ECCC GeoMet API (ECCC_API)

Upstream: ECCC GeoMet OGC API - Features collections (base URL https://api.weather.gc.ca/collections). Station GeoJSON is fetched per station and collection; hydrometric vs meteorological station lists use separate CSVs (station_csv_hydro / station_csv_meteo or legacy station_csv).

Transport: requests in coldcast.sources.eccc_api.download(). Not the generic coldcast.download.download_requests() path.

Auth: None for public GeoMet API.

CLI: coldcast download eccc_api (all configured collections), or coldcast download eccc_api --model <collection-name> to run a single collection (for example hydrometric-realtime, swob-realtime, or climate-daily / climate-hourly). climate_daily and climate_hourly are also accepted. With --model, only that collection is built and downloaded. Optional --stations-csv PATH uses that file instead of station_csv_hydro / station_csv_meteo (or legacy station_csv) from settings; the same path is used for both hydro and meteo station lists, so the file must include the column each collection expects (typically ID or MSC_ID, or a per-collection station_column override).

Output: By default, wide CSV suitable for Delft-FEWS import (see bundled default_settings.yaml under fews_csv). Layout long and column maps are configurable.

GeoMet returns at most 10000 features per /items request. For both hydrometric (STATION_NUMBER) and meteo collections, Coldcast follows OGC rel=next links until every page is fetched (fetch_all_pages), up to max_pagination_pages. The limit setting is the per-request page size (capped at 10000). With sort_output_datetime_descending: true (default in bundled settings), exported CSV rows are ordered newest first so recent observations appear at the top.

Settings: ECCC_API — collections, fews_csv, limit, fetch_all_pages, max_pagination_pages, sort_output_datetime_descending, time_range, http_timeout_seconds.

Code: coldcast.sources.eccc_api, coldcast.sources.eccc_fews_csv