Connect to the OSM data pipeline
fix(how-to): add dated Correction callouts to 20 of 28 how-to/ articles, worst ratio of any category — two systemic defects (wrong Doorman A/B/C tier semantics across 5 articles; fictional 4-tier pairing model + /v1/tokens API across 4 articles, real model is 3-role User/Admin/Interface via /v1/invite+/v1/pair), plus fictional CLI commands/routes across 11 more articles for fleet nodes, GIS API, exports, ledger reading, deployment, key rotation, and binary-download auth; several already-self-corrected articles independently re-confirmed accurate
@@ -12,6 +12,8 @@ editor: pointsav-engineering paired_with: connect-osm-data-pipeline.es.md --- **Correction (2026-08-02, verified against canonical `origin/main`):** several specific script/schema claims below are wrong. The real ingest script is `ingest-osm.py`, not `ingest-chain.py` (different CLI shape). The real taxonomy config uses `CATEGORIES`/`BRAND_FILL` dicts keyed by NAICS category (confirmed accurate in the sibling [[collect-location-intelligence-data]] guide), not a `TaxonomyEntry`/`ALPHA_HYPERMARKET` class structure, which doesn't exist. The described "rebuild" step, `build-geometric-ranking.py`, in reality only adds ranking percentiles to an *existing* `clusters.geojson` — the real full rebuild is `build-clusters.py`. The output path should be the absolute deployment path `/srv/foundry/deployments/gateway-orchestration-gis-1/www/data/clusters-meta.json`, not the relative `gateway/www/data/clusters-meta.json` given below. This guide also links to `[[pointsav-gis-engine]]`, archived earlier this session — see [[location-intelligence-substrate]] instead. **Flagged, not resolved.** The platform's location intelligence system ingests point-of-interest (POI) data from OpenStreetMap via JSONL ingest files. Connecting to the OSM data pipeline means writing or adapting an ingest script that queries the Overpass API, producing a JSONL file in the platform's schema, and registering the ingest with the taxonomy configuration. This guide covers a single-chain ingest for a new retail or service category. For the GIS engine architecture, see [[pointsav-gis-engine]]. For building a map from ingested cluster data, see [[build-a-colocation-map]].