Savav map and address data Contains information from OpenStreetMap contributors, available under the Open Database License (ODbL) 1.0. https://www.openstreetmap.org/copyright https://opendatacommons.org/licenses/odbl/1-0/ The adapted places.sqlite and address-meta.sqlite contain only public OSM-derived addresses, roads and town labels. It contains no driver routes or customer inputs. Data processing method is provided in this archive to meet ODbL section 4.6(b). This data and its transformations are offered under ODbL 1.0. Inputs downloaded November 16, 2026 (map extract dated November 15, 2026): https://download.geofabrik.de/asia/israel-and-palestine-latest.osm.pbf https://download.geofabrik.de/asia/israel-and-palestine.html https://download1.graphhopper.com/public/asia/isreal/photon-dump-isreal-1.0-latest.jsonl.zst https://download1.graphhopper.com/public/asia/isreal/index.html The upstream latest URLs change over time. Method: 1. Use Python3 with zstandard and osmium; create a fresh work directory and data/ subdirectory. 2. Place source PBF at data/region.osm.pbf and photon JSONL Zstandard file at data/addresses.jsonl.zst. 3. Run build_data.py in that fresh directory. It selects public address candidates, normalizes search terms, rounds road coordinates to 5 decimals and simplifies long road polylines. The script creates data/places.sqlite. It overwrites that generated output, so use a fresh directory. 4. Road routing graph: use GraphHopper Web10.0 and Java17, run java -Xmx1800m -jar graphhopper.jar server routing.yml in data/ with the supplied routing.yml. This imports car-access and turn restrictions from the same PBF into graph-cache. It does not add traffic or truck/security restrictions. GraphHopper source: https://github.com/graphhopper/graphhopper GraphHopper10.0 artifact: https://repo1.maven.org/maven2/com/graphhopper/graphhopper-web/10.0/graphhopper-web-10.0.jar Software notices: Leaflet1.9.4 uses BSD2Clause (static/vendor/LICENSE); Heebo font uses SIL Open Font License (static/vendor/Heebo-OFL.txt). The data license is separate from the website software. Search repair metadata: run build_address_meta.py after build_data.py with the same source snapshot. It verifies the latitude/longitude of every existing row ID, then creates data/address-meta.sqlite retaining the original street field. This prevents neighborhood or business names being mistaken for streets. No new address coordinates are fabricated. address_search.py queries these local files, strictly filters exact house numbers and city, and marks street-only or similar-name results for manual pin confirmation.