preparing search index...

    Interface ActivityStreams

    The time-series channels recorded during an activity. latlng is the track; everything else is optional and present only if the source provided it. Channels are parallel arrays — index i of each is the same sample.

    A sample a channel is missing is NaN, or null as an activity read back from JSON has it (JSON has no NaN): every reading in this package treats the two the same, as a gap, never as 0. timeSeconds is the exception: it must be finite at every sample.

    interface ActivityStreams {
        altitudeMeters?: number[];
        cadence?: number[];
        distanceMeters?: number[];
        heartrate?: number[];
        latlng: (GeoPoint | null)[];
        openWater?: boolean;
        temperatureC?: number[];
        timeSeconds?: number[];
        watts?: number[];
    }
    Index

    Properties

    altitudeMeters?: number[]
    cadence?: number[]

    Cadence, rpm.

    distanceMeters?: number[]
    heartrate?: number[]
    latlng: (GeoPoint | null)[]

    The position at each sample; null for a sample the device recorded without a GPS fix (before it found satellites, or a dropout). A GPS-less activity has an empty array. When some samples lack a fix and the device recorded distanceMeters, distance comes from that (the device kept counting without GPS); the map line joins only the fixes.

    openWater?: boolean

    An open water swim. The watch only gets a GPS fix while the arm is out of the water, so the fixes are sparse and noisy: the line is cleaned for the map (cleanSwimTrack in the geo module), and distance comes from distanceMeters (the watch's own filtered distance) rather than from the line, when the watch recorded it. Without it, distance is measured along the cleaned line, which cuts corners and so reads somewhat short.

    temperatureC?: number[]

    Temperature, °C.

    timeSeconds?: number[]
    watts?: number[]

    Power, watts (from a power meter — real, not estimated).