grabowski
779728e38b
Phase 3 of docs/RUST_REWRITE.md: rust/collector/src/web (axum) serves what
the FastAPI app (src/web_api.py) served, so one process collects, forecasts,
notifies and serves. With WEB_LISTEN=0.0.0.0:8000 it runs inside
`ping-collector run`; `ping-collector serve` runs it alone.
Same routes, parameters, JSON fields and field order, status codes,
{"detail": ...} errors and pydantic-style 422 lists, Cache-Control, gzip
(>500 B), CORS, Umami API tracking, the dashboard and static files, and the
stale-while-revalidate response caches (now with byte budgets: a history
response can be 10 MB). /docs is the same Swagger UI (+ /redoc); its OpenAPI
document is built from the route table, whose per-route parameter
declarations also validate the requests, so docs and behaviour cannot drift.
Python output conventions are kept: isoformat(), round(), numpy's pairwise
sums (catchment means), Python's summation order (skill metrics).
scripts/api_parity.py compares the two servers route by route on one
database (53 cases incl. errors, admin and static files; query values vary
per run so FastAPI's caches answer fresh). Locally: 53 identical; the
station CRUD writes byte-identical stations.json files; the skill/trend
logic matches on 9 station/horizon pairs with synthetic model versions.
Intended differences: /health counts the ThaiWater telemetry of the same
gauges while RID lags (as the notifier); /scraping/status adds the collector
version and last forecast run; an old stored forecast is served marked
X-Data-Stale instead of recomputed; /sensors/thaiwater is cached a minute;
.geojson is served as application/geo+json. Load (local, 16 clients): the
dashboard's routes 1375 vs 552 req/s, uncached history 714 vs 53 req/s;
25 MB resident when serving only.
Matrix (MATRIX_*): the notifier sends the room the transitions ntfy gets
(per station, outlook, feed; not ntfy's per-topic summaries), each once,
with its own matrix: keys in notification_state, seeded from ntfy's on first
use so nothing already announced repeats. Checked against a fake homeserver.
It replaces the `run.py --alert-check` cron job (repeated alerts every 15
min), which install_collector.sh now comments out (crontab backup kept).
install_collector.sh: --check-api (ping-collector serve on :8001 next to
FastAPI, api_parity.py), --cutover-api (refuses on a difference; sets
WEB_LISTEN, stops and disables water-monitor.service), --rollback-api;
--rollback-forecast and --rollback bring the Python app back as server.
ping-collector.service: MemoryMax 512M. Config::from_vars for tests; route
tests against a throwaway schema.
148 lines
9.0 KiB
Markdown
148 lines
9.0 KiB
Markdown
# A Rust rewrite: assessment
|
|
|
|
Status: **in progress.** Phase 1 (collector) is in production since 2026-10-17,
|
|
phase 2 (forecast + ntfy) since 2026-10-18; phase 3 (website/API + Matrix) is
|
|
built and parity-checked, cutover pending; see
|
|
[rust/README.md](../rust/README.md). Spike: [spikes/rust-forecast](../spikes/rust-forecast/README.md).
|
|
|
|
## Short version
|
|
|
|
- **Feasible.** The risky part, serving the scikit-learn forecast from Rust with
|
|
no train/serve skew, works: the spike reproduces the pandas feature matrix
|
|
and the HistGradientBoosting regression output bit for bit (classifier heads
|
|
to the last ulp).
|
|
- **Training stays Python.** There is no Rust HGB that trains like
|
|
scikit-learn, and the evaluation harness (`scripts/evaluate_variants.py`) is
|
|
where model work happens. A "Rust rewrite" means a Rust service (collectors,
|
|
API, forecast serving, ntfy) plus a Python offline job (monthly retrain and
|
|
evaluation) that exports models for Rust.
|
|
- **Performance is not a reason.** The forecast takes 158 ms per station in Python
|
|
and 0.25 ms in Rust, but it runs once an hour. Load tests showed the site
|
|
is bandwidth-bound once caches are hit.
|
|
- **Real reasons, if any:** one process instead of leader-elected uvicorn workers
|
|
(the root of per-worker state bugs like `/scraping/status`), a fraction of the
|
|
memory (324 MB today), typed parsing and SQL, one deployable binary, and the
|
|
same stack as proplan (axum + sqlx).
|
|
- **Recommendation:** no big-bang rewrite. If Rust is wanted, start with the
|
|
collector + notifier daemon (phase 1 below), which stands alone and can be
|
|
abandoned without loss. The cheap Python alternative is at the end.
|
|
|
|
## What exists (Python, `src/`)
|
|
|
|
| Area | LOC | Fate in Rust |
|
|
| --- | ---: | --- |
|
|
| Collectors: RID hourly, HII/ThaiWater, RID dams, Open-Meteo | 2,500 | port; typed serde payloads |
|
|
| Web API (20 routes), schemas, health, metrics | 2,300 | port with identical JSON (golden-file tests) |
|
|
| DB adapters: Postgres, SQLite, MySQL, InfluxDB, VictoriaMetrics | 1,300 | Postgres only via sqlx, existing 13 tables, no data migration |
|
|
| ntfy notifications; Matrix alerting (CLI path) | 1,050 | port ntfy; Matrix only if still used |
|
|
| ML serving: predict, features, rain/dam/HII fill | ~1,300 | port; parity-tested (spike) |
|
|
| ML training + evaluation | ~1,700 | **stays Python** |
|
|
| CLI (`main.py`) | 650 | small clap CLI (`fill-gaps`, test cycle) |
|
|
| `static/dashboard.html` (164 KB) | | unchanged, served as a static file |
|
|
| tests | 3,900 | Rust unit tests + parity fixtures + API golden files |
|
|
|
|
## Target shape
|
|
|
|
One binary, one process (tokio):
|
|
|
|
- tasks: RID / HII / dams / rain collectors on their schedules; forecast
|
|
precompute when new readings land; ntfy notifier. Shared state is an
|
|
`Arc<RwLock<..>>`, so there's no leader election, no status hand-off file and
|
|
no executor sizing.
|
|
- axum + tower-http: same routes and JSON, compression (gzip/brotli also covers
|
|
the 257 KB river GeoJSON), static dashboard, CORS; `utoipa` if `/docs` stays.
|
|
- sqlx against the existing schema. Python and Rust can run against the same
|
|
database during migration.
|
|
- models: `train_all()` also exports each bundle (`models/flood_<station>.json.gz`,
|
|
1.2-2.3 MB), `scripts/retrain.sh` promotes it. Rust reloads changed files on
|
|
its next run. A bundle whose `feature_names` Rust cannot build falls back to
|
|
the heuristic, as Python does.
|
|
|
|
## The permanent cost: the Python/Rust model contract
|
|
|
|
Every feature added in `features.py` must be ported to Rust in the same change,
|
|
or the new bundle is refused. That has to be enforced by CI: Python writes a parity
|
|
fixture per station (as `export_fixture.py` does), and a Rust test asserts
|
|
bit-identical features and regression output. The spike found two traps that
|
|
silently flip tree branches (HGB thresholds often equal real data values):
|
|
serde_json's default float parsing (1 ulp off; enable `float_roundtrip`) and
|
|
pandas' Kahan-compensated rolling sums/means (replicated in `rolling_kahan()`).
|
|
This slows model iteration, which is the most active part of the repo.
|
|
|
|
## Benefits and costs
|
|
|
|
| For | Against |
|
|
| --- | --- |
|
|
| Single process: no leader election, no per-worker state | Two languages forever; parity CI is mandatory |
|
|
| Memory ~324 MB -> est. 60-80 MB (mostly tree arrays; less with u32/u16 nodes) | Weeks of work with little user-visible change |
|
|
| Typed RID/HII parsing, compile-checked SQL | Re-introduces bugs into recently hardened code (tz, alerts, security) |
|
|
| One static binary; no venv on the server for the API | Server still needs Python + sklearn for the monthly retrain |
|
|
| Same stack as proplan | Feature work = Python + Rust change every time |
|
|
|
|
## Plan (each phase ships on its own)
|
|
|
|
| Phase | Scope | Status |
|
|
| --- | --- | --- |
|
|
| 0 | Spike: P.1 features + HGB inference | done, bit-identical |
|
|
| 1 | `ping-collector`: RID, HII, dams, Open-Meteo | in production since 2026-10-17 23:09 (Python: `COLLECTION_ENABLED=false`) |
|
|
| 2 | Forecast crate + ntfy | in production since 2026-10-18 15:08 (`FORECAST_IN_COLLECTOR=true`); live check on production data: 48/48 rows identical, 96/96 probabilities bit-identical |
|
|
| 3 | Website + read API in `ping-collector`, Matrix like ntfy | built 2026-10-18; locally 53 of 53 compared routes identical to FastAPI (`scripts/api_parity.py`); cutover with `install_collector.sh --cutover-api` |
|
|
| 4 | Cleanup | |
|
|
|
|
0. **Spike: done.** P.1 features + HGB inference, bit-identical.
|
|
1. **Collector daemon** (`rust/collector`, ~1.5k LOC incl. tests): RID with the
|
|
fixed semantics (store late readings, retry while behind, re-fetch
|
|
yesterday's holes), HII, dams, Open-Meteo, into the same tables. The Python
|
|
API runs with `COLLECTION_ENABLED=false` and re-forecasts when the newest
|
|
RID reading in the database advances. ntfy moved to phase 2: the outlook
|
|
alert needs the forecasts, which are still Python. Cutover and rollback
|
|
are one command each (`scripts/install_collector.sh`).
|
|
2. **Forecast crate + ntfy** (`rust/forecast`, run by `ping-collector` with
|
|
`FORECAST_IN_COLLECTOR=true`): all stations, HII/rain/dam inputs, heuristic
|
|
fallback, probabilities and stages, `forecast_history`, `forecast_current`
|
|
(served by the API) and `notification_state`. Parity: fixtures from the
|
|
Python production path (CI: P.1 model; locally all 15), a live comparison
|
|
script, and 751 notifications replayed word for word. Porting found a
|
|
Python serving bug, fixed first: forecast rain (`rain_fc24`) was always
|
|
NaN at serving time (docs/FLOOD_FORECASTING.md section 5). The ntfy change:
|
|
ThaiWater mirror readings keep level alerts and the feed check current
|
|
while RID lags at night. The Python API then only serves.
|
|
3. **Website + read API** (`rust/collector/src/web`, axum): every FastAPI
|
|
route with the same parameters, JSON fields and order, status codes,
|
|
FastAPI-style errors (`{"detail": ...}`, pydantic's 422 list), cache
|
|
headers, gzip, CORS, Umami API tracking and the stale-while-revalidate
|
|
response caches (now with byte budgets). `/docs` is the same Swagger UI;
|
|
its OpenAPI document is generated from the route table whose parameter
|
|
declarations also do the validation. One process serves, collects,
|
|
forecasts and notifies (`WEB_LISTEN`), so the status file, the leader
|
|
election and the `forecast_current` hand-off between processes are gone
|
|
in effect. Checked by `scripts/api_parity.py`, which runs `ping-collector
|
|
serve` next to the FastAPI app on the same database and compares every
|
|
route (query values vary per run so FastAPI's caches answer fresh);
|
|
`--cutover-api` refuses to switch on a difference. Intended differences:
|
|
`/health` counts ThaiWater telemetry of the same gauges while RID lags (as
|
|
the notifier); `/scraping/status` adds the collector version and the last
|
|
forecast run; `/forecast` serves an old stored forecast marked
|
|
`X-Data-Stale` instead of recomputing (same process); `/sensors/thaiwater`
|
|
is cached for a minute; database errors serve the last good response
|
|
where FastAPI answered `[]`. Matrix: the notifier's transitions go to the
|
|
room too (per station, outlook, feed; not the per-topic summaries), with
|
|
their own `matrix:` state keys seeded from ntfy's, replacing the
|
|
`run.py --alert-check` cron job that repeated alerts every 15 minutes.
|
|
Python keeps training, evaluation and the monthly retrain timer.
|
|
4. **Cleanup:** drop FastAPI/uvicorn, the Python collector, notifier, CLI and
|
|
the multi-backend adapters (SQLite, MySQL, InfluxDB, VictoriaMetrics) and
|
|
demo databases; the Python environment shrinks to the training stack.
|
|
|
|
Decided before phase 3 (2026-10-18): Matrix stays and behaves like ntfy;
|
|
`/docs` stays; SQLite dev mode goes; the monthly retrain stays on the same
|
|
machine (Python, `water-monitor-retrain.timer`).
|
|
|
|
## Cheaper alternative (stay Python)
|
|
|
|
Most of the operational gain comes from taking collection out of the web
|
|
workers: run the collector as its own systemd unit (one process) and the
|
|
uvicorn workers as read-only API. That's about a day of work. It removes leader
|
|
election and per-worker state, but doesn't bring the memory, typing or
|
|
single-binary benefits.
|