A couple of scripts for parsing OpenStreetMap .osm.pbf files, to extract the amenities contained within.

You will need at least the JAVA_JRE, properly installed and in your path. Check with:

$ echo $JAVA_HOME
/path/to/jdk1.8.0_111.jdk/Contents/Home

You also need osmosis to do the heavy lifting of converting binary .pbf files to extracted XML. Osmosis can be found here .

Get the pbf files required.

chmod +x ./update.sh
./update.sh

This will proceed to check MD5 hashes between the known sources, and the local copy. If they match, no update is required. If they don’t match, the pbf file is downloaded.

The ./extract.sh file can be made +x:

chmod +x ./extract.sh

Then, perform the parsing with the following format:

extract.sh -f <filename> -l <location> -a <amenity>

Ex:

extract.sh -f ./pbf/manitoba-latest.osm.pbf -l Manitoba -a restaurant

You’ll then receive a <location>-<amenity>-all.osm file!

The included ./parse.py can also be made +x:

chmod +x ./parse.py

Then, perform the parsing with the following format:

parse.py <filename>.osm "[filter]"

Eg:

parse.py ./Manitoba-restaurant-all.opm

or

parse.py ./Manitoba-restaurant-all.opm "McDonald's"

The parse.py file currently lists all Nodes effectively; extracting name, latitude, and longitude coordinates. Way types are dealt with separately; first by also ensuring the reference nodes are collected, and then compared with the first Way node. While this will generate a point which is not located in the centre of that item, it will at least define the starting place, and be pretty darn close.

• Error handling