The skew-T commit (30c1018) doubled @pressure_levels from 13 to 25 so
new contact fetches would cover the full troposphere. That list is
also what PropagationGridWorker pulls per forecast hour, which
doubled the GRIB footprint (~57 MB compressed + 92k points × 25
levels × 3 vars decoded through wgrib2) and pushed prod pods over
their 4 Gi OOMKill threshold. Every chain died during f00 and the
map timeline never got beyond now and now+1h because the .ntms files
for f02-f18 were never written.
Split the constant:
* @profile_pressure_levels (25 levels, 1000-100 mb) drives the
per-contact HrrrClient.fetch_profile path so the skew-T plot
keeps its full-atmosphere trace.
* @grid_pressure_levels (13 levels, 1000-700 mb) drives the grid
hot path. That's the band SoundingParams.derive reads for
min_refractivity_gradient, and native hybrid-sigma data
(native_min_gradient) takes priority over the pressure-level
fallback anyway, so upper-air levels contribute nothing to
scoring — pure memory waste on this path.
build_profile/1 still iterates the full 25-level list; grid fetches
simply populate the 13 near-surface slots and skip the rest.
Makes HrrrClient.pressure_messages public with a :grid | :profile
variant so the split is testable from outside the module.
Historical contacts showed a skew-T log-P diagram that stopped at 700 mb
because HrrrClient and Era5Client only fetched pressure-level data down
to the top of the boundary layer. The chart canvas already ran up to
100 mb, so the trace clipped mid-atmosphere.
Two complementary fixes:
1. Extend @pressure_levels in HrrrClient and Era5Client with
650/600/550/500/450/400/350/300/250/200/150/100 mb so new fetches
cover the full troposphere + lower stratosphere.
2. Prefer the native hybrid-sigma profile for the contact-detail
skew-T when one has been backfilled for the contact's hour. The
native profile already stores all 50 hybrid levels up to ~19 km,
so historical contacts covered by the native backfill get a full
trace without re-hitting S3. A new HrrrNativeProfile.to_skew_t_profile/1
converts the parallel arrays into the %{"pres","tmpc","dwpc","hght"}
list shape the renderer expects, deriving dewpoint from SPFH via the
Magnus inverse. Weather.find_nearest_native_profile/3 mirrors
find_nearest_hrrr/3 for the lookup.
CDS returns status="rejected" (distinct from "failed") when a submit
is rate-limited past the 150-job per-user cap. The poll worker's
interpret_status_body treated this as an unknown status → generic
retryable error → 10 retries → discarded. 102 jobs hit this path
overnight.
Fix:
* Era5Client.check_status/1 now returns {:rejected, reason} (and
treats "dismissed" the same way — that's the status a manually
deleted job transitions to).
* Era5PollWorker handles :rejected exactly like :not_found: drop the
DB row, clean up the sibling CDS job, re-enqueue the submit, and
discard the current Oban job so it doesn't burn attempts retrying
a terminal state.
* Widened Era5SubmitWorker cap headroom from 10 → 30 (effective
ceiling 120 not 140). The race between concurrent workers all
passing the count check simultaneously meant we were reaching 141
in-flight against the 150 hard cap; 30 slots of headroom tolerates
a ~15-worker race before clipping CDS's ceiling.
Refactored handle_row/1 into handle_row + handle_non_both_done with a
leg_outcome/1 tag helper so the cross-product of leg statuses collapses
to a single pair match (credo complexity limit).
When CDS returns 404 for a poll check, the job is gone (manually
deleted, reaped, or never existed). Retrying will always 404, so treat
it as terminal: drop the era5_cds_jobs row, delete the sibling CDS job,
re-enqueue Era5SubmitWorker for the tile-month, and discard the Oban
poll job. Fixes 30 retryable poll jobs left behind by the interrupted
manual-cleanup script.
The single Era5MonthBatchWorker pinned an Oban slot for the full 30-45
min a CDS month-tile takes to assemble, and lost the work entirely on a
rolling deploy because the in-flight Task died with the pod. This splits
the flow into two tiny workers and persists the CDS job IDs so deploys
survive.
New pipeline:
1. Era5SubmitWorker (:era5_submit queue) — POSTs both CDS requests in
parallel, writes an `era5_cds_jobs` row with the returned job IDs,
and enqueues an Era5PollWorker scheduled +5 min. ~1s of real work.
Short-circuits when the month-tile is already cached or when a row
already exists (in-flight from a previous attempt).
2. Era5PollWorker (:era5_poll queue) — reads the row, calls
Era5Client.check_status/1 for both CDS job IDs, and:
- returns {:snooze, 300} if either job is still running (Oban
re-schedules without counting an attempt and releases the slot
immediately — a pod can keep dozens of tile-months in flight
without pinning workers)
- streams both GRIB files to disk via Req into: File.stream!,
decodes via Wgrib2.extract_grid_messages_from_file, bulk-inserts
via Era5BatchClient.decode_and_insert/6, deletes the row, and
DELETEs both completed jobs from CDS to free server-side quota
- if either leg CDS-reports failed, deletes the row + both CDS
jobs and returns {:error, reason}
Era5Client gains four testable building blocks:
submit_job/2 (bare POST → {:ok, job_id})
check_status/1 (GET → :running | {:done, src} | {:failed, reason})
download_source_to_file/3 (streams {:url, href} or writes {:body, bin})
delete_job/1 (DELETE /jobs/:id, treats 200/202/204/404 as :ok)
All Req calls now route through `era5_req_options` so tests can stub
CDS responses via Req.Test.stub(Era5Client, fn).
Era5MonthBatchWorker is retained as a thin forwarder to Era5SubmitWorker
so any jobs already in the :era5_batch queue on prod pods drain cleanly
on the next rolling deploy. Safe to delete in a follow-up.
Adds era5_cds_jobs table with a unique index on
(year, month, tile_lat, tile_lon) so duplicate submits collapse.
New queue config in runtime.exs:
era5_submit: local_limit 4, rate_limit 30/hour (burst protection)
era5_poll: local_limit 20 (polls are cheap GETs)
era5_batch: kept at 1 for legacy job drain, delete next cycle
Four independent wins from a perf audit of the era5_batch worker path:
1. Parallel CDS submits per job. fetch_single_level_month and
fetch_pressure_level_month have no ordering dependency — each blocks
30+ min on CDS queue/download — so running them in two linked Tasks
halves wall time for every job in the backlog.
2. Stream GRIB2 downloads directly to disk via Req `into: File.stream!`.
A month-tile GRIB is 50–200 MB and was previously slurped into the
BEAM heap and then written back out to a temp file before wgrib2 ran.
New Era5Client.submit_and_download_to_file/3 skips both copies. Saves
100–200 MB heap per worker × 12 concurrent ≈ 1.2–2.4 GB peak under
full parallelism. Partial files are removed on HTTP error.
3. Wgrib2.extract_grid_messages_from_file/3 lets batch decoding read the
streamed temp file directly. The binary variant still exists for the
single-point path and now delegates through the same helper.
4. Composite index (valid_time, lat, lon) on era5_profiles, matching the
three-range scan used by Weather.find_nearest_era5 (contact detail +
path profiles) and StatusLive.count_era5_done (status page EXISTS
subquery). The unique index on (lat, lon, valid_time) couldn't serve
it efficiently.
Also bumps the bulk insert chunk size 2k → 4k to halve insert round-trips
per month-tile (still well under Postgres' 65k-parameter cap).
CDS rejects `reanalysis-era5-pressure-levels` requests that include
dewpoint_temperature because it's only a single-level 2m variable; no
pressure-level dewpoint exists in ERA5. The failure mode is a 2–3 second
`{"status": "failed"}` response with no message field, which is why
prod was burning through Era5MonthBatchWorker attempts without anything
landing in era5_profiles.
Switch the pressure-level moisture variable to specific_humidity (SPFH),
update the wgrib2 extraction regex, and convert SPFH → dewpoint °C via
Weather.ThetaE.dewpoint_from_spfh/2 in the profile builder. Surface
dewpoint keeps using the single-level 2m dewpoint_temperature, which is
valid and already working.
Direct queries against prod (75M HRRR profiles, 16,864 soundings, 392k
surface obs, new hrrr_native_profiles table) drive these changes:
* Reinstate shallow-BL bonus as an HPBL multiplier on the refractivity
score (1.10× <200m → 0.78× ≥2000m). The Apr 11 "shallow BL bonus
removed" conclusion was an artifact of the prior matching strategy;
with the cleaner contact↔HRRR join (n=680) the binned data goes 230 km
avg at HPBL <200m vs 100 km at ≥2000m, monotonic across all bins.
* Add 6 missing bands (142, 145, 288, 322, 403, 411 GHz) so contacts in
those bands stop being silently dropped. Coefficients extrapolate
ITU-R P.676/P.838 trends from the existing 134/241 GHz entries.
* Bump ERA5 poll timeout 10 min → 1 hour and Era5MonthBatchWorker
max_attempts 3 → 5 so CDS slowness stops discarding tiles. Also dump
the full failure body when CDS omits the message field — the prior
"ERA5 job failed: nil" was masking real error reasons in oban_jobs.
* Add scripts/recalibrate_algo.py so this analysis can be re-run any
time new data lands without manual SQL. Reads PROP_PROD_DB_URL from
.envrc, drops a Markdown report into docs/algo-reports/.
* Append a dated Part 2c to algo.md documenting the corpus expansion,
the honest accounting of historical HRRR coverage (only ~1,020 of 58k
contacts are precision-matchable), and the empty era5/rtma/climatology
tables. Update the gaseous-absorption and rain-attenuation tables to
include the new bands.
Test suite: 1,335 tests, 0 failures.
IEM's RAOB endpoint stopped keeping Canadian stations current — most
stalled around Sep 2024 — which leaves a gap in refractivity/ducting
calibration over all the CW*/CY* stations already in weather_stations.
MSC's own tephi CSV is fixed-width and rendering-focused; the WMO TEMP
bulletins would need a full decoder. University of Wyoming serves the
same stations in a clean space-delimited format, is current, and
accepts the 3-letter station code (drop the leading C from Canadian
ICAO). Its output shape matches IemClient.parse_raob_json/1 exactly,
so it drops straight into Weather.upsert_sounding/2.
- New UwyoSoundingClient with URL builder, Req-based fetch_sounding,
and a fixed-width parse_sounding_html that extracts PRES/HGHT/TEMP/
DWPT/DRCT/SKNT from the <PRE> block and DateTime from the <H2> header.
- New CanadianSoundingFetchWorker: Oban batch worker that finds all
sounding stations whose code starts with C, picks the most recently
publishable 00Z or 12Z slot (90 min publish delay), calls UWYO per
station, and upserts with SoundingParams-derived refractivity/
ducting fields.
- Oban cron at 01:30Z and 13:30Z (dev + prod) to drive the worker.
- Req.Test plug wiring in config/test.exs.
- Captured Goose Bay fixture as test/support/fixtures/uwyo_sounding_yyr.html.
Also drops two implementation plans under docs/plans/:
- 2026-04-13-rdps-vertical-profiles.md: ~2 day plan for RDPS 10 km
Canadian ducting outside HRRR's Lambert footprint.
- 2026-04-13-hrdps-canadian-prop-grid.md: ~5 day plan for the full
HRDPS 2.5 km Canadian prop grid. Explicitly recommends doing RDPS
first.
TDD throughout, 19 new tests, 1318/1318 passing, credo strict clean.
MRMS
----
Layer the NOAA MRMS PrecipRate product onto the score grid so rain fade
updates every 2 minutes instead of every hour alongside HRRR. New modules:
- Microwaveprop.Weather.MrmsClient: fetches the latest .grib2.gz off the
NCEP mirror (Req auto-decompresses so no gunzip step), writes the raw
GRIB2 to a temp file, and calls the existing wgrib2 wrapper with the
0.125 propagation grid spec to get interpolated cells. Returns a
%{{lat, lon} => mm_per_hour} map with missing-value sentinels dropped.
- Microwaveprop.Weather.MrmsCache: ETS-backed GenServer mirroring
ScoreCache/GridCache. Caches a single "current" entry keyed by
valid_time with PubSub broadcast so peer nodes stay in sync and only
the Oban leader pays the fetch + regrid cost.
- Microwaveprop.Workers.MrmsFetchWorker: cron every 2 minutes, short-
circuits when the cached valid_time already matches the newest file.
Microwaveprop.Propagation.AsosNudge.compute/4 now takes an optional
rain_grid. When a cell has MRMS rain >= 0.1 mm/hr it gets patched onto
the HRRR profile's `precip_mm` field (the scorer already reads it there)
and the cell is re-scored even with no ASOS station nearby. Cells with
MRMS rain below the threshold aren't touched so dry cells keep their
raw HRRR scores (which have the wind/sky/native-gradient signal that
isn't persisted on HrrrProfile rows and would otherwise be lost).
AsosAdjustmentWorker pulls MrmsCache on every tick and passes the grid
through to AsosNudge.compute/4. Also skips the IemClient error branch
that can never happen and handles the ASOS-empty + MRMS-empty case
explicitly. MrmsCache wired into the supervision tree; MrmsFetchWorker
cron entry added to config.exs and dev.exs.
Four new AsosNudge cases cover MRMS-only re-scoring, threshold gating,
and the wet/dry score delta.
Beacons 500
-----------
Beacon.format_freq/1 and format_mw/1 crashed on whole-number floats
(e.g. 24192.0) because `frac == 0.0` could become false under float
rounding while `trim_trailing_zeros/1` stripped the decimal point,
leaving a 1-element list that couldn't be destructured as [_, frac].
Shared format_number/1 helper handles integer input directly and
pattern-matches both the "int-only" and "int + frac" shapes.
Added stream_data property tests covering the whole microwave range for
both integers and floats to catch this class of bug before prod.
UTC clock flash
---------------
The /weather and /map UTC clocks were empty until the JS hook mounted
post-WebSocket, producing a several-second blank spot on initial load
and a clobber risk on sidebar re-renders. Mount now computes a
server-rendered `initial_utc_clock` string and the template seeds the
element with that plus `phx-update="ignore"` so LiveView morphdom won't
overwrite what the hook writes.
Adds Microwaveprop.Propagation.AsosNudge: a pure IDW bias-field module
that takes ASOS observations + HRRR profiles and returns re-scored grid
rows for every cell within 250km of a reporting station. Upper-air
fields (min_refractivity_gradient, pwat_mm, hpbl_m, profile, duct
metadata) pass through unchanged so HRRR's signal isn't clobbered.
The old AsosAdjustmentWorker was unwired and buggy — nil'd out ~22% of
the scoring weight and wrote orphan timestamps. Replaced with a slim
worker that queries the latest HRRR valid_time, fetches live ASOS
currents, calls AsosNudge.compute/3, and upserts onto
(lat, lon, valid_time, band_mhz) so nudged values overwrite the HRRR
hour cleanly instead of polluting available_valid_times. After each
upsert it warms ScoreCache and broadcasts propagation:updated so live
/map clients refresh.
Cron hooked up every 10 minutes in config.exs and dev.exs. Also cleaned
up the stale "dev has propagation disabled" note in CLAUDE.md.
13 new AsosNudge unit tests cover: residual computation (co-located,
out-of-grid, nil fields), IDW weighting (single station, far station,
two equidistant stations, nil component handling), upper-air
preservation, and the compute/3 entry point's shape and radius filter.
Drive-by Styler formatting touched a handful of unrelated files from
`mix format`.
The /weather LiveView was timing out on pod restart because
latest_weather_grid/1 ran load_weather_grid_from_db/1 synchronously on
cache miss — that query reads 176k hrrr_profiles rows with two huge
JSONB columns (profile array + duct_characteristics), and JSONB
decoding blocked the LiveView process past the 15-second pool
ownership timeout.
On cache miss latest_weather_grid/1 now returns empty immediately and
kicks off a deduped background fill through GridCache.claim_fill/1 (a
per-valid_time ETS lock so N concurrent mounts after restart don't each
fire the slow query). When the fill completes it broadcasts
weather:updated, and the handle_info in WeatherMapLive hits the warm
cache and pushes rows over the socket.
Added load_weather_grid/1 as a synchronous sibling for tests and any
caller that genuinely needs inline data. Tests updated to use it and
to clear GridCache in setup so ETS state doesn't leak between cases.
Also untrack k8s/secret.yaml and add both it and k8s/*-secret.yaml to
.gitignore — those manifests hold plaintext SMTP and database
credentials that should not live in the repo. Secrets already in git
history should be rotated separately.
- Add Weather.GridCache: ETS cache of derived HRRR grid rows keyed by
valid_time, cluster-synced via PubSub. Eagerly warmed from
PropagationGridWorker after each upsert so /weather map pan/zoom and
weather_point_detail hit zero DB on warm cache.
- Replace latest_weather_grid DB query path with cache-first lookup +
DB fallback. hrrr_profiles is 42M rows partitioned; pulling 3-10k
rows per viewport on every pan was the main cost.
- ContactLive.Show: defer the heavy enrichment loads (weather, solar,
HRRR, terrain, IEMRE, elevation profile, ITU-R propagation analysis,
data_sources) into a handle_info(:hydrate) that runs after the shell
renders. Initial mount now returns nil placeholders; template already
had :if guards for all of them. Shell-to-first-paint goes from
~500ms-2s down to ~20ms.
- Cache fetch_queue_counts for 5s in ContactLive.Show — oban_jobs group
by query was running on every contact page view.
- Backfill stats: wrap count_unprocessed, fetch_stats, fetch_db_stats
in Microwaveprop.Cache with 2-5s TTLs; bump refresh debounce from 1s
to 2s so bulk enrichment events don't thrash the DB.
- ScoreCache stores {band, valid_time} as %{{lat, lon} => score} map so
point lookups are O(1); adds fetch_point/4 and valid_times/1
- available_valid_times/1 reads directly from ScoreCache when warm,
falls back to DB on cold start
- point_forecast/3 iterates cached valid_times and uses fetch_point/4
instead of hitting the DB per click
- NexradCache: node-local ETS cache of decoded n0q PNG pixel buffers
keyed by 5-minute rounded timestamp; skips ~1-5s HTTP+decode on
concurrent/repeat clicks within the same window
- MapLive: start_async the rain_scatter fetch so point_detail renders
immediately with a pending marker; push rain_scatter_update when
NEXRAD resolves
- MapLive: preload all 18 remaining forecast hours for the current
viewport after mount/band change/propagation_updated; client caches
them and renders timeline scrubs instantly without a server roundtrip.
Adds set_selected_time event for fast-path state sync.
- Propagation map JS: forecastCache map + drawScatterMarkers helper,
timeline click uses preloaded cache when available
- Replace %Struct{} with Struct.t() in all @spec annotations
- Replace length(x) > 0 with x != [] in test assertions
- Fix multi-line spec struct references in weather.ex
Registered users can suggest edits to any contact's core fields
(callsigns, grids, band, mode, timestamp). Edits enter an admin
approval queue with field-by-field diff view. On approve, changes
are applied and enrichment re-enqueued if grids/band changed.
Users receive email notification on approve or reject.
Also updates dependabot.yml for mix ecosystem.
When clicking a grid point, fetches the latest NEXRAD composite
reflectivity and identifies rain cells within 300 km that could
enable rain scatter contacts. Shows:
- Scatter classification (excellent/good/marginal/none)
- Top 3 cells with dBZ, distance, bearing, and relative signal
- Colored circle markers on the map at rain cell locations
- Markers sized by reflectivity, colored by intensity
Uses simplified bistatic radar equation accounting for reflectivity,
frequency-dependent scattering (Rayleigh/Mie), and R^4 path loss.
NEXRAD cells sampled every ~5 km within bounding box for efficiency.
When clicking a grid point with ducting, the panel now shows each
duct layer with base-top height in feet, thickness in meters, and
minimum trapped frequency. Data flows from Duct.analyze through
the scoring factors as a ducts array.
The PropagationGridWorker now fetches native hybrid-sigma levels
(TMP, SPFH, HGT, PRES × 50 levels) alongside the standard surface
and pressure products. Native data provides 10-50m vertical spacing
vs 250m from pressure levels, detecting thin surface ducts invisible
to the standard product.
Key design: cell-by-cell reducer in Wgrib2.extract_grid_from_file_mapped
processes each of the 95k CONUS cells through a duct analysis function
inline, keeping only scalar metrics per cell. Peak memory ~86 MB
instead of ~1.8 GB for the full grid map.
Per-cell output: native_min_gradient, best_duct_freq_ghz,
max_duct_thickness_m, duct_count. The scorer prefers the native
gradient over the pressure-level gradient when available.
Native fetch is optional — if it fails, scoring continues with
pressure-level data only.
Without -s, wgrib2 -lon only outputs msg:offset:lon=X,lat=Y,val=Z
with no variable name or level. The -s flag adds the short inventory
(d=DATE:VAR:LEVEL:...) so the parser can identify which variable
each value belongs to.
Points spread coast-to-coast created a ~476k cell bounding grid
(350 messages × 476k cells × 4 bytes ≈ 665 MB), causing OOM.
Switch to -lon which extracts values at specific lat/lon points
with text output. One wgrib2 call, one file scan, negligible
BEAM memory regardless of point geographic spread.
- Stream profile storage and score upsert instead of materializing
full 20k+ item lists (propagation_grid_worker, propagation.ex)
- GC between forecast hours and store/compute phases to reclaim
~400 MB of grid data between steps
- Single-pass field extraction in scorer.ex path_integrated_conditions
instead of 6 separate Enum traversals
- Eliminate intermediate merged map in fetch_grid by combining
merge + profile build into one pipe
- Fix UUID bug: bingenerate → generate in native grid worker
(same issue previously fixed in nexrad_worker)
The parallel download was holding all ~530MB of range responses in
memory before writing to disk. Now each range is fetched and written
one at a time, so only one chunk is in memory at a time.
Instead of holding ~530MB GRIB binary in BEAM memory, download
ranges directly to a temp file and run wgrib2 on it. Peak memory
drops from ~530MB to just HTTP chunk buffers.
FrontalAnalysis module (Weather.FrontalAnalysis):
- detect_fronts/3 computes the Thermal Front Parameter (TFP) from
2D grids of surface temperature and pressure using Nx vectorized
ops. TFP = -nabla|nabla(theta)| . nabla(theta)/|nabla(theta)|.
Most negative values mark cold fronts.
- central_gradient/1 for 2D finite differences with edge handling
- nearest_front/3 finds closest front point with distance and bearing
- path_front_angle/2 computes angle between a QSO path and the
front (0 = parallel = good, 90 = crosses = dead)
Backtest feature stubs for distance_to_front and parallel_to_front
(return nil until the pipeline caches per-cell frontal features from
the hourly HRRR grid run). The FrontalAnalysis module itself is
tested and ready for integration.
NEXRAD spike docs also included in this commit.
NCEI ASOS 5-minute data client (Weather.NceiMetarClient):
- fetch/3 pulls per-station monthly .dat files from NCEI C00418
- parse/1 decodes the fixed-width METAR format including precise
T-group temperatures (T02110094 → 21.1/9.4°C)
- metar_5min_observations table: schema-identical to
surface_observations, separate table to avoid mixing cadences
Weather.recent_surface_obs/3 prefers 5-min data when available,
falls back to the hourly surface_observations table.
Data URL: https://www.ncei.noaa.gov/data/automated-surface-observing-system-five-minute/access/YYYY/MM/asos-5min-KXXX-YYYYMM.dat
Available back to 1996.
Phase 3 NEXRAD spike: IEM n0q composite available at 5-min cadence
back to 2022+. Compression-ratio proxy shows afternoon images have
13-81% more texture than dawn (directionally correct), but the n0q
product thresholds out the faint clear-air returns needed for BL
stability detection. Parked until MRMS or Level III products can be
investigated. See docs/research/nexrad_spike.md.
Phase 6: hrrr_climatology table aggregating surface_temp_c by
(lat, lon, month, hour) from the 42M+ hrrr_profiles grid-point
rows. mix hrrr_climatology builds it via a single SQL GROUP BY +
upsert. Backtest.Features.temperature_anomaly computes current_temp
minus climatological mean — the meteorologist's "temperature
deviation above normal" predictor for summer afternoon enhancement.
Inversion detection module (Propagation.Inversion):
- find_inversion_top/1 walks the native profile to locate the first
temperature inversion (surface-based or elevated)
- bulk_richardson/3 computes the Richardson number across the
inversion layer (Ri < 0.25 = turbulent, > 1 = laminar/good)
- shear_magnitude/3 computes the wind shear vector magnitude
- potential_temperature/2 for θ = T*(P0/P)^0.286
Theta-e module (Weather.ThetaE):
- Bolton (1980) equivalent potential temperature
- dewpoint_from_spfh/2 via Magnus-Tetens inversion
- theta_e_jump/3 for the thermodynamic decoupling metric
mix hrrr_native_derive_fields populates inversion_top_m,
bulk_richardson, theta_e_jump_k, and shear_at_top_ms on existing
hrrr_native_profiles rows.
First real data: 2022-08-20 12Z TX profile shows inversion at
186 m, Ri = 0.16 (turbulent), θ_e jump = 0.33 K — consistent with
marginal propagation conditions at that hour.
wgrib2 -lola ... bin writes Fortran unformatted records (4-byte
length header + data + 4-byte length trailer per message).
parse_lola_binary was treating the binary as tightly packed,
causing every message after the first to read from the wrong
offset — values came out as garbage across all grid points.
Fix: account for the 8-byte record overhead per message when
computing the data offset for each message's grid values.
This bug affects both the existing propagation grid extraction
(which may have been producing subtly wrong scores) and the new
native-level extraction (which was producing obviously wrong
values). The fix is a one-line stride change.
Also adds Backtest.Features.native_surface_refractivity for the
Phase 1 sanity check, plus a tighter wgrib2 match pattern that
selects only hybrid-level messages from the native file.
The Elixir GRIB2 decoder didn't map level type 105 (hybrid) or
variable IDs for SPFH and TKE, so native-level messages decoded as
"unknown:105:N" keys that build_native_profile couldn't find. Add
the three missing mappings to Section.identify_level/identify_var.
Also add HrrrNativeClient.extract_native_profiles/2 which uses
wgrib2's -lola on a tight bounding-box subgrid for speed (the pure
Elixir decoder takes ~70s per point on a 395 MB file; wgrib2 handles
40 points in seconds). The worker now routes through this path.
- Spike docs at docs/research/hrrr_native_levels.md confirming files
are on AWS S3 for 5+ years, 50 hybrid levels, and include TKE and
SPFH needed for Phase 2 turbulence features. Architectural finding:
per-point on-demand fetching is impractical (~530 MB/file), so
the ingestion worker batches per (date, hour) instead.
- hrrr_native_profiles schema: arrays per level plus cached surface
scalars and placeholder columns for Phase 2/4 derived fields.
Strictly additive — the existing hrrr_profiles table is untouched.
- HrrrNativeClient: pure URL/message-list helpers, build_native_profile/1
that turns a parsed wgrib2 map into the schema shape (TDD'd).
- Exposed HrrrClient.download_grib_ranges/2 so the native client
reuses the existing parallel byte-range download + disk cache.
- HrrrNativeGridWorker: Oban worker keyed on {year, month, day, hour},
unique at :infinity, pulls distinct (lat, lon) points from contacts
in the ±30 min window, downloads the native grib2, extracts per
point, bulk-upserts.
- mix hrrr_native_backfill --limit N enqueues the top-N hours by
contact count.
Phase 1 gate still pending Task 1.6 (sanity-check backtest after
live data lands).
Per-point ERA5 fetches were tragically slow because every point-hour
triggered its own asynchronous CDS job (submit → poll → assemble →
download). For backfill this meant thousands of independent jobs
queued against Copernicus. The new path groups requests by calendar
month and a 2° × 2° lat/lon tile so one CDS cycle populates ~60k
profiles at once, and Oban uniqueness on (year, month, tile_lat,
tile_lon) collapses every duplicate enqueue.
- Era5BatchClient builds the monthly CDS requests, extracts every
(lat, lon, hour) from the GRIB2 blob with wgrib2, derives
refractivity params, and bulk-inserts in 2k-row chunks with
on_conflict: :nothing. fetch_month_into_db/1 short-circuits when
the month-tile already has any cached profile.
- Era5MonthBatchWorker runs the batch on the :era5 queue with a
generous backoff (10m → 1d) and the uniqueness key above.
- Era5FetchWorker is now a thin router: cache hit → :ok, cache miss
→ enqueue the month-batch for the point's tile-month and return.
No more per-point CDS calls.
- Wgrib2 grows extract_grid_messages/3 which preserves per-message
datetimes by parsing `d=YYYYMMDDHH[MMSS]` from the inventory, so a
single GRIB2 file carrying a whole month decodes correctly.
- The era5_backfill mix task enqueues month-tile batches directly.
- Fix score_pressure crash on nil pressure_mb (coastal HRRR points)
- Set 10-min timeout on grid score upsert transaction (was :infinity)
- Single DELETE for prune_old_scores instead of N queries in a loop
- Remove dead load_hrrr_refractivity that loaded 95k rows into nil map
- Pass selected_time to point_detail to skip latest_valid_time sub-query
- Batch station existence checks (1 query per path point, not per station)
- Batch solar index upserts via insert_all in chunks of 500
- Batch backfill_distances via single UPDATE FROM VALUES statement
- Add is_grid_point boolean + partial index to hrrr_profiles (replaces
non-sargable modular arithmetic filter on every weather map query)
- Add partial index on contacts(qso_timestamp) WHERE pos1 IS NOT NULL
- Move backfill enqueue to Oban worker so UI returns immediately
- Cache raw GRIB2 byte ranges to ~/hrrr in dev (never deleted)
- mix hrrr_backfill re-fetches QSO-linked HRRR profiles with 13 levels
- Supports --all (re-fetch everything) and --limit N flags
- Groups by HRRR hour to batch requests, 500ms rate limit
- AWS archive has full HRRR history, same URL pattern as live feed
Fetch every 25mb from 1000-700mb (13 levels, up from 8). Gives ~80m
vertical spacing near the surface, enough to resolve ducting layers
that were previously invisible between the coarse 250m level spacing.
- HrrrClient.hrrr_url accepts forecast_hour param (wrfsfcfHH.grib2)
- PropagationGridWorker fetches all 19 forecast hours per run
- Propagation.scores_at/3 queries scores at specific valid_time
- Propagation.available_valid_times/1 returns all forecast times for timeline
- Pruning keeps scores with valid_time >= now - 2h (forecast-aware)
- MapLive: select_time event, timeline data pushed to JS
- JS: forecast timeline bar at bottom of map with clickable hour buttons
- PubSub broadcast sends list of valid_times instead of single time
Req default 15s receive_timeout too short for GRIB2 byte-range
downloads on prod. Also reduce HRRR queue concurrency from 20 to 5
to avoid NOAA rate limiting (was causing burst-then-stall pattern).
New AsosAdjustmentWorker runs every 10 minutes:
- Fetches latest ASOS observations from all ~2900 US stations via
IEM bulk currents API (parallel fetch across 51 state networks)
- For each grid point within 75km of a reporting station, re-scores
using fresh ASOS data (temp, dewpoint, wind, sky, pressure, precip)
with HRRR refractivity gradient from the last hourly computation
- Pushes updated scores to the map via PubSub
Also stores HRRR profiles in the database during grid computation
so the data persists for reference and ASOS blending.
Add a separate wgrib2-builder stage that compiles wgrib2 from source.
Docker layer caching means this only builds once — subsequent deploys
reuse the cached layer. The binary is copied into the final runtime
image. Also fixed wgrib2 path lookup to be runtime instead of
compile-time so it works in the Docker build pipeline.
