/** * @copyright Sister Software * @license AGPL-3.0 * @author Teffen Ellis, et al. * * `geonames`: GeoNames populated-places consumer (https://www.geonames.org/, CC-BY-4.0). * * GeoNames is a global gazetteer of ~12M features. This adapter ingests the POPULATED PLACES * (`feature_class = "P"`, excluding historical/abandoned/destroyed variants) from a per-country * dump file — global locality coverage, including the small towns and villages a coarser admin * gazetteer (WOF) lacks. It's the cheapest path to broadening the corpus's LOCALE coverage. * * Input: a per-country tab-separated dump (e.g. `US.txt` from * `https://download.geonames.org/export/dump/`, 19 columns, no header). Two sibling files in the * same directory supply human-readable names (downloaded once from the same place): * * - `admin1CodesASCII.txt` — `.` → region name (e.g. `US.VT` → "Vermont"). * - `countryInfo.txt` — ISO alpha-2 → country name (e.g. `US` → "United States"); `#`-commented. If a * sibling is missing, the corresponding component is simply omitted (graceful degradation). * * Output: per place, up to two hierarchy variants (mirroring `wof-admin`'s with/without-country * balance so the model sees both domestic and international order) — * * 1. `{ locality, region }` → "City, Region" * 2. `{ locality, region, country }` → "City, Region, Country" `reconcileComponents` drops any * component that didn't survive into the rendered `raw`. * * License: stamped `"CC-BY-4.0"` per row (GeoNames' terms); provenance is the `geonames-` key. */ import type { CorpusAdapter } from "@mailwoman/corpus/types"; /** * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset * it came from. */ export declare const GEONAMES_ADAPTER_ID = "geonames"; /** * License carried by this source (CC-BY-4.0), attached to each row so downstream consumers inherit the terms rather * than having to look them up. */ export declare const GEONAMES_DEFAULT_LICENSE = "CC-BY-4.0"; export declare function createGeonamesAdapter(): CorpusAdapter; /** * The configured adapter instance registered with the corpus builder. */ export declare const geonamesAdapter: CorpusAdapter; //# sourceMappingURL=adapter.d.ts.map