Overpass API

Querying an Overpass API endpoint

Python

The following snippet gives an idea how an Overpass API endpoint can be queried with Python:

import requests

query = """
[out:csv(::id,name)];
relation[admin_level="2"];
out;
"""

response = requests.get(
   'http://overpass-api.de/api/interpreter',
    headers = {'Accept-Charset': 'utf-8'},
    params  = {'data'          :  query }
)

response.encoding='utf-8'

with open('overpass-result.csv', 'w', encoding='utf-8') as file:
     file.write(response.text)

There is also the overpy library.

PowerShell

The following snippet gives an idea how an Overpass API endpoint can be queried with PowerShell:

$query = @'

  [out:json];
  nwr [name=Freienstein];
  out;

'@

$res     = invoke-webRequest -method post -body (@{'data' = $query}) https://overpass-api.de/api/interpreter
$resTxt  = $res.Content
$resJson = convertFrom-json $resTxt

foreach($elem in $resJson.elements) {
   "id: $($elem.id) ($($elem.type))"

   if ($elem.type -eq 'node') {
      "  $($elem.lat) $($elem.lon)"
   }

   …
}

File extension

The proposed file extension for files that store Overpass API queries is *.overpassql.

TODO

Calculating distances

Although there is a length() function, the calculation of distances (notably between two nodes) seems impossible.

Find buildings whose house number is 13

{{geocodeArea:Pfungen}};              // Search in Pfungen

nw                                    // Search nodes and ways
  [building]                          // that have a building and
  ['addr:housenumber']                // an addr:housenumber tag.
  (area)                              // Restrict search to selected area (geocodeArea)
  (if: t['addr:housenumber'] == 13 ); // and keep only those objects whose addr:housenumber value is 13

out geom;

olbricht.nrw/ovt

olbricht.nrw/ovt is a temporary instance of Overpass Turbo (and seems to have an implementation that translates osm to GeoJSON).

The source code is hosted on github (osmtogeojson)