Commit graph

4 commits

Author SHA1 Message Date
7fb340bc35 Fix all remaining credo --strict issues (0 issues)
Aliases: add module aliases for 9 nested module references
Apply: replace apply/3 with direct module attribute calls
Line length: break 1 long spec line
Refactoring: extract helpers to reduce complexity and nesting
in show.ex, radio.ex, weather workers, terrain, duct detection,
backfill dashboard, contact map, and mix tasks
2026-04-12 10:26:53 -05:00
51e390959c Add dialyzer specs and types across the codebase
277 @spec/@type annotations added to 58 files covering all public
APIs: contexts (propagation, radio, weather, terrain, beacons,
commercial), GRIB2 decoders, terrain analysis, duct detection,
rain scatter, CSV/ADIF import, weather clients, and all Ecto
schemas. Dialyzer passes with 0 errors.
2026-04-12 08:55:04 -05:00
e7a7ae073d Phase 9.3, 9.4, and Phase 3 NEXRAD pipeline
Task 9.3 - Weight recalibration via gradient descent:
- Recalibrator module fits logistic regression weights using Nx
- Trains on QSO positives vs random baseline negatives
- Cross-validates by month, normalizes weights to sum to 1.0
- Mix task: mix recalibrate_scorer --sample 5000 --epochs 2000

Task 9.4 - Side-by-side scorer comparison:
- ScorerDiff.compare/3 re-scores grid with old vs new weights
- Reports mean diff, regressions, improvements, per-band breakdown
- Mix task: mix scorer_diff --new-weights '{...}'

Phase 3 - NEXRAD ingestion pipeline:
- NexradClient fetches IEM n0q composite PNGs, extracts per-point
  box statistics (mean/max dBZ, texture variance)
- NexradObservation schema with unique (lat, lon, observed_at)
- NexradWorker on :nexrad queue for background processing
- nexrad_texture backtest feature in Features module
- mix nexrad_backfill --limit 200

All tasks added to AdminTaskWorker and Release for production use.
1116 tests, 0 failures.
2026-04-10 12:48:36 -05:00
82bf248ab7 Phase 4: Ray-traced duct geometry
Duct module (Propagation.Duct):
- refractivity_profile/1: ITU-R P.453 N at each native level
- m_profile/1: modified refractivity M = N + 157*h(km)
- detect_ducts/1: find contiguous regions where dM/dh < 0, returning
  base/top height, thickness, and M-deficit per duct
- min_trapped_frequency_ghz/1: waveguide approximation (Bean & Dutton)
  for the minimum frequency a duct of given geometry can trap
- analyze/1: full pipeline from native profile to duct list + best
  trapped frequency across all ducts

Derive task updated to also compute ducts JSONB and best_duct_band_ghz
alongside the Phase 2 turbulence fields.

Backtest features: duct_thickness, best_duct_freq, duct_usable_10ghz,
duct_usable_24ghz, duct_usable_47ghz.

Real-data validation: 2022-08-20 12Z TX profile shows 0 ducts (M
increases monotonically) — correct for a well-mixed boundary layer
on a turbulent August afternoon (Ri=0.16).
2026-04-10 08:31:16 -05:00