This project has moved to my geotools project

Extract geojson coordinates from wikipedia files

Wikipedia2geojson reads wikipedia files and prints out geo locations.

It can read compressed files, and compressed streams. Works on Linux, MacOS, and MS Windows.

Example

    wikipedia2geojson.exe file.xml.bz2

Reads from file.xml.bz2, automatically uncompressing bz2 format

[
{ "type": "Feature", "geometry": { "type": "Point", "coordinates": [ 2, 28 ] }, "properties": { "name": ""Algeria"" } }
{ "type": "Feature", "geometry": { "type": "Point", "coordinates": [ 30, 42 ] }, "properties": { "name": ""Andorra"" } }
{ "type": "Feature", "geometry": { "type": "Point", "coordinates": [ -150, 64 ] }, "properties": { "name": ""Alaska"" } }
{ "type": "Feature", "geometry": { "type": "Point", "coordinates": [ 19, 13 ] }, "properties": { "name": ""Apollo 11"" } }

Each location is on its own line, so you can pipe this stream into grep and other command line programs. Add the --strict flag if you want completely correct geojson format.

Streaming

Because w2g can read compressed streams, you can process network files on the fly. You don't need to download them completely first.

wget -q -O - http://someserver.com/enwiki-pages-articles2.xml.bz2 | wikipedia2geojson --compression=bz2 -

e.g. from the wikipedia download site (don't do this, it's better to download the file once and use it)

	wget --no-check-certificate -q -O - https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles1.xml-p10p30302.bz2 | ./wikipedia2geojson --compression=bz2 --strict -

Installation

    go get -u github.com/donomii/wikipedia2geojson
	go install github.com/donomii/wikipedia2geojson

More examples

    wikipedia2geojson.exe file.xml

            Read from file.xml


    wikipedia2geojson.exe file.xml.bz2

            Read from file.xml.bz2, automatically uncompressing bz2 format


    wikipedia2geojson.exe file.xml.gz

            Read from file.xml.bz2, automatically uncompressing gz format


    wikipedia2geojson.exe --compression=bz2 file

            Read from file, force uncompressing bz2 format


    wikipedia2geojson.exe --compression=gz file

            Read from file, force uncompressing gz format


    wikipedia2geojson.exe -

            Read from stdin.


    wikipedia2geojson.exe --compression=bz2 -

            Read from stdin.  Stdin is in bzip2 format


    wikipedia2geojson.exe --compression=gz -

            Read from stdin.  Stdin is in gz format

Known bugs

You can't stream straight from the network on MS Windows, because Windows fiddles with the data as it goes through the pipe, and most download programs don't know how to stop that.

So this won't work

wget -q -O - http://someserver.com/enwiki-pages-articles2.xml.bz2 | wikipedia2geojson.exe --compression=bz2 -

W2g does not print out fully compliant geojson. Instead of printing an array of points, it just prints the points. To change the output into fully compliant geojson, add the --strict flag to the command line.

Bonus

A perl one liner to unpack the wikipedia geodata files in sql format

type enwiki-20171103-geo_tags.sql | perl -pe "s/\),\(/\r\n/g" | perl -ne "@c=split/,/;if($c[8]ne'NULL'){print '{ \"type\": \"Feature\", \"geometry\": { \"type\": \"Point\", \"coordinates\": [ '.$c[4].', '.$c[5].' ] }, \"properties\": { \"name\": '.$c[8].' } };'.\"\n\";}"

cat enwiki-20190501-geo_tags.sql | perl -pe "s/\),\(/\n/g" | perl -ne '@c=split/,/;if($c[8]ne"NULL"){print "{ \"type\": \"Feature\", \"geometry\": { \"type\": \"Point\", \"coordinates\": [ ".$c[4].", ".$c[5]." ] }, \"properties\": { \"name\": \".$c[8].\" } }"."\n";}'

Wikipedia's geodata extraction appears to have trouble identifying points, so wikipedia2geojson will be useful for a bit longer yet.