feat: report nav-message data frequency for a decoder state

[giove-a]

feat: report nav-message data frequency for a decoder state

What

Bumps the GNSSSignals compat to 3.3 and adds a get_data_frequency method
for GNSSDecoderState: given a decoder, it returns the navigation-message
symbol rate of the signal that decoder demodulates.

using GNSSDecoder, GNSSSignals
get_data_frequency(GPSL1CADecoderState(1))  #  50 Hz
get_data_frequency(GPSL5IDecoderState(1))   # 100 Hz
get_data_frequency(GPSL2CMDecoderState(1))  #  50 Hz

Why

Downstream code (e.g. PositionVelocityTime.jl, which converts
num_bits_after_valid_syncro_sequence to seconds) needs the nav-message symbol
rate for a decoder. Today it hand-maintains a decoder→signal→rate map — one
method per signal — purely because GNSSDecoder exposed no way to recover a
decoder's signal. This lifts that mapping into GNSSDecoder, where it belongs,
so callers can just write get_data_frequency(decoder).

How

• Extends GNSSSignals.get_data_frequency on GNSSDecoderState (not piracy —
  the type is ours; Aqua confirms), one small method per signal, each next to
  its existing GNSSDecoderState(::Signal, prn) constructor.
• Each method forwards to the corresponding signal's rate in GNSSSignals
  rather than hardcoding a literal, so the value stays single-sourced.
• Dispatch is on the constants type (1:1 with the signal). This is what
  keeps GPS L5-I (100 Hz) and L2C-M (50 Hz) distinct even though both decode
  into a shared GPSCNAVData. The full-CBOC E1B and its BOC(1,1) approximation
  share one method (identical 250 Hz I/NAV stream).

Scope notes

• The compat lower bound moves to 3.3. The accessor itself doesn't require
  3.3 (get_data_frequency predates it), but this pins the intended baseline.
• Deliberately limited to get_data_frequency. It does not introduce a
  generic signal-type escape hatch, because ranging quantities (code/carrier
  frequency) belong to the tracked signal, which may differ from the decoded
  data component — a distinction a decoder state shouldn't blur.

Testing

• New test/data_frequency.jl: each decoder matches its signal's rate; L5-I and
  L2C-M stay distinct; E1B BOC(1,1) matches full E1B.
• Full suite green (2491/2491), including Aqua (no piracy/ambiguity).

Follow-up (separate PR, PositionVelocityTime.jl)

Delete the six nav_data_frequency methods in sc/multi-gnss and call
GNSSSignals.get_data_frequency(decoder) directly.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.15%. Comparing base (f592d02) to head (a05aba2).

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #72   +/-   ##
=======================================
  Coverage   98.14%   98.15%           
=======================================
  Files          14       14           
  Lines        1513     1519    +6     
=======================================
+ Hits         1485     1491    +6     
  Misses         28       28           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.