Module transforms 

Transform pipeline that converts a RawMessage into a typed Message.

The pieces here are deliberately decoupled and individually testable so the M5 typed decoder (crate::TypedDecoder) can compose them in the protocol-mandated order:

  1. SubField selection           (subfields)
  2. Components unpacking         (components + bit_stream)
  3. Accumulator running totals   (accumulator)
  4. Scale / Offset application   (scale_offset)
  5. Enum int → string            (enum_strings)
  6. DateTime conversion          (`crate::datetime`)

Reference: guide/fit_binary_learning_notes.md §“解码处理顺序（关键！）”.

Re-exports

pub use accumulator::Accumulator;

pub use bit_stream::BitStream;

pub use hr_merge::merge_heart_rates;

pub use memo_glob::decode_memo_glob;

Modules

accumulator

Accumulator — wraparound compensation for FIT counter fields.

bit_stream

LSB-first bit reader for the Components transform.

components

Components — split a packed integer field into multiple LSB-first lanes.

enum_strings

Resolve a numeric raw value to its canonical snake_case enum name.

hr_merge

Heart-rate merge — interpolate HR samples onto record messages.

memo_glob

MemoGlob reassembly — stitch multi-part text onto target messages.

scale_offset

Profile-level Scale/Offset transformation.

subfields

SubField selection — pick the correct semantic interpretation of a field based on the value of another field in the same message.