Repository contains scripts for fetching and importing State Land Service of the Republic of Latvia open data (addresses and parcels).
Python script main.py downloads and imports Latvian addresses into PostgreSQL database. Data contains parishes,
counties, cities, towns, and streets.
Shell script import-kadastrs.sh
downloads and imports parcels shapefiles (cadastral groups, buildings, engineering structures, parcels, parcel borders,
parcel errors, parcel parts, surveying statuses, and way restrictions)..
- PostgreSQL with PostGIS extension
- Python3 with httpx and psycopg2 modules
- shp2pgsql (shipped with PostGIS), curl, and jq command-line JSON processor for parcel import script.
- Add import of addresses related shapefiles (same dataset, different archive file; contributors welcome, I have no use case for it just now). You can contribute via a PR.
- Add a utility script to download and import of parcels data (there's a lot of shapefiles)
These are welcome (use issues to report or discuss, pull requests to implement).
To fetch and import addresses, create schema (psql schema <schema.sql) and then
run VZD_DBNAME=schema python3 main.py --verbose. For more options see python3 main.py --help.
To download and import all parcel shapefiles, run VZD_DBNAME=schema ./import-kadastrs.sh. This will take a while.
Python script checks against
Latvian address register open data
, using If-Modified-Since header, which it saves to a file for future reference, so data is being downloaded only if
it has been updated. This means that it can be dropped into cron job to download data when it is updated.
If data has been downloaded, it's unzipped into data/csv and then imported into PostgreSQL. Schema has to be created
(it can be found in schema.sql)
If an entry has invalid coordinates (latitude or longitude is not a number), it's skipped.
For aw_eka table column geom is created, and a spatial index is added. SRID 4326 is used, so some offsets may arise.
If you would like to have a denormalized view of the data, a materialized view aw_full_addresses is created. Resulting
dataset contains only existing houses and schema is as follows.
| column | type | description |
|---|---|---|
| code | int | Code, primary key |
| name | text | Name (house number or name, if not on a street) |
| iela_code | int | Street code |
| iela_name | text | Street name |
| ciems_code | int | Vilalge code |
| ciems_name | text | Vilalge name |
| pilseta_code | int | Town code |
| pilseta_name | text | Town name |
| pagasts_code | int | Parish code |
| pagasts_name | text | Parish name |
| novads_code | int | County code |
| novads_name | text | County name |
| full_name | text | Full address of the building |
| parent_code | int | Building's parent code |
| parent_code | int | Building's parent type |
| geom | geom | Geometry (Point) |
Parent type decodes as:
| type | description |
|---|---|
| 101 | Latvijas Republika |
| 102 | Rajons |
| 104 | Pilsēta |
| 105 | Pagasts |
| 106 | Ciems/mazciems |
| 107 | Iela |
| 108 | Ēka, apbūvei paredzēta zemes vienība |
| 109 | Telpu grupa |
| 113 | Novads |
Shell script does nothing fancy. It just fetches JSON metadata, extracts list of all shapefile zip's, downloads them,
then imports into database (via dropping, creating and populating tables). Spatial index on geom is also created.
Data has been released by the State Land Service of the Republic of Latvia under goverment's OpenData initiative, and is available at this data.gov.lv page . It's released under CC BY 4.0 license.
Spec for address data , spec for parcel data.
CC BY 4.0, which means that it can be used for free, however attribution is required, and no additional restrictions on this data can be imposed. This script follows suit.